Loss of quality through compression

When you edit your LP tracks and mp3s to try to recreate the 45 version, or indeed when you import a track from a CD to iTunes and then burn it again to a CD-R, do you keep losing large amounts of quality?

In my naivety, I always thought that since digital info is all “1’s and 0’s”, transferring that info back and forth did nothing to diminish sound fidelity.

But with recent discussions about file compression and kbps, I am left wondering: If ripping a song from a CD and thus transferring it to an mp3 file significantly lessens the sound quality, what about then burning that file again to CD?

Does all that stretching back and forth keep degenerating the purity even further?

I am reminded of my frustration in the eighties and nineties over taping and re-taping songs from audio cassettes and listening to the generational quality slip away …

You are absolutely correct that you will lose sound quality when you convert a song to an mp3. Converting it back to a wav file or burning it on a CD does NOT bring back the sound quality; however, it does not make it any worse either. It will only get worse if you convert it to an mp3 again. (For example, rip a CD to an mp3 = 1st generation loss; convert mp3 to wav or burn to CD = no further loss; convert that same file back to mp3 again = 2nd generation loss.)

When editing my audio, I always start off with the best source I have available, which is never an mp3. I always use wav files, and I never convert the finished product to mp3 for the "master." If I make an mp3 out of a song, I always have the original wav file somewhere, too.

You should never experience any sound quality loss when you are going back and forth from CD to wav file, no matter how many generations it is. The only danger with that is making sure you didn't pick up any errors in the process. On occasion, I'll find a CD-R that my drive/recorder doesn't like to read. When I play back the file, it skips or has a short blank space somewhere in the middle where data is missing. Slower ripping speeds usually reduce or eliminate any errors that could occur, but there are many other factors that could also give you read errors.

The important thing to remember about mp3s is that they will never sound as good as an original CD or wav file that has not been compressed. In order to fit the same amount of music in a smaller file size, some of the audio information gets thrown away (hence the sound quality loss).

It's true that you lose quality when converting to mp3, but you can lose a tiny bit of quality or a great amount. For instance, I make mp3s at 256 kbps. Further, when making the mp3s, my software asks me "High quality, medium quality, or low quality?" (Low quality is transferred at the fastest speed. High quality takes longer to transfer (although not much) but is higher quality!)

My friend MMathews, who contributes to this board, has told me that if you select "high quality" and "256" (and btw, 320 is even better quality), the ear really only senses a slight loss of treble, about 0.5 dB at 12k according MMathews. Therefore, if you really didn't want any noticable quality loss, you can pre-raise 12K by 0.5 dB, then make a 256 mp3, choosing "high quality" if your software gives you that option.

Now, while it's true that this mp3 has less info in it - if you A-B'd this mp3 with the original WAV, would someone like MMathews or me even hear the difference and even know which is which? He has this theory (and now me, too) that if the ear can't tell which is which, it doesn't really matter that one is less quality. After all, it's really what our ears are detecting that matters, not what the file LOOKS like on the computer.

BTW, I don't pre-raise the 12K area by 0.5 dB, but I do make 256 mp3s and choose "high quality" -- and I can't really hear any difference or that I'm listening to mp3s! (And people say I have very good ears. In fact, I rather wish I didn't!)

Edited by EdisonLite

Wonderfully explained, guys - thanks so much!!!

Lots of good additional info here, guys. There is one point that I would like to clarify, though:

EdisonLite wrote: ...if you select "high quality" and "256"...the ear really only senses a slight loss of treble, about 0.5 dB at 12k according MMathews.

This is not actually true. While it may sound like you are losing a slight loss of treble, converting to an mp3 does not adjust the EQ of a song at all. Instead, it throws away entire frequencies that are less important to the ear. If the range of frequencies we can hear starts at 20 Hz and goes to 20,000 Hz, an mp3 compressor will discard any number of frequencies in that range.

Yes, you may hear a loss of treble; however, you cannot get that frequency back by boosting the high-end of your EQ. Whichever frequencies were discarded are gone forever. This is the reason that you can never make a low-quality mp3 sound good no matter how much you adjust the EQ.

I agree with EdisonLite that I do not really hear a decrease in treble on a 256kbps mp3; however, I can certainly hear the compression, especially in the high end. The encoder usually leaves behind some audible artifacts that weren't there before. It's hard to describe, but it's almost like a fluttering of the high end frequencies, and it gets worse as you use lower bit rates.