Sound adds a useful dimension to corporate intranets
August 30, 1999
by Cynthia Morgan
(IDG) -- The day is fast approaching when some executive will have a singularly bright idea: Why not fill in the blanks of his otherwise terse slide show with the recorded speech that went with it, and post the whole shebang on the corporate intranet? It would make a more informative presentation and save him from repeating the speech to every new employee.
He's right. Audiovisual experts say sound paired with still images or text can have nearly as much impact and information as video without the high costs of producing and transmitting video clips across the network. It's also a good step toward preparing your network for the inevitable onslaught of rich media.
I've been experimenting with adding useful sound to network presentations for the past few months. Despite vendor claims, standard desktop software can't always cope with multimedia demands. Slide presentation software has a difficult time synchronizing text and speech. And building sound into animated slide shows, a corporate mainstay, is difficult. Unless you use the right tools, your audio efforts could wind up being a lot of effort for not much benefit.
A WAV file, the typical format for digitized sound, tends to be very large and unsuited for transmission across a network, in part because it contains a great deal of information that the human ear can't perceive. A coder/decoder (codec) application can compress those files into smaller encoded files (for more information on networking and audio terminology, see "Glossary of terms," link below).
One technique, perceptual audio encoding, confines its compression activities to sounds outside the range of normal human hearing. Those codecs, including RealNetworks Inc.'s G2, Microsoft Corp.'s Windows Media and the nonproprietary MP3, leave a high percentage of the original sound data in the file after compression.
But your choice of codec depends less on the quality of its output than on its production considerations.
In my tests, G2 shaved a few seconds off total encoding time and also streamlined the production process. Xing Technology Corp.'s MP3 encoder was the hardest to use, although it was generally a fast processor once the settings were correct. Windows Media spent more time in encoding, and its setup was sometimes confusing -- there are a lot of subtle steps you can take to enhance the final result, but learning where they are can be time-consuming. If you plan to compress large batches of sound files regularly, those small differences really add up.
Encoded sound files in hand, I set out to marry them to a slide presentation and publish the package on the Web. I turned a few extra hairs gray as I built a simple Microsoft PowerPoint 2000 presentation to augment a Computerworld Minute -- an audio news report we publish daily on our Web site. Then I tried to synchronize it with my encoded sound files.
Right away, I discovered that, although PowerPoint can play a single sound file across multiple slides, syncing images with appropriate spoken sentences is a tedious matter of trial and error. Lotus Development Corp.'s Freelance Graphics wasn't much better, although it's clearly better suited for building online shows. Rather than spend hours trying to match up the entire presentation, I broke the sound file into components I could sync with a particular slide.
The typical corporate speech includes unwanted dialog or noise. So in addition to PowerPoint, I needed a sound editing tool. I chose Sonic Foundry's $499 Sound Forge 4.5.
Sound Forge is almost the de facto "word processor" of sound editing; you can cut, paste, drag and drop a visual graph of the file, and you can apply effects such as fade-in or fade-out. I used it, for example, to clean up the heavy breathing of one speaker, fade music to silence and fill in dead air. You can save money by opting for the $99 "lite" version, AudioAnywhere, which includes a less extensive version of Sound Forge. That's probably all you'll need, plus Sonic Foundry's Acid Style music authoring tool and hundreds of public domain sound clips.
PowerPoint doesn't handle G2, unfortunately. To make up for it, RealNetworks offers RealSlideshow, a free download that offers more powerful rich media features but is less than stellar at anything else. It's far better at creating online presentations because it tracks overall download time and will automatically publish to a streaming G2 server on the Web. It also lets your presentation carry two separate tracks of audio, so you can have background music and a voice-over simultaneously, something PowerPoint had trouble with.
But RealSlideshow wouldn't accept my imported presentations; it works only with JPEG or bit-map images. Music and image editing tools were almost nonexistent. Although I could specify which image linked to which sound file, fine-tuning timing was impossible. The program mangled our slide images no matter how I tinkered with compression settings, making text almost unreadable. You'll have to add legible text later using RealNetworks' RealText editor, which is a lot harder than simply typing words on the page and adds considerably to production time.
Of course, PowerPoint and RealSlideshow aren't the only programs that can place sound-and-image files on the Web. I had better luck with Kai's PowerShow, $50 from MetaCreations Corp. It handles presentations reasonably well and offers precise sequencing of sound, image and simple animation.
If you'll be augmenting a lot of existing presentations with sound, you should strongly consider switching to PowerShow or other low- to midrange multimedia authoring tools. For higher-end uses, such as building interactive tutorials, consider Digital Lava Inc.'s video publishing system (see "Emerging companies: Digital Lava," link below), which not only joins sound and images but also indexes files by keyword and provides navigation points throughout the presentation.
Computerworld covered the ins and outs of setting up streaming audio servers in a previous issue (see "Get rich (media) quick," link below). Until recently, the choice of a streaming server was dictated by the particular codec you were using. But that's rapidly becoming a moot point: New streaming server software products, such as Sonic Foundry's StreamAnywhere, now in beta, can stream music created by the most popular codecs. (Note: Computerworld uses RealNetworks' G2 encoding and streaming media server to record and transmit the Computerworld Minute. It's an easy-to-use, low-maintenance system, and the player is readily available. Our sound engineer, Aaron Bishop, says there hasn't been much call for the Windows Media editions; because the players are free, most people simply download both. And either player can handle MP3.)
The final results? You can see (and hear) sample files on our Web site. In the end, I chose to put up with PowerPoint's limitations; I simply don't do enough to warrant the steep learning curve of a multimedia authoring tool.
Sound quality: Not a corporate issue
Any of the codecs I tested produced far better audio quality than the typically cheap set of corporate speakers can detect.
The more compression the codec performs on a sound file, the smaller the file size... and the higher the chances for poor-quality sound. Because sound quality is a key strategy in the battle for audio codec supremacy, codec developers have spent a lot of time pointing out deficiencies in rival codecs.
I was curious as to just how much difference the average corporate user would find between leading codecs, so I tested 10 sound and voice files encoded from CD and in the studio.
I encoded the samples at bandwidths suitable for a slow (28.8K bit/sec.) modem, Integrated Services Digital Network and Ethernet network connections, using the latest versions of G2, Windows Media and Xing's MP3.com.
We then judged sound quality using different output devices:
The high-end headphones easily detected differences between original and encoded/compressed sound files. There were marked differences between the encoded samples. I detected subtler differences with the Bose speakers, sometimes drowned out by environmental noise.
None of this mattered once we'd tried the same tests with our "typical" speakers and headphones. Aside from slight differences in volume, the files were nearly indistinguishable.
Lousy speakers produce lousy sound, and on most corporate desktops, that's as good as it gets. Base codec choice on user demand, platform and ease of administration.
Corporate portals grow, get outsourced
RELATED IDG.net STORIES:
Glossary of terms
|Back to the top||
© 2001 Cable News Network. All Rights Reserved.|
Terms under which this service is provided to you.
Read our privacy guidelines.