TechBongo Solutions

Is vWorker.com Hacked? I see an XSS Attack Alert on Login Page

2011-11-22T12:24:00.001-08:00

When I click on the Login link on vWorker.com following page is displayed. Really worried... Is it hacked??!!

Tips For Sales Conversion In Debut Adwords Campaigns

2011-04-09T00:11:00.000-07:00

The most amazing thing is Google is offering Gift Coupon worth Rs 2500/- of Adwords for free to promote their most profitable service. And new webmasters around are taking full benefits of this promotional offer. It helps them starting up pay-per-click advertising for their websites. It significantly boosts traffic to the website who never managed to bring even 10 real visitors a day in past. But this debut PPC campaign hardly results in conversion. They fails to get Sales from these 'free' visitors. Here in this article I'll try to highlight few of its reasons:

Insufficient idea of PPC vitals like Cost Per Click rate, Audience Targeting, Keywords etc. An online advertising expert can help you choose best keywords, cost effective and perfect CPC rates etc.
First PPC Campaign - Too catchy Ad Title and Description. Ultimately results in 100% disappointment of visitors.
False offerings, tricky offers, hidden charges, silenced terms and conditions. I feel amazed to see those $ 3/Month unlimited hosting offers; you get $5-$7 per month discount when you place an order for 5 years or even more.
Unknown Brands, New Agencies are depot of unreliability. Buyers never like to buy from new providers; even I often prefer to read atlest 4-5 public reviews before spending 1000+ rupees online.
Unavailability of a proper landing page for PPC campaigns. And the potential customers reaches a page where they get a huge 'Order Now' button but no 'Detailed Product Description', 'Warranty Terms', 'Support Availability', 'Clients' Testimonials', 'Picture of Items' etc. All they need to do is to digg the site to collect all these details. But surely they won't do that since many other sites appear in the same Google Search result and they'll open the next link.
Lack of online ordering system or instant order processing commitments. I've often seen sites offering $5 product; many of the visitors won't hesitate to give a try IF an online ordering or instant activation system is available
Not availability of Live Chat / Phone Support when campaigns are running. Because buyers will always have questions and you need answer all of them to get conversion.
Server outage, Server Busy issue. Although it's not usual but your homely web server should be able to handle sudden huge traffic. You need to choose a good web hosting company, who providers high-performance servers.
Language problem for improperly targeted campaigns. A Portuguese won't spend more than 2 seconds to close the browser window of your Vietnamese website.
Poor appearance of website, incomplete sites. Last but not least is your overall website outlook. A professional website developer can make a beautiful website, which itself attracts visitors. Visitors like to bookmark your site, 'Stumble' it, 'Twitt' it or to return back to your site if it's nice, effective, complete and extra-ordinary.

In my opening, if you try to follow the points mentioned above, you'll surely be able to bring out significant sales out of these 'free visitors'.

PHP Class to Compose and Generate (Export) Excel Files in XLS Format

2010-10-29T13:30:00.000-07:00

TechBongo Solutions has developed a new PHP class (can be termed as a PHP component) to export contents into spreadsheets. This class can be used to compose and generate xls files on the fly. Even you can add formula to the excel file. It's really easy to use.

This class can be used to fetch data from the database and export them into xls files (Microsoft Excel). Previously, we were using any to csv export feature. But we found number of issues, while using this feature in real time applications. For example, we couldn't add commas (",") into any cell of the exported csv file. But xls file overcomes all these limitations and works fine. You can open these exported xls files in any version of MS Office, Open Office and Google Spreadsheets.

Download the XLS PHP Class and generate XLS files using PHP.

Direct Download Link:

XLS PHP Class

How To win first Bid Request (job) on vWorker.com

2010-08-15T07:49:00.000-07:00

Some people take a bit of time to land their first job. Others find instant success. What's the difference between the two? PhpORcaffeine shares the techniques that enabled him to win a job on his first DAY on the site...and land five more during just his first week! (...at the time of writing this article, he had won only two of the five).

I created my account on Dec. 8th 2006 and I won and completed my first job on Dec 8th 2006. Right now I have two other bids that I have won and am working with.

I'm not naive though, I know it isn't always that easy but I do believe that some things I did directly contributed to my quick success and I would like to share those methods with you (for article / knowledge base purposes).

Your first bid on a project SHOULD NOT be a monetary bid, you want the employer to see that you are interested in their project, not just their money. Ask questions and make sure you get clear explanations to EVERYTHING. Make sure everything is in the project message thread. If the employer doesn't respond to your first bid (comment) then I would consider pursuing a different project. If the employer does respond then by all means continue. Once you have a CLEAR definition of the scope of the project, THEN make your monetary bid (it's always easier to back out before you make a monetary bid and get accepted).
Do not copy/paste your bid's or use generic bid text. I wouldn't buy services from a bot (or generic text), neither will most other people. The employer gets the sense that you didn't even bother to read the description and is usually hesitant to select you. Some will just ignore you. I want a worker that will reference things I said in my description, that way I know they read the description that I took time to write and that I am talking to a real person.
DO NOT bid on projects that you are 'iffy' about. Unless you are 100% sure that you CAN and WILL do it, don't make a monetary bid on it. Bidding on things you can't or don't end up doing only serves to hurt the effectiveness of thevWorker.com system and hurts your reputation if the project goes to arbitration.
Listen and follow the vWorker.com rules about communicating off site DO NOT DO IT. I realize off site communication is sometime faster and most of you will still do it. My rule of thumb is that if you go off site, do not discuss details or additions/subtractions to the project. Save that for the vWorker.com project message thread.

Above all, treat the vWorker.com system and vWorker.com users on the system as co-workers in an office environment. With discipline, this can become a full time job! Treat it as such!

How To win first Bid Request (job) on vWorker.com

2010-08-15T07:41:00.000-07:00

Your first bid on a project SHOULD NOT be a monetary bid, you want the employer to see that you are interested in their project, not just their money. Ask questions and make sure you get clear explanations to EVERYTHING. Make sure everything is in the project message thread. If the employer doesn't respond to your first bid (comment) then I would consider pursuing a different project. If the employer does respond then by all means continue. Once you have a CLEAR definition of the scope of the project, THEN make your monetary bid (it's always easier to back out before you make a monetary bid and get accepted).
Do not copy/paste your bid's or use generic bid text. I wouldn't buy services from a bot (or generic text), neither will most other people. The employer gets the sense that you didn't even bother to read the description and is usually hesitant to select you. Some will just ignore you. I want a worker that will reference things I said in my description, that way I know they read the description that I took time to write and that I am talking to a real person.
DO NOT bid on projects that you are 'iffy' about. Unless you are 100% sure that you CAN and WILL do it, don't make a monetary bid on it. Bidding on things you can't or don't end up doing only serves to hurt the effectiveness of thevWorker.com system and hurts your reputation if the project goes to arbitration.
Listen and follow the vWorker.com rules about communicating off site DO NOT DO IT. I realize off site communication is sometime faster and most of you will still do it. My rule of thumb is that if you go off site, do not discuss details or additions/subtractions to the project. Save that for the vWorker.com project message thread.

Above all, treat the vWorker.com system and vWorker.com users on the system as co-workers in an office environment. With discipline, this can become a full time job! Treat it as such!

Banking CRM, Retail Banking | Chordiant

2010-08-12T12:22:00.000-07:00

Banking CRM, Retail Banking | Chordiant: "Solutions: Retail Banking
Grow customers and account value with customer-focused conversations

When it comes to banking CRM solutions, retail banking firms face a host of challenges if they are to achieve higher customer satisfaction and profitability. Multiple products and channels, increased regulation, rapid mergers and acquisitions, high staff turnover and geographically dispersed operations all increase the difficulty of delivering customer-centric experiences that promote higher customer loyalty and increase lifetime value.

Chordiant solutions for banking CRM help you make sure every customer experience is unique, relevant and appropriate to your business goals. Chordiant solutions unify customer data, analyze customer behavior, dynamically guide representatives, automatically generate personalized recommendations and enable real-time monitoring and adaptation. Whether customers are opening an account, changing account information or receiving an offer, Chordiant solutions for banking CRM help you optimize the conversation to increase customer loyalty and wallet share, while reducing the cost of each interaction."

Brand New Human Resource & Project Management Tool ... Coming Up Next July

2010-05-08T09:52:00.000-07:00

Hi well wishers,

Thank you all for being with us in this long journey. After long time we're feeling once again to write on this blog. We're really happy to announce that finally TechBongo Solutions is now committed in the field of product development. Apart from our website development, web hosting, domain registration and software and web based consultancy services, we're launching our brand new product next July. The product is one of the most useful tools for small and medium business concerns. This tool integrates a Project Management System, Time and Expense Management System and a Human Resource Management System for your company. We're sure, there are some companies with some 10-200 employees, who are still now using papers for storing employee details, HR related information and project related documents. Our tool will bring them all under a same umbrella. In this blog, unfortunately, we won't reveal the name of the application. This is our trade secret. But it's our promise, we'll show you screenshots, demo and more detailed features within one month.

Stay with us and enjoy the Excellence!

Changing Default User Agent String Using PHP

2009-10-03T15:04:00.000-07:00

It's easy to change the default user agent string of your php script (web server). Just use this piece of code in your crawler/scrapper script.

ini_set("user_agent","TechBongo Crawler V1.0");
?>

If you want to test this piece of PHP code, write the following PHP code in a PHP file and execute the file. Now check the visitor log of the web page and you'll see your custom user agent string there, instead of the default user agent string.

ini_set("user_agent","Your User Agent String");

$data = file_get_contents("http://you-page-url.com",0);

Print strip_tags($data,"");

?>

Free Reciprocal Link Checker Script in PHP

2009-07-30T11:44:00.000-07:00

This time, we're publishing a short simple code for backlink checking. Webmasters, who are doing lot of link exchange via forum, chats, mails etc. often need a reciprocal link checker script. Because, often one of the link exchanging party removes other's backlink from his website. Now, it's really hard for you to test each and every reciprocal link for your website. This PHP script will make your work easier.

You can edit the files as per your requirement and add features like database connectivity.

2 files are used to write the script. One is a backlink checking class and the other one is an example file.

Code for PHP Class:

= 4.0.5*
---------------------------------------------------------------------*
EXAMPLE:** [ TO PARSE A STRING ]* $myLinks = &new LinkExtractor(); // create
a LinkExtractor Object* $myLinks->parseString( $myStringWithLinks ); // parse
a string* for( $a = 0, $b = count( $fetchLinks = $myLinks->getLinks() ); $a
< $b; $a++ ) {* echo $fetchLinks[$a]." ";* }** [ TO PARSE AN URL ( or if you want, a file ) ]* $myLinks = &new LinkExtractor(); // create a LinkExtractor Object* if( $myLinks->parseUrl( "http://www.techbongo.com/index.php"
) == true ) {* for( $a = 0, $b = count( $fetchLinks =
$myLinks->getLinks() ); $a < $b; $a++ ) {* echo $fetchLinks[$a]." ";* }* }** [ TO PARSE A FILE ]* $myLinks = &new LinkExtractor(); // create a LinkExtractor Object* if( $myLinks->parseFile( "myTextFile.txt" ) == true ) {* for( $a = 0, $b =
count( $fetchLinks = $myLinks->getLinks() ); $a < $b; $a++ ) {* echo $fetchLinks[$a]." ";* }* }* ---------------------------------------------------------------------class LinkExtractor { /* private Array variable: $linkReg [ contains pregs to parse links ]*/ var $linkReg = Array( "/(?i) /** * Public constructor. * Create a global
Array with no value, used for parsing * and an internal array with valid
pregs for links parsing. */ function LinkExtractor()
{ global
$__linkExtractor_linkRecipient; $__linkExtractor_linkRecipient =
Array(); }
/** * Private method, popolate internal Array with
preg matches * . * @Param String String to push into
internal array *
@Return nothing */ function
__manageLinkRecipient( $replacement ) { global
$__linkExtractor_linkRecipient; array_push(
$__linkExtractor_linkRecipient, htmlspecialchars( $replacement[2] )
); }
/** * Private method, call preg_replace_callback
function with string. * . * @Param String String to
parse * @Return nothing */ function __callBackCaller( $st )
{ preg_replace_callback( $this->linkReg, Array( &$this,
'__manageLinkRecipient' ), $st ); }
/** * Public method, read
remote page or file and parse them * . *
@Param String valid url address to parse *
@Return Boolean true if readed , false in other
cases */ function parseUrl( $url ) { if( @$fp = fopen(
$url, "r" ) ) { $st = ''; while( $text =
fread( $fp, 8192 ) ) { $st .=
$text; } fclose( $fp
); $this->__callBackCaller( $st ); return
true; } return false; } /** * Public
method, parse links in a file * . *
@Param String string to parse *
@Return nothing */ function parseFile( $st ) { return
$this->parseUrl( $st ); }
/** * Public method, parse links
in a string * . * @Param String string to parse *
@Return nothing */ function parseString( $st )
{ $this->__callBackCaller( $st ); } /** *
Public method, return an array with all found links * . *
@Param no no params need * @Return Array Array with all
links ( if there're ) */ function getLinks() { global
$__linkExtractor_linkRecipient; return
$__linkExtractor_linkRecipient; }}?>

