Cut MB from 36 to under 25 to enable email

[kozikowski]

MP3 and other compressed format damage is sneaky. You can slide into low bit rate damage before you ever realize you're doing it. This was intentional.

At extremes, a highly compressed show will sound honky (speaking through a megaphone) gargly and bubbly. Sometimes like a bad cellphone call and for the same reasons. Compression is everywhere.

My compression joke: Two cheap violins and a Stradivarius. After compression, it's three cheap violins.

That's the quality that gets thrown away in compression. It's not smack you in the face obvious.

Koz

[steve]

Koz has already answered most of your questions, but for a choice of explanations (take your pick, they both mean the same thing):

1: the possible negative effects of using "stereo to mono" when the source mp3 is 2 channel mono (IE; a 2 channel mp3 of a pre-stereo era recording).
From your prior comments, I assume the chance of a negative effect is virtually zero--one instance in your entire experience.

Depending on how the 2 track mono recording was made, you may notice a very subtle difference between the 2 track mono and the 1 track mono versions (probably only just noticeable if you listen with headphones). Essentially the difference is that any background noise (for example tape hiss or record crackle) may have been true stereo on the 2 channel version and will now be true (1 channel) mono background noise. If this occurs, the result may be subjectively slightly better sounding than the original because the noise may be less obvious. (Note the emphasis here that this may or may not be the case, but the difference, if any is likely to be very subtle).

In the 1 case that I've seen on the forum, the problem was that the original 2 channel version was faulty. The left and right channels were "out of phase" (don't worry if you don't know what this means) - the result was that the mono version was almost silent - the problem was really obvious and you would notice the problem immediately by listening to 1 second of the exported mono file.

Giving a quick listen to the exported file, preferably before closing the Audacity project, is (in my opinion) always a good idea, juts to check that nothing has gone wrong (such as a silly operator error that completely messes up the export).

So short answer - the possible negative effects of using "stereo to mono" when the source is 2 channel mono are in 99.9% of cases likely to be none, or insignificant. The exception to the rule is if the two channels are out of phase, then the negative effect will either be a silent (or near silent) exported file, or a "phasing" wooshing sound throughout the recording.

I have it in my head that it is bad form to open and re-export an mp3 file repeatedly because a little more data is thrown out (forever and irretrievably) on each new export.

Absolutely correct.
The "damage" done to a recording by the MP3 format occurs when the audio is encoded (compressed). The higher the quality (higher bit-rate) of the MP3, the less damage there is, (VBR will usually produce less damage overall than CBR or ABR of an equivalent bit rate/file size), but there will always be some damage, even if it is not noticeable.

Decoding (un-compressing) MP3s does not produce any additional damage.

When you open an MP3 in Audacity, there will already be some damage to the sound due to the original encoding, so let's say that the original MP3 file has 0.1% of "damage". When you import the file into Audacity, it is decoded, producing an exact replica of the MP3 file, but as uncompressed data. Thus the track in Audacity faithfully reproduces the MP3 and still has the 0.1% damage. When you export the track from Audacity - if you export it as WAV, then it will still have that 1% damage, but no more. However, if you export as MP3, then it will again loose a bit of sound quality and the sound will be damaged further. So there is an additional 0.1% damage due to encoding losses. What makes it even worse, is that the original 0.1% damage is damaged further, so the exported MP3 file will have worse than 0.2% damage.

In this example we are looking at very small amounts of "damage" (data loss), so it is assuming that each MP3 is a reasonably high quality MP3. In this case, you can probably re-encode the track several times before the sound becomes unacceptable. If higher compression amounts are used (lower bit-rates/lower sound quality) then the "damage" will be more severe and will build up much more rapidly. The damage is irreparable.

If you want to test this out - try opening a stereo audio file, then export it as "test1.mp3" at say 64kbps. You may notice that it does not sound quite as good as the original.
Now open test1.mp3 and export it as test2.mp3 (again at 64kbps). It will sound worse. Repeat this a few times and it will soon sound terrible.

If you repeat this test using 320kbps (preset = "Insane"), you will notice that the damage increases very much more slowly, but it does still increase and will eventually become noticeable.

[ignatz]

Koz and Steve:

Thanks for the additional detail. It pretty well confirmed my suspicions and corresponded to what I had previously understood.

I just checked on some info. My average mp3 file size is almost exactly 4 MB and the average running time of my mp3s is 2'33. The catalog is split pretty equally between mono, joint stereo, and stereo.

I can't compute the average bitrate, but the median bitrate is near 190 kbps (half above; half below).

Nearly all of the sub 100 kbps items are mono.

I don't use portable devices and have no need to further compress files so they will fit on a portable.

Most of my mp3s are never reopened or edited in Audacity. Perhaps 15% are and in virtually every case it would be a one-time edit, resaved as joint stereo quality 2 in Audacity.

Given the above, I think I am on safe ground so far, but I am continually listening to songs and making note of any fidelity issues. Thus far, I have not noticed any general loss of fidelity that is likely linked to re-exporting.

[steve]

Your welcome