# What Happens To Your Recording?

Discussion in 'Recording Gear and Equipment [BG]' started by Jay2U, Jan 10, 2019.

1. ### Jay2UNot as bad as he lóòks

Dec 7, 2010
22 ft below sea level
A few years ago I performed a quick test to find out about the playback quality of MP3 and YouTube. In order to keep things simple, I generated a 600 Hz square wave. Square waves consist of the fundamental frequency plus all odd harmonics. The amplitude of each harmonic with regard to the amplitude of fundamental wave, is the reciprocal value of the number of the harmonic. So the amplitude of the third harmonic is 1/3, of the fifth harmonic 1/5 and so on. If all harmonics (say up to 20 kHz) are present, a neat square wave will be the result. If some harmonics are left out or have an incorrect amplitude, the wave will look distorted.

I recorded the 600 Hz signal with a digital recorder which produces a 16 bit uncompressed wav-file. Then I converted this to a high quality 320 kB/sec MP3 file. I also uploaded the original signal to my YouTube account and downloaded it from there. Hereunder the results can be seen, as captured with Audacity. The difference can be heard, even with speakers or headphones of lesser quality.

2. ### bholderAffable SociopathGold Supporting MemberSupporting Member

Sep 2, 2001
Vestal, NY
Glad my music isn't square waves.

3. ### Jay2UNot as bad as he lóòks

Dec 7, 2010
22 ft below sea level
It's the easiest waveform for this experiment. Other complex waveforms get equally distorted during the compression process. By the way, the waveform of a heavily overdriven guitar comes pretty close to a square wave.

Last edited: Jan 10, 2019
bholder likes this.
4. ### bholderAffable SociopathGold Supporting MemberSupporting Member

Sep 2, 2001
Vestal, NY
Well, a pure sine wave signal at 600hz (with no high frequency components) will get distorted a whole lot less, but yeah. It's interesting to see what happens. Waveforms with "corners" or "points" always have high freq components that will get mangled.

In my case, the manglement mostly happens out of the range of my half-deaf hearing, sooooooo......

Jay2U likes this.
5. ### silky smooveSupporting Member

May 19, 2004
Seattle, WA
It's an interesting thing for sure, but I don't think a 600hz square wave is a very representative tool. There's so much more going on with, for instance, a stereo mixdown that would not be as visually apparent when compared in this fashion. You stated as much, but it bears repeating. The functional end of this is that a mixdown of a full composition in WAV format compared to the same mixdown done as a 320kbps MP3 will sound so close that in a double blind listening test I would wager that no one would be able to accurately and repeatedly be able to choose which was which.

And I likes this.
6. ### Jay2UNot as bad as he lóòks

Dec 7, 2010
22 ft below sea level
Square waves represent complex wave forms in a measurable manner. Just encoding and decoding doesn't have to have notable impact on quality, but once compressed, the original composition lost some information that can't be retrieved. Whether this can be heard depends on compression ratio, playback equipment and one's hearing.

7. ### silky smooveSupporting Member

May 19, 2004
Seattle, WA
...and the source material, which was my entire point. The quality of encoding these days is so good that a composition in either format would be audibly identical, or at least audibly unidentifiable as one or the other to the end user despite measurable differences. It’s a case of something being measurable, but not having a practical difference in most cases. I appreciate the science behind it, but as an engineer I put it in the category of “interesting factoid” rather than something that will influence my audio work. Don’t get me wrong, I applaud your work, I’m just trying to give it a working context.

Jay2U likes this.
8. ### Jay2UNot as bad as he lóòks

Dec 7, 2010
22 ft below sea level
I get your point, this is more theoretical than practical. In the 70's I designed and built FM radio transmitters. Modulation of the audio onto the carrier was limited because of allowed bandwidth. Signal processing technology then was all analogue. I kept comparing the original sound with the (modulated, transmitted and detected) sound from the receiver, tweaking the modulator circuit till it couldn't be improved. I think I trained my hearing to detect differences most people don't notice.

9. ### silky smooveSupporting Member

May 19, 2004
Seattle, WA
Which is something my engineering brain omitted on my first read

Jay2U likes this.
10. ### Ukiah Bass

May 10, 2006
YouTube is not a "virgin" repository for music. When the uploaded video is processed by YouTube, it adds AAC compression (I believe at an audio bit rate of 192kbps for a high quality upload and lower rates for lower quality audio). It also "normalizes" audio volume to its own standard of 16 LUFS. So if the upload is overcompressed to boost volume above 16 LUFS and "be louder than other videos," YouTube will lower the volume. This attempt to trick YouTube ends up making your music sound worse in quality.

Essentially, music downloaded from YouTube will never sound exactly the same as a well-produced original format audio file on your local device. Some type of degradation will occur depending on variables. Which makes YouTube a terrible place to "store" high quality audio files you want to re-use later.

ddnidd1 likes this.