Code For Example File:

"LinkExtractor.class.php" );
function linkExtractor( $what, $url = false )
{
// linkExtractor( $string_url_or_file [, true/false ] ) // true
if $string_or_url is an url or a file // false or anything if is a
string $myLinks = &new LinkExtractor(); if( $url == false )
{ $myLinks->parseString( $what ); } elseif( $url == true
) { if( $myLinks->parseUrl(
$what ) == false ) {
return false; } } return
$myLinks->getLinks();}
// EXAMPLE$url = "http://techbongo.com/resources/"; //
site to parse or check the reciprocal link
$testURL="http://techbongo.com/"; // the reciprocal url
to check
$arrayLinks = &linkExtractor( $url, true );
if( $arrayLinks
!= false ) { for( $a = 0, $b = count( $arrayLinks ); $a < $b; $a++ )
{ if($arrayLinks[$a]==$testURL) { echo
$arrayLinks[$a]."
"; //you can add your piece of code
here. if a reciprocal link is found, the link will be available via
$arrayLinks[$a] } }}?>

Hope this script helps you. You can download the files from the link given below:
Download Free Reciprocal Link Checker Script Using PHP

Free Reciprocal Link Checker Script Using PHP Simple Backlink Checker Script PHP Code To Check Reciprocal Link Link Exchange Script

Tricks For Better Adsense Targeting and High CTR

2009-07-10T12:49:00.000-07:00

What is AdSense?
AdSense ads are ads displayed on your site, on position where provided code is added. That is pay per click model, what means that every click on links/banner is cash in your pocket. Adsense ads are automatically targeted according to theme (keywords) of your page, and on the first sight, it is looking easy to monetize your site on this way. But, it is not that easy.
Maybe some of readers will be disappointed, but purpose of this article is not how to buy your first car from adsense checks. Instead of that, this article can help you to pay hosting expenses at least, making happy Google, visitors, advertisers and you.
So, forget MFA (made for adsense) sites, high paying keywords and arbitrage, for this kind of optimization, there is only small effort required.

AdSense Preview tool
High traffic page, with no matching product or CPA offer to cover theme of that page, is what we need for test.
For initial check, you need to download AdSense preview tool, i.e plug-in for Internet Explorer. that is small *.reg file and installation is easy (double click on file).
So, go to AdSense Preview tool page, read instructions, download and install that plugin. More about AdSense Preview tool you can read on this page
Installed? Now, go to that test page. If that page is already indexed, and you can find it in Google index (search engine results), it is better. Now, right mouse click (anywhere on page), select Google AdSense Preview Tool, and, after small window is displayed, select country (see pic below, geographic targeting). At this time, Google spider is checking your page for keywords, according to those keywords (theme), you will see ads inside preview tool window. Now, compare ads inside widow with your page theme.
Below is screenshot of adsense preview tool window, taken after uploading this page. As you can see, ads are related to this page theme.

Web Page optimization
If your test page receives traffic from search engines, every touch can change your position. If it is not that important, you can change or replace word here and there, using preview tool as optimization tool. According to ads displayed, you can guess what keywords are targeted. Now, tweak your page in small steps to get ads what will match theme, adding or replacing keywords in title, headings and body (top middle, bottom).

What ads are displayed on site page?
If your site is targeted to USA visitors, and you are from outside USA, you will not see same ads as USA visitor. You have to use proxy in targeted country to control what ads are displayed to your targeted visitors.
Now, to fine tune ads displayed on your site, get leather board or skyscraper code and add it somewhere on your page. If it is necessary, connect using proxy. Look at sites displayed in ads block. High paying ads are on top, but if you want to remove some sites from that block, use "competition filter" on your adsense account and add those domains there.

Ads position
Net is covered with advises for adsense ads position, how to blend ads into content, how to push visitor to click on ad...
In our test page, purpose to add adsense block is to fill unused space, and to get some return for our effort. In that case, we still want to keep our visitor, not to push him out for a few cents. That is different story, site content is what is attracting visitors, not ads. Overloading site with adsense ads is not recommended, not only due to lower payment from lower positioned ads. Adsense ads are all over the Net, and they are recognized as banners before. If you care for your visitors, do not blend (mix) ads with content, so that they are looking like site menu (like I did it on bottom of this page), or links to the part of your site. Visitor don't like to feel cheated or used (you too?).
So, we will offer adsense to visitor as an option. On the end of this article, as example, is one block of ads, what could be interesting for you, or maybe not. If you are reading this now, that means that content is interesting, and your next focus could be on those ads. And, if ads are positioned in the middle of this article, or on the right side of this page will be (probably) skipped. The catch is to present visitor something what will offer more information about theme. In that case ads will be effective and visitor will not leave your site forever.

Most successful AdSense ads formats
Simple enough, three factors are important to target adsense ads better and to get better CTR (click-through rate.). Ads format, page theme (keywords), and Ads block position on page. These examples below are not suitable for websites where you are trying to build trust, and keep visitors. On marketing theme websites, where visitor is familiar with AdSense, my suggestion is to use other kind of advertising, due to possible low CTR.
First, to blend ads better, font size should the same as AdSense ads (10px), link color the same as links on your site. Most used type is the large block. Position of two blocks above the fold (looking like only content on the page) shows very high CTR (MFA sites). On directory type websites, this large block could be blended in directory listing (on the end of listing). Also, on this type of websites, a few single ad blocks can be mixed into listing. (between listed websites)
One more format, often used on "visitor, go away" websites is "menu". Positioned on top left corner of page, where visitor expect to see links to other sections of website, it gets good CTR if it is blended good in website menu (positioned above real menu). Horizontal menu block could be added on top or bottom of page.
Along with those three factors mentioned above, more are in the game. Colors on your website, hot spots and content are playing here also. The best way to find good ad format and position on page is testing. Read below.

AdSense tracking
For simple tracking (CTR) you need to create channel. So, for your test page, create channel for every ad format you want to test on that page. Testing ads on that page, you will spot most effective format and position. Don't forget, if some format is working on one site or web page, it doesn't mean that it is working on every site. Test adsense format and position on your site until you get best CTR.
There some JavaScripts available, what can help to find what ads are clicked, and , according to that, some ads can be excluded using "competition filter", leaving only ads what attracts clicks, but that is an other story, not covered here.
Anyway, to target AdSense ads better, you need to tune your page for related set of keywords, especially those paragraphs above and below block where AdSense ads are displayed. And, those links at bottom, if they are, on the first sight, looking like menu, that is not menu.

Facebook-Style Modal Box Using MooTools

2009-07-01T22:46:00.000-07:00

In my oh-so-humble opinion, Facebook’s Modal box is the best modal box around. It’s lightweight, subtle, and very stylish. I’ve taken Facebook’s imagery and CSS and combined it with MooTools’ awesome functionality to duplicate the effect.
Facebook uses a funky sprite for their modal box. Pretty cool though.

/* from facebook */
.generic_dialog { height:0; left:0;
overflow:visible; position:fixed; /*dw*/ top:0; width:100%; z-index:101;
}
#generic_dialog_iframe { left:0; position:absolute; top:0; z-index:3;
}
.generic_dialog .generic_dialog_popup { height:0; overflow:visible;
position:relative; }
.generic_dialog div.dialog_loading {
background-color:#F2F2F2; border:1px solid #606060; font-size:24px;
padding:10px; }
#generic_dialog_overlay { display:block; left:0;
position:absolute; top:0; width:100%; z-index:100; }
.dialog_body
.dialog_content_img { float:left; margin-right:15px; }
.dialog_body
.dialog_content_txt { float:left; padding-bottom:5px; width:300px;
}
.dialog_body .dialog_content_body { padding-bottom:13px; }
.dialog_body
.form_label { padding-right:5px; }
.dark_dialog_overlay {
background-image:url(facebook-overlay.png); background-repeat:repeat; }
*
html .dark_dialog_overlay { background-color:transparent;
background-image:url(blank.gif); }
.full_bleed .pop_dialog_table
td.pop_content .dialog_body { padding:0; }
table.pop_dialog_table {
border-collapse:collapse; direction:ltr; margin:auto; table-layout:fixed;
width:465px; }
td.pop_topleft, td.pop_topright, td.pop_bottomleft,
td.pop_bottomright { height:10px; overflow:hidden; padding:0 !important;
width:10px !important; }
td.pop_topleft { background:transparent
url(facebook-pop-dialog-sprite.png) no-repeat scroll 0 0; }
td.pop_topright {
background:transparent url(facebook-pop-dialog-sprite.png) no-repeat scroll 0
-10px; }
td.pop_bottomleft { background:transparent
url(facebook-pop-dialog-sprite.png) no-repeat scroll 0 -20px;
}
td.pop_bottomright { background:transparent
url(facebook-pop-dialog-sprite.png) no-repeat scroll 0 -30px; }
td.pop_top,
td.pop_bottom { background:transparent url(facebook-pop-dialog-sprite.png)
repeat-x scroll 0 -40px; }
td.pop_side { background:transparent
url(facebook-pop-dialog-sprite.png) repeat-y scroll -10px 0; }
td.pop_content
{ background-color:white; direction:ltr; padding:0; }
.pop_dialog_rtl
td.pop_content { direction:rtl; }
td.pop_content h2.dialog_title {
background:#6D84B4 none repeat scroll 0 0; border:1px solid #3B5998;
color:white; font-size:14px; font-weight:bold; margin:0; }
td.pop_content
h2.dialog_loading { background:#6D84B4 url(facebook-indicator_white_small.gif)
no-repeat scroll 400px 10px; padding-right:40px; }
td.pop_content h2 span {
display:block; padding:4px 10px 5px; }
td.pop_content .dialog_content {
background:#FFFFFF none repeat scroll 0 0; border-color:#555555;
border-style:solid; border-width:0 1px 1px; }
td.pop_content .dialog_body {
border-bottom:1px solid #CCCCCC; padding:10px; }
td.pop_content
.dialog_summary { background:#F2F2F2 none repeat scroll 0 0; border-bottom:1px
solid #CCCCCC; padding:8px 10px; }
td.pop_content .dialog_buttons {
background:#F2F2F2 none repeat scroll 0 0; padding:8px; text-align:right;
}
td.pop_content .dialog_buttons input { margin-left:5px; }
td.pop_content
.dialog_buttons_msg { float:left; padding:5px 0 0; }
td.pop_content
.dialog_footer { background:#F2F2F2 none repeat scroll 0 50%; }
/* david
walsh custom */
#fb-modal { display:none; }
#fb-close { cursor:pointer;
}
.info { width:280px; float:left; font-size:11px; color:#666; }
.info b {
color:#000; }
.image { width:200px; float:left; margin-right:10px; }

Most of this CSS was copied straight from Facebook’s stylesheets. You’ll see a bit of custom CSS at the bottom but that’s just me customizing the content in the modal box.

window.addEvent('domready',function() {
/* hide using opacity on page load
*/
$('fb-modal').setStyles({
opacity:0,
display:'block'
});
/*
hiders */
$('fb-close').addEvent('click',function(e) {
$('fb-modal').fade('out'); });
window.addEvent('keypress',function(e) {
if(e.key == 'esc') { $('fb-modal').fade('out'); }
});
$(document.body).addEvent('click',function(e) {
if($('fb-modal').get('opacity') == 1 &&
!e.target.getParent('.generic_dialog')) {
$('fb-modal').fade('out');
}
});
/* click to show */
$('fb-trigger').addEvent('click',function()
{
$('fb-modal').fade('in');
});
});

Just because Facebook has more resources than you doesn’t mean your site can’t look as good as Facebook!

