What happens to your recording?

[Jay2U]

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.

 

[ahpook]

Interesting, but I'd ask is using a square wave to test the fidelity of MP3 encoding a reasonable test ? It was designed to encode music and takes into account the psychoacoustic effects of human hearing, it's not designed to create a perfect copy.

Edited January 10, 2019 by ahpook

[JapanAxe]

I get what the OP has done, but no audio kit made has ever been made that can reproduce a square wave faithfully!

[Jay2U]

14 minutes ago, ahpook said:

Interesting, but I'd ask is using a square wave to test the fidelity of MP3 encoding a reasonable test ? It was designed to encode music and takes into account the psychoacoustic effects of human hearing, it's not designed to create a perfect copy.

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.

[Jay2U]

20 minutes ago, JapanAxe said:

I get what the OP has done, but no audio kit made has ever been made that can reproduce a square wave faithfully!

Some audio kits that come close within the audible spectrum. Those are the one's that reveal the difference between original and compressed recordings.

[ahpook]

47 minutes ago, Jay2U said:

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.

Well, yes....MP3 is a lossy format, I'd expect to see signal degradation.

[JapanAxe]

13 hours ago, Jay2U said:

Some audio kits that come close within the audible spectrum. ...

Exactly. To generate or reproduce a true square wave, you would need devices with infinite slew rates. A true square wave contains a portion of every odd harmonic, way  beyond what is audible to anyone or anything!

I do take the point of your test though.

[bazzbass]

so you're saying MP3s are poorer quality than the original uncompressed wav file?

MIND BLOWN 

 

[Jay2U]

52 minutes ago, bazzbass said:

so you're saying MP3s are poorer quality than the original uncompressed wav file?

MIND BLOWN 

 

Yes, for sure! Any compression takes data away from the original. The size of MP3-files often is less than 1/5th of the original file. This can be compared with images. A jpg-file, derived from an uncompressed bitmap or bmp-file, is less detailed.

[bazzbass]

14 hours ago, Jay2U said:

Yes, for sure! Any compression takes data away from the original. The size of MP3-files often is less than 1/5th of the original file. This can be compared with images. A jpg-file, derived from an uncompressed bitmap or bmp-file, is less detailed.

lol I was being sarcastic