When discussing audio, the term “compression” can have more than one meaning.
This is especially useful when streaming audio over the internet, as the bandwidth required for acceptable sound can be reduced to less than 10 percent of the amount that would be needed to pass the original “uncompressed” signal. Data reduction is also central in HD radio broadcasting.
Despite the fact that you have “tossed out” over 90% of the original information, the sound can be quite listenable.
It’s important to start with the best quality audio. Whenever possible, use the original, full-sized (uncompressed) audio direct from a CD or other high quality medium. Only convert the audio to .mp3 after you have completed all the editing processes. With hard drive costs as low as they are today (1TB – that’s 1000 gigabytes – is commonly less than $100), there is no reason to use anything less than uncompressed linear audio as your source material. Yes, the difference is clearly audible to the end user.
When it’s time to convert the audio to .mp3 (or any other data reduced format), use an encoder that includes processing techniques to minimize the “swirly” sounds often heard on low-bandwidth streaming audio. We have software encoders like Omnia A/XE and F/XE, and hardware encoders and processors by Telos, Omnia, Inovonics, Wheatstone, and others. Careful attention to the encoding process can make even a very low-bitrate stream sound impressively good.
The second type of compression is Dynamic Range Compression. This type reduces or “compresses” the range between loud and soft. There is still a difference… just not a much as on the original source material.
Most radio listening is done in a car, office, or on a mobile device. The listener is dealing with noises in their surroundings that would make it difficult to hear quiet sounds, so we work to increase the quiet parts, while at the same time controlling the loud parts so they don’t distort.
So – two completely different types of compression. One reduces the size of the file; the other reduces the range between loud and soft. Both are very necessary in broadcasting. You use both types every day.