Creating Thumnails Of JPEG Images Using PHP GD2 Library

2009-06-28T21:56:00.000-07:00

We were doing a website development project of Goenka Jewellers Private Limited. The website contained a image gallery to display their design collections. We used this function to create thumbnails of large JPEG images. The function is using PHP GD2 Library to create thumbnails of JPEG Images. This is why thumbnails are not very high in quality.

function make_thumb($src,$dest,$desired_width)
{

/* read the source image */
$source_image = imagecreatefromjpeg($src);
$width = imagesx($source_image);
$height = imagesy($source_image);

/* find the "desired height" of this thumbnail, relative to the desired width */
$desired_height = floor($height*($desired_width/$width));

/* create a new, "virtual" image */
$virtual_image = imagecreatetruecolor($desired_width,$desired_height);

/* copy source image at a resized size */
imagecopyresized($virtual_image,$source_image,0,0,0,0,$desired_width,$desired_height,$width,$height);

/* create the physical thumbnail image to its destination */
imagejpeg($virtual_image,$dest);
}

Free Website Templates Download Part 1

2009-06-24T00:40:00.000-07:00

Game zone template

Download
Preview

Hair salon template

Download
Preview

Fashion template

Download
Preview

Music shop template

Download
Preview

Travel template

Download
Preview

Transport template

Download
Preview

Beauty salon

Download
Preview

Charity template

Download
Preview

Architecture template

Download
Preview

Media template

Download
Preview

Food online template

Download
Preview

Dog care template

Download
Preview

Gadgets template

Download
Preview

Clothes fashion template

Download
Preview

Model portfolio template

Download
Preview

Music store template

Download
Preview

My family template

Download
Preview

Beauty secrets template

Download
Preview

Party time template

Download
Preview

College music template

Download
Preview

Architecture template

Download
Preview

Counterstrike template

Download

Preview

Video Streaming Technique

2009-05-24T01:08:00.000-07:00

A Robust Abstraction for First-Person Video Streaming:
Techniques, Applications, and Experiments
Neil J. McCurdy

