Someone said on this forum that if you convert 128 kbps to 128 kbps, it will effectively cut the quality in half to 64. Is there any conversion loss table that shows what the loss would be with other combinations? I commonly save 192 from mp4 to 320.
That sounds like an exaggeration, though there will be some loss of sound quality.
“Quality” is highly subjective, and among other things, it depends on the type of material being encoded. For a listener, orchestral music is more likely to show the effects of MP3 compression than say generated test tones, but it may be easier to measure the loss in the generated test tones than in the orchestral music.
High bit-rates will preserve the sound quality better than low bit-rates, and “lossless” formats (such as WAV and FLAC) are best of all, but at the end of the day it is your ears that are the only judge.
That’s an approximation. It could happen. It’s the illustration of why it’s a terrible idea to do production in MP3 or any other compression format. Successful modern compression algorithms can pay attention to content and it’s not simple to publish a conversion table.
As a fuzzy approximation, pick the midpoint between the two compression rates and take half.
There are rules of thumb.
You should take the AudioBook compression values of 192 Minimum at their word. If you can get your chapters to fit in their other requirements at higher compression, do it. The ACX post-production products will come out better and more reliable the higher you go.
You get another significant quality improvement if you submit mono instead of stereo.
It’s true if you stick to the higher compression rates, you can go a very long time before you experience any audible damage at all. But. Post production or other unexpected sound management can just kill you.
There is a poster who was doing a show for a small local radio station. To cut a long story, they couldn’t make an MP3 podcast of his show because the MP3 music he used turned to bubbly trash.
Do all your work in WAV and only convert to compression if you’re reasonably sure it’s the final product. I have one personal recorder with no provision to record at perfect quality WAV. It’s in a box somewhere in the bottom of my desk. It’s pretty to look at.
And there’s another super fuzzy minimum rule. Right around 32, most people can just start to hear some sound damage in a mono show if they’re not exposed to the original work. 64 for Stereo. The transitions are pretty steep. Lower than 32, Mono sound turns to garbage immediately. Above 32 the damage becomes very much less apparent. Audacity stereo MP3 default export used to be 128 which is “safe.”
But not safe for production. Safe to listen to with little or no apparent damage.
Compressed formats in production are a land mine. Nobody knows it’s there until Blammo!
“My show sounds like a bad cellphone when I made the MP3 to upload…”
Thanks for the thoughtful replies. I understand the advice to use lossless, but that isn’t practical or even possible in this situation, I use 192 kbps audio I extract from the standard HD online music video (720p) using Audacity. I save it at 320 kbps and that is what goes on the air. I have good reason to believe that the 192 is made directly from original commercial quality audio. The station equipment does not support WAV, and for prerecorded shows, the program manager doesn’t want enormous files like WAV. I need to investigate the other lossless types. The sound I get from the mp3 is good enough, though probably could be better. The questionable part is whether it retains fidelity after conversion to the station’s 128 stream. So the specific question is whether the 320 from 192 would technically (leaving subjectively aside) be reduced below the 128 quality capacity of the stream, or (I’m hoping) be higher than the 128 stream can communicate, thus be as good as it can be. The “fuzzy approximation” would make it 128. This is assuming “all other factors being equal”, meaning that I am actually getting good 192, and not just 192 made from trash. I may be the local programmer you mentioned, Kos. But the station management never complained about my audio, it is I who am so determined to do it the best way possible. To do my shows, which features the lesser known, I must use online, because the station library doesn’t have most of what I need, and I can’t afford thousands of CDs. Hopefully, the artists are happy for the exposure. Anything you think I could do better, under the circumstances?
I will look into whether FLAC will work on our equipment. Even if it does, if FLAC is a very large file like WAV, that can also be an issue. What is the smallest lossless type? The mp3DirectCut looks interesting, although I don’t understand it. It will do things Audacity won’t do?
As a programmer, I cannot afford to buy the music I use. $0.99 or $1.29 adds up fast, thousands of dollars per year it would be. I need to collect several times as much as I use in order to have the selection options on hand for immediate review or use, because it is not practical (or even possible) to recall and chase down each item online at the time I need it. Much of what I need is international or otherwise obscure, and not available through the usual commercial sources. Although I am not a commercial promoter, what I do is effectively free promotion. The station subscribes to a system which pays the royalties for what we play.
From your description, you have achieved the quality limit of your pipeline. The only way to get an observable increase in quality is to change one or more steps.
— Deliver in WAV.
— Start with uncompressed music.
Gently massaging the compressed deliverable in the middle will not give you any listening quality increase in the face of damage from the other steps. Further, it could cause problems if you settle on a celebrity format and Something Goes Wrong.
Weren’t we going to deliver on a physical Audio CD rather than the music file? That is perfect quality WAV deliverable. What happened to that?
I need to provide programs in hour segments. WAV files proved too large to burn to a disk. Also, flash drive is more dependable (doesn’t mess up) and the program manager prefers it. WAV on a flash drive will not read on the studio equipment.
It does seem that I have done about the best I can for audio quality with the available options, from what you all explained.
The question about WEBM OGG and WEBM audio only OPUS is: what are they and how do they compare to other file types discussed?
I’m talking about these as original source files, not conversions to them.
I need to provide programs in hour segments. WAV files proved too large to burn to a disk.
Maybe a Data CD. An Audio CD like you buy in a store will record 74 minutes. Audio CD is a special, space-saving, loss-less version of WAV format. For example, one of the ways they did that was to eliminate options. Audio CDs always have the same sound format and the player always looks for that one format.
An Audio CD burning program will ask you about the space between songs (default is two seconds). A Data CD burning program doesn’t. A Data CD is a flat, shiny hard drive. You can put Photoshop files on there if you want. An Audio CD is limited to sound files.
If your computer will burn a disk by just dragging files to a disk icon, chances are very good you got a Data CD. Audio CDs are created by an Audio CD Authoring and Burning Program such as Windows Media or iTunes.
WEBM OGG and WEBM audio only OPUS is: what are they
First I’ve heard of them. Maybe one of the other elves will chime in.
Well, you can listen to the files after importing them. For the same file size, OGG is probably higher quality than all but the highest bit rate MP3’s. OPUS files are very small. For classical music I don’t find them as good as high bit rate MP3. You can have terrible sounding files in any of those formats, however high the encoding settings are.
The file size of the original files is immaterial, because Audacity will expand them to lossless PCM when it imports them.