Some in the Part 15 endeavor insist on having their programming encoded in lossless FLAC. Others, if they use lossy mp3, will use the definite overkill encoding of 320kbps.
I've mentioned this topic in the past, but thought I'd go into a bit more detail. If a lossy encoding is transparent, or clear, that just means that there is no discernable difference to the ear between the original and the new encode.
For stereo mp3, the bitrate for a transparent encode is around 128kbps. You'd have to have a superb recording, expensive equipment, and a good ear (including being young, as hearing degrades with age) to hear a difference between 128kbps and 192kbps (and even the experts disagree - sometimes you hear what you want to). Anything over that latter bitrate is just wasted bits and space. A transparent mono encode would be 64kbps.
Going to the other extreme, Opus at 64kbps stereo will provide a full range sound (up to 20Khz) and be transparent to virtually everyone. Opus 32kbps is more than sufficient for mono recordings.
Now, FM broadcasting generally produces higher fidelity sound than AM, but even it has limitations. The FM high end is around 15Khz (so it won't interfere with the stereo carrier), and that can be achieved by even lower bitrates than mentioned here, both for mp3 and Opus.
Of course, streaming can be higher bitrates and thus higher fidelity, but it is bound by the same transparency limitations - there is absolutely no listening quality reason to have a stream with more than 128Kbps mp3 stereo.
So why would you store lossy encoded audio files at higher bitrates? Or even use FLAC? There is one valid reason. Every time you do a reencode to a lossy encode, you lose something and degrade the end result. If you're going to do a lot of audio editing, which is usually done in lossless formats requiring back and forth conversion, you will lose quality. Therefore, it doesn't hurt to store files in a higher bitrate to minimize that loss. How much depends, but I usually use 192kbps or 256kbps maximum (less, of course, for Opus). I've never gotten much in FLAC and it certainly isn't as common as mp3 (although getting more so).
But for storing files on your automation computer and playing them, either over the air or in a stream, there's no reason to use those larger, high bitrate files.
As complex as everything Artisan Radio talked about with respect to encoding and bit-rates, there is a whole other part of the story of audio quality pertaining to microphone placement, mixing, and time differences, also known as 'phasing'. No matter what bit-rates are involved in the final digital product, sound must begin and end on the analog plane and whole college courses or lengthy books can be written about the intricacies involved. Looking back to the time when analog recordings were produced in both monaural and stereophonic sound, huge differences in quality resulted from different recording sessions and session engineers and no digital re-sampling can fully overcome the initial flaws. If we take it on as an educational mission of this website we will be trying our hand as an unaccredited audio college.
I am guilty of thinking that doing a couple of OTR series at a high bit rate(insane quality as it was called)) would have it sound better(less muffled) but it can't be made better if it is not at source to begin with so a lot of space is wasted with the higher bit rate for nothing. And yes artisan is right that after a certain point(bit rate) and age you wouldn't notice it anyways. Agree with artisan.
And Carl brings up a very good point also....you can only hear (and see) analog. Digital is just for storage and transmission like shorthand used to be and all MP3s or whatever format is converted back to analog in the end so that with the quality of what you are listening through is actually what you hear. It just made me feel better to think the higher bit rate "may" help the audio quality with some episodes, at least it wouldn't be worse.
The only reason I have everything in MP3 is it's compatible with everything. In fact the volume leveller I have only works with MP3. The automated programs like it the best and way back when when I converted my record collection to digital to do this hobby it was done in MP3. MP3 seems good enough. I have no need to try any other formats. As for "full" 20-20 Hz/kLz nothing in music is that high(or low), even cymbals, tambourine, triangle, etc would just produce a bit of bzz bzz bzz intermittently if at all at 15K and if you are over 60 or 65 you won't even hear that. Well, you could with hearing aids but even hearing aids only care about to 8K. This "full" range from 20Hz to 20Khz is if you are young. And for music, redundant. Nothing is up there.
Improvements are sometimes possible.
Re-sampling to a different format or bit-rate is what we've been talking about, but in some cases 'apparent' improvements in perceived quality CAN be achieved with EQ (equalization). In many instances a +3 dB bump at around 2.5 kHz can add 'presence' to speech. This is a common method used by the copper-wire phone company to make voice calls stand out.
Yes, audio processing is something different entirely. I was talking about attempting to encode the source material as transparently and accurately as possible.
I used to stream at 40kbps mono mp3, and found that was also sufficient quality also for FM. I could notice a bit of a difference streaming between 40 kbps and 48kbps mono (96kbps stereo), and no apparent difference at 64kbps with good headphones and listening intently. For FM, upping the bit rate made no difference at all.
Nevertheless, I do now stream at 64kbps on the off chance that it might make a bit of a difference to some music and some people with excellent ears and great equipment. I still stream mp3, because as Mark states, it is common, and everyone uses it. Earlier versions of IOS, in fact, would not natively support what I wanted from the stream without going to mp3 (I believe, by default, Apple uses AAC). I'm not up on the latest versions, and they may be better now with other formats. Even Windows 10 (the Media Player) has difficulties natively with OPUS, and you have to use alternative applications.
I also found out the hard way that the IceCast directory doesn't support OPUS metadata.