Abstract. The emergence of personal mobile computing and ubiquitous wireless
networks enables powerful field applications of video streaming, such as
vision-enabled command centers for hazardous materials response. However, experience
has repeatedly demonstrated both the fragility of the wireless networks
and the insatiable demand for higher resolution and more video streams. In the
wild, even the best streaming video mechanisms result in low-resolution, lowframe-
rate video, in part because the motion of first-person mobile video (e.g.,
via a head-mounted camera) decimates temporal (inter-frame) compression. We
introduce a visualization technique for displaying low-bit-rate first-person video
that maintains the benefits of high resolution, while minimizing the problems typically
associated with low frame rates. This technique has the unexpected benefit
of eliminating the “Blaire Witch Project” effect – the nausea-inducing jumpiness
typical of first-person video.We explore the features and benefits of the technique
through both a field study involving hazardous waste disposal and a lab study of
side-by-side comparisons with alternate methods. The technique was praised as a
possible command center tool, and some of the participants in the lab study preferred
our low-bitrate encoding technique to the full-frame, high resolution video
that was used as a control.
1 Introduction
The emergence of personal mobile computing and ubiquitous wireless networks allows
for remote observation in uncontrolled settings. Remote observation is powerful in situations
in which it is not possible or too dangerous for an observer to be present at the
activity of interest. These include coverage of breaking news, emergency response, or
grand parents joining the grandchildren on a trip to the zoo. The application investigated
in this paper is video-support for a supervisor overseeing hazardous materials disposal.
Despite incredible advances in wireless networking and the mobile devices connected
by it, our repeated experience is that wireless networks in uncontrolled settings
are fragile, and there is seemingly unlimited demand for more video streams
at higher resolution. Modern video streaming techniques heavily depend on temporal
(inter-frame) compression to achieve higher frame rates, while minimizing the impact
on resolution when operating at the network’s capacity. Unfortunately, the panning motions
common to first-person mobile video (captured from a headcam, say) virtually
Fig. 1. Snapshots of two transitions in progress. The top row depicts a camera pan from left
to right where the frames do not overlap. The bottom is a morph from the frame on the left to
the frame on the right as the camera pans down and to the right to look at the child. The live
experience is one of smooth camera movement.
eliminates inter-frame compression. To stay within the available bandwidth, either the
frame rate or the resolution must be reduced. In applications like hazardous materials
disposal, image resolution cannot be sacrificed, making a low-frame-rate encoding
the only viable option. Ironically, lower frame rates further reduce the likely overlap
between frames, further reducing inter-frame compression.
The problemwith low-frame-rate video is that a one-second interval between frames
is long enough to disorient the viewer. This is especially true with head-mounted cameras
because it may only take a fraction of a second for the view to rotate 180 degrees.
With little or no overlap between successive frames, the viewer lacks the information
required to understand how the frames relate to one another. Even in a relatively unchanging
outdoor environmentwhere there is a large field of view, a viewer can become
disoriented looking at the camera’s view of the ground when the camera operator looks
down to avoid obstacles.
In this paper, we present a visualization technique that minimizes the confusion
caused by low-frame-rate video, using modest hardware and processing. If the orientation
of the camera is known – either by attaching tilt sensors and an electronic compass
to the cameras, or by using an online vision processing algorithm on the cameras –
we can generate a visualization that shows the viewer how subsequent frames relate to
one another. The visualization takes the form of a dynamic transition similar to those
described for switching between two streaming cameras located in the same environment
[1]. A transition (Fig. 1) has two components:movement (rotation) of the viewing
portal from one frame to the next, and a gradual alpha-blend between the overlapping
portions of the frames. If the frames do not overlap at all, the current frame rotates off
of the screen, a spherical grid (as viewed from the center of the sphere) continues to
show the degree and direction of orientation, and finally the next frame rotates onto the
screen. The net effect is a high-frame-rate interpolation of the camera’s motion between
the frames. These transitions intuitively convey the relative positions of the frames, and
no users in our user study reported anything more than occasional temporary confusion
when watching long sequences of these transitions. Due to the visual nature of this
2
work, we encourage the reader to view some short video clips of transitions downloadable
from the web (http://ubivideos.homeip.net).
No knowledge of the camera’s position is required, unlike the previous work involving
inter-camera transitions [1]. The assumption is that the amount of positional
change in the interval between two frames is not significant, and the results of our user
studies confirms this. Even without the explicit representation of position, however, the
viewers still have a sense of movement through the environment. Not only is there the
illusion of movement similar to the illusion experienced when watching any sequence
of frames, but there is real movement as well. The manner in which we align the subsequent
frames when there is frame overlap, and the transition between the frames, creates
the sensation of movement. At times the alignment will cause the entering frame to start
off smaller than it really is, and then grow in size (zoom in) until it fills the screen. This
zooming creates the appropriate sensation of moving forward (or conversely, backward)
through the environment.
We explore the features of this approach in part with a field study of a hazardous
materials (hazmat) supervisor remotely monitoring a live video feed – transmitted over
a “broadband” cellular network – of two hazmat workers disposing of hazardous chemicals.
The camera was mounted on the mask of one of his team members. Such a system
configuration is motivated by a response in a damaged and chaotic environment. The
supervisor’s impressions of our visualization techniquewere surprisingly favorable, and
he dismissed the alternative encodings that were available. The unmodified low-frame
rate video left him feeling disoriented, and the low-quality 5fps (frames-per-second)
video was so choppy and disorienting that it interfered with his thinking and made him
nauseated.
We explore the finer distinctions among the various approaches to low-frame-rate
video with a laboratory study in which 14 subjects were asked to view video clips
of three different scenes that were encoded in four different ways. A surprising result
of this study is that four of the subjects actually preferred watching our 1fps transitionenhanced
video over full-frame (12fps), high quality video.Nearly all of the participants
preferred our visualization to the 5fps video clip that was encoded at a comparable
bitrate. One further interesting result is that nearly all of the participants were unable to
discern the difference between a clip that performed a simple alignment and blending
between frames, and one that also performed a morph between the frames to produce
more seamless transitions. This result can be explained by the brain’s ability to commit
closure with minimal cognitive load when modest amounts of visual information are
missing [2].
The remainder of this paper is organized as follows: In section 2 we motivate the
use of video in a disaster response setting, and describe the constraints that such an
environment places on technical solutions. In section 3 we describe our solution in
detail, and section 4 discusses related work. We present our field and lab studies in
sections 5 and 6, and then conclude.
2 Motivation
There are many situations in which high-panning low-bit-rate video can have value.
Consider, for example, CNN coverage of hurricanes or remote war-torn areas where
CNN resorts to satellite-phone video segments. These feeds are tolerable for the talkinghead
shots, but panning of the surrounding environment to show viewers what is hap-
3
pening results in a dissatisfying choppy, grainy image. There are also man-on-the-street
news reporting scenarioswhere it might be desirable to look at low-bitrate video. Breaking
news such as an accident, prior to the arrival of traditional television cameras, could
be viewed through citizen cameras with feeds transmitted over cellular networks, or
overlooked news could be streamed direct to the internet by citizen mobile phones.
On a more personal level, our user study revealed that people may be interested in
viewing live first-person streams of their distant loved-ones.Grandparents, for example,
may want to join the grand kids on a trip to the zoo.
2.1 HazardousMaterials Disposal
The use of video during the early stages of a disaster response, or even during the
late stages of a chemical clean-up are scenarios that can be impacted today. This paper
focuses on this latter scenario, and we have used the requirements of a hazardous
materials (hazmat) supervisor as the requirements for our video streaming solution.We
interviewed the Hazardous Materials Business Plan Manager (hereafter referred to as
Tod) at the University of California, San Diego (UCSD) to determine how he thinks a
live video feed could be used in managing a hazmat scene.
Although Tod’s team is trained and ready to handle disaster scenarios, a typical incident
is thankfully fairly mundane. On a university campus, the most typical emergency
condition is a fire in a laboratory that contains toxic chemicals. After the fire has been
put out and the situation has been stabilized there is often a substantial cleanup effort
that can take anywhere from days to weeks – all of which must be performed in full
suits with masks and respirators. Tod’s primary concern during one of these responses
is ensuring the health and safety of his team. As a commander, it is his job to know
what is going on, to interface with the various entities on scene (such as fire fighters,
witnesses, and lab managers), and to supervise the stabilization and cleanup of the environment.
Currently, Tod does not operate with any visuals of the scene, and because
of this, he often rushes to an incident so that he can be the first person to enter the environment.
This way he can use his memory of the conditions to help him make future
judgments.
As one of themost experiencedmembers on his team, Tod asserted that he could use
the information he receives from video feeds to help identify hazards, to coach the less
experienced members who are on the inside, and most importantly to see that his team
is active and healthy. Fatigue and heat exhaustion is a serious problem in this line of
work, and the hero mentality that is common among first responders often causes these
symptoms to go unreported. Tod said he would go so far as to make his team members
hold their air gauges to the camera since he does not necessarily trust their self-reports –
not because the team members are untrustworthy, but because the symptoms of fatigue,
the conditions of the environment, and the cumbersome suits that are being worn could
cause errors in the reading of the displays.
Our first reaction to the requirement for health and safety readings was that sensors
that report such things as heart rate, body temperature, etc. could transmit this data at
a much lower bandwidth. Tod, however, was eager to invest in video because the same
information could be conveyed by observing the body posture and the activity of his
team, as well as support the other functions cited above. He said that feeling like he is
actually there, in the environment, is important to him.
There are other benefits to having video. For example, what should be a simple
task, like finding a shut-off valve, becomes difficult when the people who know the
4
environment (lab managers, for example) cannot see what the people on the inside are
seeing. Not only would the lab manager be operating from memory, but the memory is
likely outdated since the conditions in the lab may now be very different.
The preference for head-mounted cameras over pan-tilt-zoom tripod-based cameras
is motivated by shortcomings of fixed cameras. There is no reason why fixed cameras
cannot also be supported, but the dynamic nature of a hazmat response suggests the
need for mobile cameras. It would be very difficult to position fixed cameras, even
with pan, tilt, and zoom capabilities in a place that would give enough coverage of the
environment while still providing detail of the activities of the team members.Multiple
cameras would be required if there are multiple work environments with obstructions
(e.g., walls, furniture, or debris) between them. A head-mounted camera has the benefit
that it is almost always focused on something that is of importance to at least someone
(the camera operator).
2.2 Networking Challenges
Operating in real-world environments is always challenging, but we are continually
been surprised by just how challenging the conditions are for wireless networking. In
prior work, we attempted to stream video during county-sponsored disaster drills across
an 802.11b wireless mesh network [3]. We were confident of success, given that we
were using a network that was carried on scene, self-configures, and is battery powered.
Two drills later, in which thousands of people were involved, with helicopters,
fire trucks, and the media interfering with the network in numerous ways, we are still
learning how to deal with the realities of wireless networking in disaster response. And
being just drills, the conditions of a real disaster scene would be different still. The
following list of network challenges is derived from our experience operating in these
environments.
Weak infrastructure support. One cannot rely on an existing network to exist, with
the possible exception of cellular networks, which have cell towers far removed from
the incident. In large scale disasters, though, cellular networks have been overloaded,
rendering them them useless for extended periods [4].
Unreliable networking. For networks brought on site, expect frequent network congestion
and failures, causing device disconnects. Interference is caused by both natural phenomena
and competition with other networks deployed in the same space (pre-existing
or imported for the response).
Low bitrate. Even on an 802.11b wireless network which has an effective single-source
bitrate of 6.2Mbps,with three cameras on a wireless mesh, we have not been able to rely
on much more than 100kbps per stream. Noise in a real-world environment contributes
to this loss of throughput, as does the fact that the total throughput drops drastically
as more nodes are added to the system. Empirical studies have shown that with more
than eight nodes, total throughput decreases to roughly 2Mbps [5]. Also, since each
camera is mobile, the perceived signal strength of the local access point (and thus the
bandwidth of the connection) will vary depending on the location of the camera and
intervening obstructions.
5
2.3 Video Compression Challenges
The conditions that have been outlined so far present a significant challenge for video
compression. Let us firstmotivate the need for compression.UncompressedCIF (352x288
pixels in RGB24 format) video playing at a typical 30fps requires a 69Mbps pipe.With
spatial compression, each frame can be reasonably compressed from 297KB to a 12KB
JPEG image. This compression technique, called MJPEG (motion JPEG) reduces the
bandwidth requirements to 2.8Mbps. Temporal (inter-frame) compression like that provided
by MPEG makes it possible to reduce the bitrate to the 300kbps range without
sacrificing much in terms of quality.
At the most basic level of detail, an MPEG-style encoder works roughly as follows:
A group of pictures (GOP) begins with an I-frame (which can be thought of as a JPEG
encoded image), and is followed by multiple P-frames (a predicted frame which encodes
the difference between the current frame and the previous frame). A P-frame is
more than just the simple difference between frames, though – the motion of objects
between the frames is taken into account and is encoded as motion vectors. We will
ignore B-frames since they are not important for this discussion. The length of a GOP
is usually specified as a parameter to the codec, but may also be determined by scene
changes. None of the P-frames are useful if any of the previous frames are lost, so Iframes
are important for error recovery in streaming scenarios, or to facilitate random
access playback for locally stored media.
If there is significant redundancy between frames, the difference between P-frames
will be small, and a high level of compression will be possible. If, on the other hand,
there is a lot of rapid panning (a common characteristic of first-person video), the differences
between frames will be great, and the P-frames may offer no better compression
than the I-frames. With traditional video codecs, there is no way around this. Without
temporal redundancy, there is little chance of doing better than MJPEG.
There are three options for rate-limited video that does not have temporal redundancy
(assuming the use of traditional codecs). (1) The frame rate can be reduced, (2)
the encoding quality of each frame can be reduced, or (3) a combination of these two
alternatives can be used. Reducing the frame-rate makes the video choppy and jittery –
a condition that many users in our user study could not tolerate. Reducing the framerate
below 5fps changes the experience of video to a sequence of still images. As the
frame rate drops, it becomes difficult to track objects, and eventually it is even difficult
to orient yourself in the scene. The other option, reducing the image quality, has similar
problems. While the motion will be smooth, the blurriness of the image may make it
difficult to identify objects.
To get the bitrate within 100kbps (to support three streams on our mesh network),
we have to sacrifice both the frame rate and the image quality in order to have the feed
continue to look like video (in other words, stay above 5fps). We found, and our user
studies confirmed, that video at this quality is really not acceptable in most circumstances.
The choppiness and blurriness induce nausea and headaches.
The remaining choice, then, is to drop below 5 fps and optimize instead on image
quality. Referring back to our MJPEG calculation above, the average JPEG encoding
size of 12KB at 1fps translates into a bitrate of 96kbps. Dropping down to one frame
per 1.5 seconds comfortably keeps us under 100kbps and even gets us under the 64kbps
1xEVDO cellular network limit.
In an unreliable network, MPEG has an additional problem: packet-loss results in
a garbled image until the next I-frame is received. If temporal rendundancy is low,
6
anyway, it may be best to limit the effects of packet-loss to individual frames by using
an MJPEG-style compression scheme instead. This is our motivation for usingMJPEG.
3 Our Approach
To reduce the disorienting effects of low-frame-rate video, our concept is to perform a
dynamic visual interpolation between frames using meta data captured from a digital
pan/tilt compass or inferred using vision techniques. In particular, we align the frames
in a spatially consistent way in a 3d graphics environment, and then use rotational and
translational motion to segue between the frames, producing a high-frame-rate experience
that captures the effects of camera motion. Because precise frame stitching is
impossible in real-time using 2D data, we use a dynamic crossover alpha-blend to help
the viewer correlate the information in the overlapping parts of the frames.
An imperfect alignment between two frames, due to, say, inaccurate sensor readings
is less of an issue than might be expected. Closure is a property of the human visual
system that describes the brain’s ability to fill in gaps when given incomplete information
[2]. It is a constant in our lives; closure, for example, conceals from us the blind
spots that are present in all of our eyes. So while there is ghosting, and maybe even
significant misregistration between frames, the human brain easily resolves these ambiguities.
An interesting result of our lab study is that almost no users were able to discern
the difference between segues that involved roughly aligned frames and those that were
more accurately aligned. In fact, of the few users that could discern a difference, some
of them actually preferred the rough registrations. Closure is that powerful.
The rest of this sections describes the details of our approach.
Creating a Panoramic Effect. Our approach can be described as the creation of a dynamically
changing and continually resetting spherical panorama. Each incoming frame
is positioned on the panorama, and projected onto a plane that is tangential to the sphere
to avoid distortion. A dynamic transition then moves the user’s viewpoint from the current
position within the panorama to the incoming frame’s position (Fig. 1, bottom).
The user’s viewport has the same field of view as the source camera, so the frame fills
the entire window once the transition is complete.Movement between frames looks like
smooth camera panning. Theremay also be a translational (shifting)motion effect if the
camera moves forward or backward through the scene.
A new panorama is started when consecutive frames do not overlap (Fig. 1, top).
The frames are positioned at their relative locations on the sphere, with an appropriate
gap between them. To help the user stay oriented, a wireframe of the sphere that serves
as the projective surface is displayed. Horizontal and vertical rotations are thus easily
recognized. The grid wireframe could be further enhanced by including markers for the
equator and the cardinal directions.
The planar simplification of 3d space only works for a short interval when cameras
are mobile. For this reason, at most five frames are placed in a given panorama. The
oldest frame is discardedwhen this limit is reached. This is not a significant compromise
because the source and target frames of a transition mostly fill the viewport, and any
other frames in the panorama are filling in around these two.
Frame placement in the panorama is managed through a robust two-level scheme,
as described in the rest of this section.
7
Image-based Frame Placement. When inter-frame rotations are not too large, we use
an implementation of Lowe’s SIFT algorithm [6] to find matching points between a
new frame and the previous frame, and then do a best-fit alignment of the frames to fit
the new frame into the panorama. The point-matching is performed on the camera units
in real-time, and the list of matched points between the current frame and the previous
frame are transmitted with each frame. In order to perform the matches in real-time on
our camera devices, the frame is downsampled to a quarter resolution (QCIF instead of
CIF) prior to analysis by SIFT. The result is good even at this lower resolution.
For each incoming frame, our rendering engine looks at the list of matching points
and determines if there is a match with the previous frame. To help remove erroneous
matches that made it through the RANSAC filter [6], we further filter the data based
on the expected usage pattern of the camera. For example, since the camera is mounted
on a person’s head, we may be able to assume that if matching points correspond to a
side-to-side tilt of the camera by more than 45 the matching points are erroneous.
The new frame is aligned to the previous frame by determining an affine correspondence
between the frames. We look at the relative position, orientation, and zooming
based on the two control points in each frame that are furthest apart. After aligning the
new frame, the frame is warped so that the matching points are exactly aligned. Surrounding
points are warped by an amount proportional to the inverse of the distance to
the neighboring control points. A transition to this new frame thus involves a morph
as well as the standard rotation and alpha-blend. At the end of the transition, the new
frame will be unwarped, and all of the other frames will be rotated and warped to match
the control points in the new frame.
Even with point matching, the alignment is not fully precise. Our planar simplification
of 3d space makes objects in the scene that are at depths different to those of
the points that have been matched be less accurately aligned. Even if the depths of the
matching points were recovered, the number of matching points (10-20) is very small
relative to the number of objects and object depths in the scene, so any recovered geometry
would be coarse. Also, since we are operating in real environments, dynamic
objects that move between frames will not have any point correspondences, and thus
will not be accurately aligned. Nonetheless, with the help of closure this technique can
produce very pleasing results.
Sensor-based Frame Placement. When SIFT fails to produce matching points for a
new frame, the frame’s placement depends on sensor data gathered from the camera
rig. The camera units we use are integrated with tilt sensors and electronic compasses
that record the tilt, roll, and yaw of the cameras at 15hz. This information allows us to
position the frames on the sphere. However, the sensor accuracy is not good enough for
generating a multi-frame panorama. Thus, the placement of such a frame initiates a new
panorama with the single frame. The rotational part of the transition is still performed
with the dynamic alpha-blend, using the previous frame’s and new frame’s relative sensor
data. However, since we do not have information about the relative or absolute
locations of the frames, we are unable to determine the relative translational positioning
between frames. The resulting experience mitigates the confusion caused by low
frame-rate video, but often lacks the aesthetics of the panorama and higher precision
placement.
8
4 Related Work
We are not aware of any related work that directly addresses the conditions we have
set out to handle in this paper, but there is some work that handles subsets of these
problems.
RealityFlythrough, which provides ubiquitous video support for multiple mobile
cameras in an environment, uses visualization techniques similar to the one we propose
in this paper, but for inter-camera transitions [1]. It requires knowledge of the positions
of the cameras, as well as the orientations, effectively limiting its use to environments
where ubiquitous location sensors are available, such as outdoors.
Irani, et al. directly address the problem of encoding panning video [7]. They construct
a photo mosaic of the scene, and are then able to efficiently encode new frames
by using the difference between the frame and the mosaic. Using this technique, it no
longer matters if consecutive frames have much overlap, because the assumption is that
similar frames have overlapped enough in the past to construct the mosaic. Irani reports
significant compression improvements over MPEG which was the standard in 1996,
and even using today’s standards the quality achieved at 32kbps is impressive. Unfortunately,
mosaic-based compression cannot be used for encoding mobile first-person
video because the cameras are mobile. Mosaic-based compression works well as long
as the camera remains relatively static and pans back and forth over the same scene, but
if the camera moves through the scene, there will be little opportunity to find matches
with previous images. Essentially mosaic-based compression extends the search window
for similar frames. If there are only a few similar frames, it does not matter how
big the search window is, as there will rarely be a match.
There are many examples of codecs that are designed to work in wireless, lowbit-
rate environments, although these codecs generally rely on the significant temporal
compression that is possible in “talking-head” video. H.264 [8] (also known has
MPEG4-10) represents the current state-of-the-art. MPEG4-2 [9] (commonly referred
to simply as MPEG4), the previous state-of-the-art, is more established and is more
likely to be supported in media players. There is little perceptible difference between
codecs that support these standards when compressing first-person video at low bitrates.
This is not surprising considering the absence of opportunities for temporal compression.
A non-traditional approach to video compression proposed by Komogortsev, varies
the quality of the video based on where the viewer is looking [10]. By using eye-gazetrackers
on the viewer, and predictingwhere the viewer will look next, the overall image
quality can be low, but the perceived quality would be high. This approach would be
difficult to implement in the environments we support because the network latency is
high (4-5 seconds); the gaze direction would have to predicted far in advance.
There has been substantialwork on generating panoramas fromstill photographs [11,
12]. Real-time dynamic creation of panaoramas on a handheld camera device has been
used to help with the creation of a static panorama [13]. Panoramas can also be efficiently
created frommovie cameras assuming the camera’s position is relatively static [14].
All of these techniques require some way to match points between images. We rely
heavily on Lowe’s SIFT algorithm [6], specifically the Autopano implementation of it
(http://http://autopano.kolor.com/).
9
5 Hazmat Field Study
We had several goals for our field study. First, we wanted to know if our visualization
technique was suitable for a hazmat command center. Second, we wanted to see if our
system could work in a realistic environment for an extended period of time. And third,
we wanted to discover the motion model of a head-mounted camera afixed to someone
doing a real job, oblivious to the presence of the camera.
5.1 Experimental Setup
The Scene. Every week, two members of the the UCSD hazmat team perform a maintenance
task that that doubles as an training exercise for response to an accident. All of
the hazardous waste that has been collected from labs around the university is sorted,
and combined into large drums in a process that is called bulking of solvents. This task
serves as an exercise, as well, because full hazmat gear must be worn during the procedure,
giving the team members (we will call them bulkers) experience putting on,
wearing, and performing labor-intensive tasks in gear that they will use at an incident
site. The bulkers also learn how to handle hazardous materials and obtain first-hand experience
with the properties of the chemicals with which they are dealing. For example,
it is not uncommon for labs to mislabel their materials, which can result in a dangerous
chemical reaction when the materials are combined in the drums.
The bulking process typically takes one to two hours depending on the amount of
waste material (roughly 250 gallons on average). During this time the bulkers are isolated
in a closed room because they are the only ones wearing equipment to protect
them from the noxious fumes. Tod, the team leader introduced in section 2, often worries
about the health of his team during these exercises, and looks forward to using a
video system similar to the one tested so that he can check on the bulkers periodically.
When we suggested that a permanent, wired camera might be more appropriate for this
particular situation, he re-emphasized how important preparation and training were in
his field. He wants his team to be training in the actual equipment that will be worn
during emergencies. They need to feel comfortable using and wearing it, and Tod needs
to have enough experience with it to trust it.
The Equipment. It was important to make the camera system as wearable and unobtrusive
as possible, given our desire to discover the real motion models of the camera. It
was also clear after our first interview with Tod that the bulkers were not going to tolerate
any setup that would impede their work. This is a dirty, tiring job, and they were
going to have little patience for anything that made them stay suited up for longer than
normal.
We attached a disassembled Logitech webcam ($100) to the front of the mask,
and sewed a tilt sensor manufactured by AOSI ($600) into the netting of the mask
that rested on the top of the head. These devices connected to a Sony Vaio U71P handtop
computer ($2000) which was placed in a small backpack. Consistent with Tod’s
dictum that his bulkers work with the same gear as in incident response, we chose to
transmit the video across the Verizon 1xEVDO network, which might be the only readily
available network if, say, a burnt out lab were being cleaned up. The Vaio lacked
the PC Card slot we needed for a 1xEVDO modem, however, so we connected via
802.11 to one of our wireless mesh network nodes, and had the traffic routed through
10
1xEVDO from there. The video feed was transmitted to our server, a standard VAIO
laptop (FS-790P $1600) connected via 802.11 to the campus network.
The 1xEVDO upstream bitrate was measured at between 60 and 79Kbps, and the
campus downstream bitrate at 3.71Mbps. We fixed the frame rate of the video feed to
.5fps to ensure that we would stay within the range of the 1xEVDO upstream speed.
The Task. Tod’s task was to use the video that was being transmitted by one of his
bulkers to explain to us the bulking process. This think-aloud interaction is realistic in
that Tod needs to train others in how to conduct his task for times when he is on vacation
or out sick. For us, this interaction served several purposes: (1) It would give Tod a
reason to be viewing the video, (2) it would encourage him to verbalize his impressions
of the system, and, (3) it would allow us to observe the effectiveness of the video stream
as a communicative device. Did the video provide enough detail to help illustrate what
he was describing, and at a fundamental level, did he understand what was going on?
As an expert, Tod’s subjective opinion of the system was important to us. The requirements
are domain specific, and only someone who has experience operating in a
command center can know if the quality of the video is appropriate for the task.
5.2 Results
Our camera system was worn by one bulker for the entire exercise which lasted for
roughly 64 minutes. We detected what looked like severe congestion on the 1xEVDO
link at the 50 minute mark. All of the results reported in this section discount the first 6
minutes of data (setup time), and the last 14 minutes of congested data.
The bulking task turned out to be very demanding for our visualization system.
Bulking is not only labor intensive with constant activity and motion, but is also conducted
in close quarters (a roughly 3-5 foot zone), leaving little opportunity for the
viewer of the video to get the perspective that comes with a wider field of view. The
camera motion was also unlike anything we had seen before in artificial drills. The
bulker was constantly bending down to his left to pick up a drum (weighing up to
27KG), and then placing it in the sink. This caused the view to move back and forth
between what we will call the origin (0 longitude, 0 latitude) and -90 longitude,
-90 latitude. These are quite extreme movements given the limited field of view of
our camera (44 long, 33 lat). The bulker indicated that he barely noticed our equipment,
so we judge that these extreme movements are representative for this task. It is
likely that similar motion models would be found in the cleanup phase of a burned-out
laboratory, where the task resembles a demolition effort.
Tod reacted favorably to the visualization. The lack of reaction is probably most
telling, considering the novelty of the visualization for him. He paused for a second
as he absorbed what he was seeing, and then began: “Ok, so this is following Sam as
he’s moving around the room. And as you can see they have a lot of work ahead of
them.” Tod then began describing the bulking process, at first just giving background
information that did not require access to the visuals. After this brief interlude, I asked
him if he could tell what was going on. “Yeah, I can tell that Sam is doing the bulking...
This gives me a good look at the funnel area so that we would see reactions if there was
a chemical reaction. Normally that’s displayed through vaporization. If you’re lucky
here (pointing to the screen) you might get a little bit of that. This tells me a little
bit about the viscosity. I can see the liquids, whether they’re plugging up. Sometimes
11
you get some chunks in there. And the thicker stuff – gels – looks like we had a little
bit in there... This definitely lets me know that they’re still working. That’s really the
important part.”
Tod expressed an interest in flipping through the still photographs so that we could
really study individual pictures.We showed him how he could pause the feed and move
back and forth through the images while still getting the benefit of the visualization.
The visualizations helped us stay oriented as he was describing the process, and saved
him from having to explain the relative positions of the images. Tod was also intrigued
by the time-stamping and indexing of all the images, as reconstructing timelines of an
event for post mortems is currently difficult because time pressures and distractions
undermine recordkeeping.
Tod then noticed the birdseye view that uses arrows and cones on a black backround
to represent the orientations of the camera views that are currently active. This helped
him orient, and he started using this view to illustrate where things were and what the
other bulker might be doing.
We then discussed how our visualization compared to the the normal view of low
frame-rate data which at these speeds looked more like a sequence of still photos. “Literally
for me, at the moment I would just go full screen on this particular moving one
(our visualization). I like this here (pointing to the birdseye view). This is telling me
the orientation in the room. I like these two. I’m not really even paying attention to
this one (the low frame-rate stream). The individual photos clicking through. I could be
disoriented with that one... It would tell me that they’re moving around, but after that
it’s not giving me anything that I really need for decisions.”
Throughout the exercise, Tod weighed in on why our visualization tool would be
effective in a control center environment. “...to have a visualization that adds credibility
to your discussion, ‘this is what we were doing, and this is the size of the equipment’,
and evaluating what resources are going to be needed for subsequent entries, and hopefully
we can get this from that single entry. You can’t get past the benefits of the visual.
It brings us to a whole other level of safety, assurance, and competence.”
He went on to explain how decisions are made by gut feelings, based on the skills
of the people that are involved and on his comfort level with those people. It is all
about contact, he said, and anything that increases contact is going to improve these
decisions. Contact is especially important when the lives of people you deployed are at
stake. “Video gives you a better gut feeling to what is going on. Video gives you another
form of contact. It builds trust in your decision making.”
Tod concluded with his assessment of the system: “Let me tell you what I like about
it. It’s not overwhelming. It’s appropriate. It’s not a huge distraction. That’s one of the
things you have to be concerned about – the level of distraction.... Yeah, I think you got
it. It really is the combination of the fact – it’s another piece. It’s not the all-empowering
‘this is the tool’, you know.You don’twant that, because if it didn’twork you don’twant
to all of a sudden – ’oh we can’t do anything because it’s not working’ You don’t want
that. What you want is good components that can go in and help add, and help make
better decisions... It’s appropriate. It’s not overwhelming. It doesn’t seem to be large,
cumbersome, overly difficult.”
Tod also had recommendations for improvement: he would like to have multiple
cameras so that he could see the scene from multiple angles, he requested wider-angle
lenses, and he wondered if he could set up fixed cameras as well: “I can see where I
would put in some more wide angles. I assume I could take one of these cameras and
12
just set it [in the environment]... Most of the events are quick and dirty. You wouldn’t
go with a stationary camera unless it was a prolonged cleanup.”
5.3 Followup
During the study we were of course unable to show Tod other possible encodings of
the data. Thus, we returned a few days after the experiment and presented him with
a re-creation of the experiment with 5fps encoded at bitrates comparable to the original
experiment, using FFMPEG’s MPEG4 codec, which was as good as the experimental
H264 codec described earlier (http://ffmpeg.sourceforge.net).His
reaction surprised us: “I don’t have a problem with the resolution on the right (the
5fps video), but it’s almost flipping through so fast that you’re not orienting yourself to
what’s going on... Yeah, I like the slower frame rate. It’s not so much because of the
resolution, it’s the amount of time that it takes me to know what I’m looking at... [The
5fps video] is snapping too fast – it’s too busy – it interferes with my thinking, literally,
it’s messing with my head.”
Even after showing himthe high quality 6.67fps feed that had been captured directly
at the camera, Tod still thought our abstraction was more appropriate for a command
center considering everything else that is going on. A command center needs to maintain
a sense of calm [15]. “This is just one piece of information that you’re going to
be getting. The phone is going to be ringing, people are going to be giving you status
reports. The [higher frame-rate video] is just too busy.”
We hypothesize that this intermediate frame rate overtaxed Tod’s closure capabilities.
At 30fps, the motion between frames is small enough for the result to appear
smooth. At 0.5fps, with high-frame-rate segues, there is both smoothness and ample
time to dwell on each target frame. At 5fps, there is little time to take in any individual
frame, and there is too much happening between frames, too quickly.
6 Lab Study
Intrigued by Tod’s observations during the field study that our visualizationmethodmay
actually be more pleasurable to watch than high fidelity first-person video, we increased
the scope of our planned lab study. We were now curious if our visualization method
would have broader appeal. Might it actually be an alternative to the sometimes nauseating,
“Blair Witch Project” [16] quality of first-person video? Would people choose
to watch first-person live video feeds of distant loved ones if given the opportunity?
Would Grandma want to virtually join the grandkids on a trip to the zoo?
We were interested in uncovering people’s subjective reaction to different encodings
of first-person-video. Very simply, Do you like it or not? This means that we had to
somehow divorce the content and any perceived task from the judgments. Obviously,
if the goal of watching the video was to read the text of a poster on a distant wall, for
example, then image clarity would be the most important quality. Likewise, if the goal
was to detect whether or not a big red ball bounced through the scene, the frame rate
would be most important. Our task, then, was to impress upon the subjects that it was
the quality of the video that they were judging, and assure them that it was okay for the
judgment to be purely subjective and even instinctive. The scenarios that were viewed
and the questions that were asked were designed to achieve this.
13
6.1 Experiment Setup
We recorded three 2-3 minute first-person video segments using a camera setup similar
to the one described in the previous section, but with a baseball cap replacing the
hazmat mask. Groceries was a video of a trip through the grocery store (representing a
crowded environment), Breakfast was video of someone making breakfast for the kids
(representing an indoor home environment), and Garbage was video of someone taking
out the garbage (representing an outdoor scene). Each of these videos was designed to
record a task to make the camera motion and the activities as natural as possible.
The three videos were then encoded in four different ways. encFast (eF) was sampled
at 1fps and run through our visualization system. encSlow (eS) was similar, but
sampled at .67fps. encIdeal (eI) was the “ideal” version, encoded at roughly 11fps (the
fastest our camera system could record raw video frames) with an infinite bitrate budget.
And encChoppy (eC) was encoded at 5fps at a comparable bitrate to the corresponding
encFast.
Starting with Groceries, the subjects were asked to begin watching each of the encodings
sequentially, but were then encouraged to watch them all in parallel so that they
could do side-by-side comparisons. The subjects were allowed to resize the video windows,
and could pause, rewind, and fast forward through the clips as desired. Breakfast
was viewed next, followed by Garbage. The following questions were given to the subjects
prior to the start of the experiment, and answers were solicited throughout. The
subjects were encouraged to alter their answers if subsequent clips revealed something
new.
– What is your gut reaction? Rank the video feeds in order of preference.
– Describe the characteristics of each of the video clips. Why do you like it? Why
don’t you like it?
– If it was your job to watch one of these clips all day long, and there was no specific
task involved, which would you choose?Why?
– Would you enjoy watching any of these clips (assuming interesting/relevant content)?
For example, to see kids, grandkids, friends, etc.
– Do any of these clips cause you physical discomfort?Which ones?
– Do any of the clips create confusion? If so, is it temporary or perpetual?
– Discounting the content, how do each of the clips make you feel?
– Have your preferences changed? If so, what is the new ranking?
These questionswere designed primarily to encourage the subjects to think critically
about each of the clips. Obtaining a carefully considered ranking of the clips was our
main goal. However, we also wanted to analyze the responses to help shed light on their
underlying reasoning.
6.2 Hypotheses
We expected the encodings to be ranked in the following order of preference: (1) eI, (2)
eF, (3) eS, and (4) eC, but thought some may prefer eF or eS over eI. eS was a lastminute
addition to the study after one of the authors felt a little queasy while watching
eF. We hypothesized that dropping the frame rate a little may make the difference for
some subjects prone to motion sickness. eS also encodes close to 1xEVDO network
bitrates.
14
6.3 Results
Subject Sex Age Game Exp Initial Pref Final Pref Nauseating
1 M 20 T eI, eF, eS, eC eI, eF, eS, eC eI, eC
2 F 60 F eI, eS, eF, eC eI, eS, eF, eC eC
3 M 40 F eI, eF, eS, eC eS, eF, eI, eC eI
4 F 40 F eF, eS, eI, eC eF, eS, eI, eC eI
5 F 60 F eI, eS, eF, eC eI, eS, eF, eC eF, eC
6 M 30 T eI, eC, eF, eS eI, eF, eC, eS eC
7 F 30 T eI, eC, eF, eS eI, eF, eS, eC eI, eC
8 M 30 F eI, eS, eC, eF eI, eS, eC, eF eI
9 M 20 F eS, eI, eF, eC eS, eI, eF, eC eI
10 M 30 T eI, eC, eF, eS eI, eC, eF, eS eF
11 F 30 F eS, eF, eI, eC eS, eF, eI, eC eC
12 M 30 T eI, eF, eC, eS eI, eF, eS, eC eI, eC
13 M 20 T eC, eF, eS, eI eC, eF, eS, eI eI
14 M 60 F eI, eF, eS, eC eI, eF, eS, eC eC
Table 1. Summary of results. Game Exp stands for 1st-person-shooter game experience. Inital
Pref is the gut reaction ranking given to each of the encodings, and Final Pref is the final ranking.
References to our encodings appear in bold.
The following summarizes the data found in Table 1. 14 subjects participated in this
study, 10 male, and 4 female, ranging in age from 20 to 60. All but two of the subjects
preferred at least one of our visualizations to the choppy encoding, and 4 of the subjects
actually preferred our visualizations to the ideal encoding that was used as a control. All
of the subjects reported that some of the video clips caused some physical discomfort
(nausea, mostly). eI and eC were the common culprits for this, but two individuals
had trouble with eF. 6 of the subjects preferred eS to eF, and in all of these cases the
preference was very strong. None of these 6 subjects had first-person-shooter game
experience. 4 of the subjects changed their ranking of the encodings midway through
the experiment, and in all cases our visualizations were ranked higher.
6.4 Analysis
We were surprised by how well our visualizations were received. Not only did 4 of
the subjects rank our visualizations higher than eI, there were also 4 others who were
explicitly on the fence, and saw definite benefits to the visualizations. Our visualizations
also seemed to grow on people. 4 of the subjects changed their rankings towards the
end of the experiment, moving our visualizations higher in preference. Everyone in the
study liked the visualizations, regardless of how they ranked them. The following is a
sampling of the positive qualities voiced by our subjects: calm, smooth, slow-motion,
sharp, artistic, soft, not-so-dizzy. There were of course some negative characterizations,
too: herkey-jerkey, artificial, makes me feel detached, insecure.
The clearest pattern was the subjects’ dislike of eC. We will discuss the two exceptions
to this a little later. Most stopped paying attention to eC early in the experiment
15
because the quality, to them, was obviously much poorer. This lack of consideration
may explain the occasional absence of eC but the presence of eI in the Nausea column
in Table 1.
Many of the subjects had a strong personal criterion that they used for judging
the videos. For some, it was clarity of the images and for others it was the lack of
choppiness. There were also those who were most influenced by nausea.
The clarity camp (subjects 2, 5, and 14) is interesting because it was not until the
end of the study that we realized what bothered them about our visualizations. Subject
5 kept reiterating that the characteristics she sought were “slow and clear”, and yet
she chose eI over eS. The image quality of eI and eS should have been identical, and
eS did not have the fast, jittery quality that the “slow” request was an obvious reaction
to. Subject 14’s similar responses solved the puzzle. Although the image clarity
of the individual images is high in our visualizations, the alpha-blend performed during
transitions causes a temporary blurriness since the alignment between images is not perfect.
Transitions that do not use an alpha-blend have a certain appeal, but we ultimately
chose to include the alpha-blend in the clips used for this study because, in our opinion,
the alpha-blend makes the transitions feel smoother and calmer, as well as assisting in
closure. It would be interesting to get the clarity camp’s reaction to non-alpha-blended
transitions. For those who used the lack of choppiness as their main criterion, the nonalpha-
blended transitions would probably be unfavorably received.
It was fortunate that we added eS to the study, because eS made all of the difference
for some. Subject 8 actually ranked eF the lowest because it was just too “herky-jerky”.
Subject 9 liked eS the best, but ranked eF below eI. eF “had a jolting, motion sickness
feel.” Others, on the other hand, had strong negative reactions to eS, because it was too
slow and boring. There appears to be a strong correlation between an individual’s lack of
first-person-shooter game experience and their preference for eS. None of the subjects
who preferred eS had any game experience. First-person-video is not something that
people get a lot of experience watching, unless they play first-person-shooter games.
We surmise that with more experience, people may actually prefer the speed of eF.
This study helped us understand why first-person-video can be so difficult to watch.
It mostly boils down to control and expectation. Obviously we all have experience
watching our own first-person-video every day of our lives.Why are we not bothered by
it? We are controlling where we look, and because we are controlling it, we anticipate
what the motion is going to feel like, and we have a pretty good idea of what to expect
when the motion stops.When watching something through another person’s eyes, however,
that expectation is lost, so we are always playing catch-up. Subject 4 preferred our
visualizations over eI precisely for this reason. She said that eI was moving so fast that
she could not pick up any of the details. Just as she was about to focus on the current
scene to comprehend it, the view moved to something else. She liked that eF gave her
the extra time to actually absorb what was going on.
Subjects 10 and 13 were the only ones who preferred eC over our visualizations.
Their reasons were quite different so we will consider them independently. Subject 10
simply preferred traditional video to the visualizations. He could see the value in the
visualizations, and was not confused by them, but he felt detached watching them.
Subject 13 is an interesting outlier. Not only did he rank eC the highest, but he
ranked eI the lowest! In a post-experiment interview we learned that he preferred the
artistic quality of eC. It was edgy. He was bored by eI and found it a little bit nauseating.
He also liked the artistic feel of our visualizations, but ultimately the “predator” feel of
16
eC is what drew him towards that one. Clearly, people are different – there is no way
we could ever create a solution that appeals to everyone
The different scenes did not appear to make any difference to the subjects’ preferences.
None of the scenes were responsible for a change in ranking. People seemed
to enjoy watching the Breakfast video the most because of the presence of the kids.
This video was probably responsible for all but one of the subjects indicating that they
could see themselves enjoying watching live first-person video of their loved ones. In
this context, some of the subjects who liked eI the best thought our visualizations would
be more appropriate. This can be attributed to the fact that many found our visualizations
easy to watch. Details are probably not very important in this context, so the lower
frame-rate would not be a factor.
6.5 Secondary Study
During this lab study we took the opportunity to investigate the subjective value of the
morphing performed when transitioning between frames that were aligned via point
matching, as described in section 3. It was not clear that morphing was providing much
benefit, and when the vision algorithm occasionally returned incorrect matching points,
the morph looked startlingly bad.
We had our subjects do side-by-side comparisons of a morphed and non-morphed
version of the Garbage video encoded as eF. They also made a similar comparison with
the Groceries video, although this time the transitions were slowed down by a factor
of 8. None of our subjects could discern any difference between the morphed and nonmorphed
versions of the Garbage video. After watching the Groceries video, most of
the subjects still barely noticed a difference, but many had a vague preference for one
over the other. These preferences are not surprising in light of the range of preferences
cited above: 4 preferred the non-morphed version because it was softer and rocked less,
and 4 the morphed version because it was sharper.
All of the subjects were able to see the differences once they were pointed out, and
stop-motion revealed that the alignment between the morphed images was much better.
So why is it that the subjects had such a difficult time seeing the differences themselves?
We hypothesize two explanations. (1) Our brains are so good at committing closure that
unless there is perfect alignment between images, varying degrees of misalignment (to a
point) are perceived as being the same. There are times when closure is being performed
consciously, but for the most part this is a process that happens unconsciously, and
people are only vaguely aware of it happening. (2) The interesting content of the scene is
the dynamic elements – the very content that does not getmorphed because no matching
points are found on them between frames.
7 Conclusion
We have presented a visualization technique for displaying low-bit-rate first-person
video that maintains the benefits of high resolution, while minimizing the problems
typically associated with low frame rates. The visualization is achieved by performing
a dynamic visual interpolation between frames using meta data captured from a digital
pan/tilt compass or inferred using vision techniques.We have demonstrated with a field
study that this technique is appropriate in a command center, in contrast with traditional
low-bitrate encodings which may cause disorientation and physical discomfort. Our lab
17
study showed that people may actually choose to watch such video for entertainment
since it has the unexpected benefit of eliminating the “Blaire Witch Project” [16] effect
– the nausea-inducing jumpiness typical of first-person video. Indeed, 4 out of 14
subjects in our study actually preferred this visualization to the high frame-rate, high
quality video that was used as a control.
8 Acknowledgments
Special thanks to Tod Ferguson and the UCSD Hazmat team. This work was supported
in part by contract N01-LM-3-3511 from the National Library of Medicine and a hardware
gift from Microsoft Research.
References
1. McCurdy, N.J., Griswold, W.G.: A systems architecture for ubiquitous video. In: Mobisys
2005: Proceedings of the Third International Conference on Mobile Systems, Applications,
and Services, Usenix (2005) 1–14
2. McCloud, S.: Understanding comics: The invisble art. Harper Collins Publishers, New York
(1993)
3. Arisoylu, M.: 802.11 wireless infrastructure to enhance medical response to disasters. In:
Proc. AMIA Fall Symp. (2005)
4. McKinney, Company: Post 9-11 Report of the Fire Department of New York. (August 2002)
5. Vasan, A., Shankar, A.U.: An empirical characterization of instantaneous throughput
in 802.11b wlans. http://www.cs.umd.edu/˜shankar/Papers/
802-11b-profile-1.pdf (2006)
6. Lowe, D.G.: Distinctive image features from scale-invariant keypoints. Int. J. Comput.
Vision 60(2) (2004) 91–110
7. Irani, M., Anandan, P., Bergen, J., Kumar, R., Hsu, S.: Efficient representations of video
sequences and their applications. Signal Processing : Image Communication (1996) 327–
351
8. Wiegand, T., Sullivan, G.J., Bjntegaard, G., Luthra, A.: Overview of the h.264/avc video
coding standard. IEEE Trans. Circuits Syst. Video Techn. 13(7) (2003) 560–576
9. : International Organisation for Standardisation: ISO/IEC JTC1/SC29/WG11MPEG
98/N2457. (1998)
10. Komogortsev, O., Khan, J.I.: Predictive perceptual compression for real time video communication.
In: ACM Multimedia. (2004) 220–227
11. Szeliski, R.: Image mosaicing for tele-reality applications. In: WACV94. (1994) 44–53
12. Brown, M., Lowe, D.G.: Recognising panoramas. In: ICCV ’03: Proceedings of the Ninth
IEEE International Conference on Computer Vision,Washington, DC, USA, IEEE Computer
Society (2003) 1218
13. Baudisch, P., Tan, D., Steedly, D., Rudolph, E., Uyttendaele, M., Pal, C., Szeliski, R.:
Panoramic viewfinder: providing a real-time preview to help users avoid flaws in panoramic
pictures. In: OZCHI ’05: Proceedings of the 19th conference of CHISIG of Australia on
CHI, Narrabundah, Australia, Australia, CHISIG of Australia (2005) 1–10
14. Steedly, D., Pal, C., Szeliski, R.: Efficiently registering video into panoramic mosaics. In:
ICCV ’05: Proceedings of the Tenth IEEE International Conference on Computer Vision
(ICCV’05) Volume 2, Washington, DC, USA, IEEE Computer Society (2005) 1300–1307
15. Weiser, M.: The computer for the 21st century. Human-computer interaction: toward the
year 2000 (1995) 933–940
16. Myrick, D., Sanchez, E.: Motion picture: Blair witch project. (1999)
18

Creating Your First Flash Website : Flash Web Design Tutorial

2009-05-09T04:47:00.000-07:00

This tutorial will teach you how to create a simple flash website that has three sections. We are not going to use any advanced ActionScript and will rely only on the GotoAndPlay method to control our movie clip. You can view an example of what we're going to create in the Flash movie below.
AC_FL_RunContent( 'codebase','http://download.macromedia.com/pub/shockwave/cabs/flash/swflash.cab#version=7,0,19,0','width','600','height','400','title','Simple Flash Website Demo','src','index','quality','high','pluginspage','http://www.macromedia.com/go/getflashplayer','movie','index' ); //end AC code

This tutorial uses ActionScript 1/2. If you would like to learn how to create a similar project using ActionScript 3.0 then check out our other Basic Flash Website tutorial.

Setting Up the Stage
Start off with a new Flash movie, access the Properties Inspector to set the dimensions of the movie to 600x400px, the background color to white, and the frame rate to 12 fps.
On the timeline, create four layers and name them Background, Content, Buttons, and Actions. Place them in the order illustrated in the image below.
Creating a Preloader
We are not going to discuss how this works as the topic was covered in a previous tutorial which you can check to learn more about preloaders. We'll use a very basic preloader that shows the word Loading while the movie loads. To do this, click once on the name of the layer labeled Content, and then right-click the second frame on that layer and select Insert Frame. Now use the Text Tool to write the word Loading on the scene. Once you have the word Loading written on the stage somewhere, you can change the font and the size of it through the Properties Inspector, choose whatever suits your taste. Make sure that the word is placed in the center of the scene. (Hint: Use the Align Panel Ctrl+K)

That should do the visual part of the preloader, back to our timeline, right-click the second frame in the layer labeled Actions and select this time Insert Keyframe. Now right-click the same frame once again and select Actions to open the Actions Panel. Paste the following code to complete your preloader:
if (_root.getBytesLoaded() != _root.getBytesTotal()){gotoAndPlay(1);}
You can learn more about Preloaders from this tutorial.
Adding the Content
Our website is going to have three sections (i.e. pages), we are going to put these different sections in three separate frames, those will be the third, fourth, and fifth frame on the timeline.
The first thing that we'll add to our scene is the background, we'll make this extend through all the three frames we're going to use for content. Click once on the name of that layer first, and then Right-click the third frame on the Background layer and select the Insert Keyframe command, we'll expand this keyframe all the way to the fifth frame by right-clicking the fifth frame and then selecting the Insert Frame command.

We'll now insert our background image, save this JPEG and your desktop, simply go through File>Import>Import to Stage and then select your image to import it onto the stage. You will have to place it in the middle of the stage.
Moving on to the actual content, click on the name of the layer labeled Content and then right-click the third frame on that layer and select the Insert Blank Keyframe command. This frame should contain the content of our first section. We are going to have an introduction text in here, select the Text Tool from the toolbar and write whatever text you want to have in the page. You can just write Welcome in a huge size.
We will now add the content for our second section, this was titled 'Tutorials' in the example above. Right-click the fourth frame on the same layer you've just edited, the Content layer, and then select the Insert Blank Keyframe command. Use the text tool to write different text in this page. I simply wrote, 'tutorial links coming soon!'.
We will repeat the last step for the fifth frame on the Content layer to add the content of our third section. Right-click the fifth frame on that layer and then select the Insert Blank Keyframe command. Use the text tool to write 'Click here to visit our Forum'. You can any part of this text a link by simply selecting the desired word and then accessing the Properties Inspector and then assigning a URL in the URL field.

Creating the Buttons
Our buttons will remain on the stage all through out the different sections the user views, for this reason we will place them on a frame that spans across the three frames that have our content. Click once on the name of the layer labeled Buttons and then Right-click the third frame on it to select the Insert Keyframe command. This keyframe shall span all the way through the fifth frame, so right-click the fifth frame and select the Insert Frame command.

It is now time to add the actual buttons, we will use the button component instead of creating our own. Access the Components Panel by going through Window>Components. You will find the button component under the User Interface category, simply drag and drop three copies of it onto the stage.
Select the button at the top and then access the Properties Inspector and click on the Parameters tag. You will have to change the label of the button to something descriptive, I used Home. You will then also have to assign an instance name to the button so that we can control this button via ActionScript. I assigned the name home_btn to it. (Notice: Instance names are case-sensitive)
Repeat the last step for the other two buttons, label the second one Tutorials and assign the instance name tutorials_btn to it. Label the third one Forum and assign the instance name forum_btn to it.
ActionScript
All the visual elements of our little website are now present, we need to write the command codes to tell those buttons to go to the specified section when clicked. Right-click the third frame in the layer labeled Actions and select the Insert Blank Keyframe command. Once you have a new keyframe in their, right-click that frame and then select the Action command to open up the Actions panel. We are going to tell our movie to stop at the home frame (frame number 3) and then tell the buttons to go to and stop, each at the specified frame for its content. You can do this by simply pasting the following code in the Actions panel:
stop();home_btn.onRelease = function () {gotoAndStop(3);}tutorials_btn.onRelease = function (){gotoAndStop(4);}forum_btn.onRelease = function (){gotoAndStop(5);}
I think that the code is self-explanatory, you have done it, test the movie to see our simply website functioning!
AC_FL_RunContent( 'codebase','http://download.macromedia.com/pub/shockwave/cabs/flash/swflash.cab#version=7,0,19,0','width','600','height','400','title','Simple Flash Website Demo','src','index','quality','high','pluginspage','http://www.macromedia.com/go/getflashplayer','movie','index' ); //end AC code

This concludes our tutorial, I hope that you learnt something helpful, please feel free to post at the Oman3D Forum if you have any questions regarding this tutorial. You can read the Tutorial for Creating a Portfolio Scene in Flashor the Tutorial For Creating a Dynamic Portfolio in Flash for anybody wishing to learn other relatively advanced techniques for flash navigation.
- End of Tutorial

Free SEO Tools : Free Tools For Search Engine Optimization

2009-04-18T01:12:00.000-07:00

Free SEO Tools by SEO Chat
The SEO Tools found in this section were designed to assist you in configuring your website(s) for search engine optimization within the different search engines. Please select one of the SEO Tools from below to begin optimizing your website(s) for top placement within the search engines. Our SEO tools do not support any secured server sites due to authentication issues.



Adsense Calculator Google AdSense affiliates can use this calculator to gain a better understanding of what affects their earnings by experimenting with values. AdSense Preview This preview utility will give you a sense of which ads would be placed on a given page. Advanced Meta Tag Generator This tool helps you to add meta tags to your site. Alexa Rank Comparison Tool The Alexa Traffic History Graph allows you to create a traffic history graph for any site. Check Server Headers Check your server to make sure the proper HTTP Status Codes (200, 301, 302, 304, 307, 404, 410) are being returned in the server headers. Class C Checker This Class C Checker tool allows you to find out whether some sites are hosted on the same Class C IP Range. Code to Text Ratio This tool will help you discover the percentage of text in a web page (as compared to the combined text and code). CPM Calculator This calculator measures the ROI (return on investment) of the CPM (cost per thousand) impressions advertising model. Domain Age Check Use this tool to find the age of a domain. Domain Typo Generator Enter a domain name into the box, and this tool will generate a list of suggestions of likely human misspellings and typos for the given domain. Future PageRank This tool will query Google's various data centers to check for any changes in PageRank values for a given URL. Google Dance This tool will query Google's three main web servers, showing the different statuses during updates. Watch Google Dance. Google Keyword Suggestions The keyword suggestion tool for Google will help you choose relevant and popular terms related to your selected key term. Google Search for Multiple Datacenter This tool searches your keyword/phrase in different Google data centers. Google Suggest Scraper Shows frequently search for phrases starting with the words and letters in your query. Google vs Yahoo This tool will run a search query in Google and Yahoo search engines and then graphically compare the results. Indexed Pages This tool will return the total link count for each URL from major search engines (Google, Yahoo, MSN, Alta Vista, and AlltheWeb). Keyword Cloud This tool provides a visual representation of keywords used on a website. Keyword Density This tool will analyze your chosen URL, and return a table of keyword density values for one-, two-, or three-word key terms. Keyword Difficulty Check Use the Keyword Difficulty Check Tool to see how difficult it would be to rank for specific keywords or keyword phrases. Keyword Optimizer Enter a list of keywords and this tool will remove any duplicate entries and re-order the list alphabetically.		Keyword Position Check for Multiple Datacenter This tool will help you to find the position of your site in Google for a specific keyword/phrase. Keyword Typo Generator Enter a keyword or key term into the box above and this tool will generate a list of suggestions for likely human misspellings and typos. Link Popularity Enter a valid URL, and this tool will query all the major search engines (Google, Yahoo, MSN, and Teoma) and then return that URL's total link count for each one. Link Price Calculator This tool can help you determine the approximate amount you should be paying per month for a text link. Meta Analyzer This tool will analyze a website's meta tags. Analyzing a competitor's keyword and description meta values is a good way to find ideas for key terms and more effective copy for your site. Meta Tag Generator If you're new to web development and search engine optimization, you may find this tool useful to ensure that your meta tags are correctly formed. Multiple Datacenter Link Popularity Check This tool will give a back link count for a URL from multiple Google data centers. Page Comparison Compares the page titles, meta information, and common phrases occurring on different pages. Page Size This tool will help you to determine HTML web page size. PageRank Lookup This tool streamlines the process of checking PageRank for your sites. Enter a list of URLs and it will return the PageRank value for each one. PageRank Search Enter your Google search here, and our tool will search Google and display the PageRank next to each resulting answer. Robots.txt Generator Use this tool to generate a simple robots.txt file for your website, which allows you to hide files or directories that you donâ€™t wish to be spidered. ROI Calculator This calculator measures the ROI (return on investment) of a CPC (cost per click) campaign. Search Engine Comparison This tool allows you to perform your own comparisons, and displays the results visually, making it easy to see both the rankings and comparative positions of pages in search engine results. Search Engine Keyword Position This tool checks the search engine result pages of Google, Yahoo, and MSN to see what position your site holds for a particular keyword phrase. Site Link Analyzer This tool will analyze a given web page and return a table of data containing columns of outbound links and their associated anchor text. Spider Simulator This tool simulates a search engine spider by displaying the contents of a web page in exactly the way the spider would see it. URL Redirect Check This tool checks for valid HTTP 301 headers. URL Rewriting This tool converts dynamic URLs to static URLs. You will need to create an .htaccess file to use this tool.

SEO Expert Tips : Cheap SEO Service in Kolkata

2009-03-08T12:00:00.000-07:00

Small businesses are growing more aware of the need to understand and implement at least the basics of search engine optimization. But if you read a variety of small businesses blogs and Web sites, you’ll quickly see that there’s a lot of uncertainty over what makes up “the basics.” Without access to high-level consulting and without a lot of experience knowing what SEO resources can be trusted, there’s also a lot of misinformation about SEO strategies and tactics.
This article is the second in a two-part SEO checklist specifically for small business owners and webmasters. Last week, I shared 20 “don’ts.” Naturally, this week addresses the “Do’s”—things to make sure you include whether you’re hiring an SEO company or doing it yourself.
Small Business SEO Checklist: The Do’s
1. Commit yourself to the process. SEO isn’t a one-time event. Search engine algorithms change regularly, so the tactics that worked last year may not work this year. SEO requires a long-term outlook and commitment.
2. Be patient. SEO isn’t about instant gratification. Results often take months to see, and this is especially true the smaller you are, and the newer you are to doing business online.
3. Ask a lot of questions when hiring an SEO company. It’s your job to know what kind of tactics the company uses. Ask for specifics. Ask if there are any risks involved. Then get online yourself and do your own research—about the company, about the tactics they discussed, and so forth.
4. Become a student of SEO. If you’re taking the do-it-yourself route, you’ll have to become a student of SEO and learn as much as you can. Luckily for you, there are plenty of great Web resources (like Search Engine Land) and several terrific books you can read. Aaron Wall’s SEO Book, Jennifer Laycock’s Small Business Guide to Search Engine Marketing, and Search Engine Optimization: An Hour a Day by Jennifer Grappone and Gradiva Couzin are three I’ve read and recommend.
5. Have web analytics in place at the start. You should have clearly defined goals for your SEO efforts, and you’ll need web analytics software in place so you can track what’s working and what’s not.
6. Build a great web site. I’m sure you want to show up on the first page of results. Ask yourself, “Is my site really one of the 10 best sites in the world on this topic?” Be honest. If it’s not, make it better.
7. Include a site map page. Spiders can’t index pages that can’t be crawled. A site map will help spiders find all the important pages on your site, and help the spider understand your site’s hierarchy. This is especially helpful if your site has a hard-to-crawl navigation menu. If your site is large, make several site map pages. Keep each one to less than 100 links. I tell clients 75 is the max to be safe.
8. Make SEO-friendly URLs. Use keywords in your URLs and file names, such as yourdomain.com/red-widgets.html. Don’t overdo it, though. A file with 3+ hyphens tends to look spammy and users may be hesitant to click on it. Related bonus tip: Use hyphens in URLs and file names, not underscores. Hyphens are treated as a “space,” while underscores are not.
9. Do keyword research at the start of the project. If you’re on a tight budget, use the free versions of Keyword Discovery or WordTracker, both of which also have more powerful paid versions. Ignore the numbers these tools show; what’s important is the relative volume of one keyword to another. Another good free tool is Google’s AdWords Keyword Tool, which doesn’t show exact numbers.
10. Open up a PPC account. Whether it’s Google’s AdWords or Yahoo’s Search Marketing or something else, this is a great way to get actual search volume for your keywords. Yes, it costs money, but if you have the budget it’s worth the investment. It’s also the solution if you didn’t like the “Be patient” suggestion above and are looking for instant visibility.
11. Use a unique and relevant title and meta description on every page. The page title is the single most important on-page SEO factor. It’s rare to rank highly for a primary term (2-3 words) without that term being part of the page title. The meta description tag won’t help you rank, but it will often appear as the text snippet below your listing, so it should include the relevant keyword(s) and be written so as to encourage searchers to click on your listing. Related bonus tip: You can ignore the Keywords meta altogether if you’d like; it’s close to inconsequential. If you use it, put misspellings in there, and any related keywords that don’t appear on the page.
12. Write for users first. Google, Yahoo, etc., have pretty powerful bots crawling the web, but to my knowledge these bots have never bought anything online, signed up for a newsletter, or picked up the phone to call about your services. Humans do those things, so write your page copy with humans in mind. Yes, you need keywords in the text, but don’t stuff each page like a Thanksgiving turkey. Keep it readable.
13. Create great, unique content. This is important for everyone, but it’s a particular challenge for online retailers. If you’re selling the same widget that 50 other retailers are selling, and everyone is using the boilerplate descriptions from the manufacturer, this is a great opportunity. Write your own product descriptions, using the keyword research you did earlier (see #9 above) to target actual words searchers use, and make product pages that blow the competition away. Plus, retailer or not, great content is a great way to get inbound links.
14. Use your keywords as anchor text when linking internally. Anchor text helps tells spiders what the linked-to page is about. Links that say “click here” do nothing for your search engine visibility.
15. Build links intelligently. Submit your site to quality, trusted directories such as Yahoo, DMOZ, Business.com, Aviva, and Best of the web. Seek links from authority sites in your industry. If local search matters to you (more on that coming up), seek links from trusted sites in your geographic area—the Chamber of Commerce, etc. Analyze the inbound links to your competitors to find links you can acquire, too.
16. Use press releases wisely. Developing a relationship with media covering your industry or your local region can be a great source of exposure, including getting links from trusted media web sites. Distributing releases online can be an effective link building tactic, and opens the door for exposure in news search sites. Related bonus tip: Only issue a release when you have something newsworthy to report. Don’t waste journalists’ time.
17. Start a blog and participate with other related blogs. Search engines, Google especially, love blogs for the fresh content and highly-structured data. Beyond that, there’s no better way to join the conversations that are already taking place about your industry and/or company. Reading and commenting on other blogs can also increase your exposure and help you acquire new links. Related bonus tip: Put your blog at yourdomain.com/blog so your main domain gets the benefit of any links to your blog posts. If that’s not possible, use blog.yourdomain.com.
18. Use social media marketing wisely. If your small business has a visual element, join the appropriate communities on Flickr and post high-quality photos there. If you’re a service-oriented business, use Yahoo Answers to position yourself as an expert in your industry. With any social media site you use, the first rule is don’t spam! Be an active, contributing member of the site. The idea is to interact with potential customers, not annoy them.
19. Take advantage of local search opportunities. Online research for offline buying is a growing trend. Optimize your site to catch local traffic by showing your address and local phone number prominently. Write a detailed Directions/Location page using neighborhoods and landmarks in the page text. Submit your site to the free local listings services that the major search engines offer. Make sure your site is listed in local/social directories such as CitySearch, Yelp, Local.com, etc., and encourage customers to leave reviews of your business on these sites, too.
20. Take advantage of the tools the search engines give you. Sign up for Google’s webmaster Central and Yahoo’s Site Explorer to learn more about how the search engines see your site, including how many inbound links they’re aware of.
21. Diversify your traffic sources. Google may bring you 70% of your traffic today, but what if the next big algorithm update hits you hard? What if your Google visibility goes away tomorrow? Newsletters and other subscriber-based content can help you hold on to traffic/customers no matter what the search engines do. In fact, many of the DOs on this list—creating great content, starting a blog, using social media and local search, etc.—will help you grow an audience of loyal prospects and customers that may help you survive the whims of search engines.
Just like last week, this list could continue well beyond these 21 “DOs.” Your additions are welcome in the comments.
With this checklist and last week’s list of “Don’ts,” you should be able to develop a good plan of attack for your SEO efforts for your small business.
Matt McGee is the SEO Manager for Marchex, Inc., a search and media company offering search marketing services through its TrafficLeader subsidiary. The Small Is Beautiful column appears on Thursdays at Search Engine Land.

How To Design A Good Website : Basic Tips

2009-02-12T13:25:00.000-08:00

Listed below are some useful and rather important tips for designing a professional and high quality web site:
Neat and Easy Navigation: Navigation of links on your site plays a big role in determining the stickiness of your site (how long your visitor stays and explores your site). Ask yourself this, What do visitors do as soon as they open your site? They would probably read the content of the present page and then look around to find any other page that interests them. Read our article on Web site Navigation Tips.
Clean Layout Design: A clean layout that uses a lot of white space enhances a site's looks. Try to keep the focus on your content, use a template for this. Use fonts that will be available on all computers to prevent your site looking messed up.
Optimum Load Time: Make sure your load time is low. For this you must: Minimize Graphics, Flash and scripts: They hugely increase your file size. Optimize your HTML & script code: Make sure that your site doesn't have any unwanted tags or unused scripts.Use Server Side Include (SSI) files where ever possible. SSI files once called from the web server reside in its cache so on subsequent requests they load faster. Read our article Tips for a Fast Loading Site for more.
Design for all Screen Resolutions: A site that is easy-to-use always encourages visitors to stay and read your content. For site with long pages of content this is very crucial as the amount of scrolling required is reduced. Suppose your site doesn't look good for a particular resolution it is very probable that the visitor will close the browser window feeling that the web page is not for their viewing. Designing stretch layouts that fit any screen resolution ensures that you know all your visitors see a visually appealing and professional site. Read our article Designing for all Screen Resolutions for more.
Ensure Web site scalability: Make sure your code and design is scalable. As technology advances and configuration of computers & their monitors keep increasing and varying it is impossible to test your site in all screen sizes and platforms.
Cross Browser Compatible: Make sure you check your site for Internet Explorer 5+, Mozilla Firefox 1.0, Opera 7.0 and Netscape Navigator 6+ as they constitute 95% of the worlds browsers.

For Best Price Web Design Visit http://techbongo.com
Phone: +91 9883246001
Email: techbongo@gmail.com

Huge Increase in Business 2009

2009-01-12T22:58:00.000-08:00

70.5% increase in Business. A record increase in profit than the opening year. TechBongo Solutions has made 70.5% more profit compared to last year.

Free Domain Name Resellers in Kolkata

2008-11-15T19:32:00.000-08:00

We're offering free domain name resellers account for our clients. This works in pay and buy policy. That is, you'll have a wallet in your account. You can store any amount in that wallet. Now, whenever you'll register a domain name for your client, you have to pay from that wallet. There is no minimum threshold amount, that must be kept in the wallet. This is how, we are claiming it to be a Free Domain Resellers offer.

In addition, you can also have your own reseller clients. This works as a chain. The clients join as domain reseller under you, will be your sub-ordinate in the chain. You'll have a free site from us. This free site can be branded for your company. Also, you can set up one of your existing site's sub-domain (or may be domain) accordingly, so that it opens that free resellers site.

For any kind of further assistance, contact techbongo@gmail.com or call +91 9883246001.
For domain pricing and domain name registration, visit http://domains.cheaphostindia.info/

WE ARE ICANN AFFILATED DOMAIN RESELLERS IN KOLKATA, INDIA. YOU'LL ALWAYS FEEL SAFE TO WORK WITH US.

Contact TechBongo Solutions

2008-10-03T15:15:00.000-07:00

Here is the contact information of TechBongo Solutions

General Email : techbongo@gmail.com
Support Emal : techbongo@gmail.com & support@techbongo.com

Mobile : +91 9883246001
Mobile : +91 9903029085

Website : http://techbongo.com
Website: http://cheaphostindia.info

Address :
B-12/134, Kalyani
Nadia, West Bengal
India - 741235