How can you tell if your music files are compressed or not?
Mar 17, 2012 at 3:28 PM Thread Starter Post #1 of 15

Rezound Sound

100+ Head-Fier
Joined
Jan 19, 2012
Posts
356
Likes
13
I just got this music player app on my phone, and it says how many KBPS each song is. Most of my music ranges from 100-400 KBPS. Are Kbps what tell you if your music is compressed?
 
Mar 17, 2012 at 4:24 PM Post #2 of 15
That can tell you how large the file is. That usually, but not always, means how much it's been compressed.
 
Bitrate of uncompressed audio depends on the number of channels in the file, the sampling rate, and the bit depth. CD audio is stereo (2 channels) and has a bit depth of 16 and sampling rate of 44100 Hz. You just multiply all of that together to get the bitrate, which is 1411 kbps in this example. Some audio will have a sampling of 48000 Hz, or a bit depth of 24, or have just one channel. So you can't say for certain that a smaller file or one with a lower bitrate is more compressed, but most audio is CD quality. If you multiply everything out and the real bitrate is lower, it's compressed.
 
Then you run into the question of what kind of compression. Lossless compression makes the file smaller, but it will be bit for bit exactly the same data during playback. Think of it as a ZIP file, you can throw documents in and get them back out later without losing any words. Lossy compression has a still lower bitrate, but achieves this by discarding information we probably won't hear. Some codecs do a better job than others, and some formats are variable bitrate which are smaller than their equivalent constant bitrate brethren, and theoretically sound just as good.
 
There's other problems, too. Someone can take a low bitrate file, that's already been heavily compressed, and repackage it as a high bitrate file. This file will actually sound worse than the original, because the lossy codec discards additional information and pads it with meaningless data to get the desired bitrate. Lossy files can also be placed in lossless containers, which won't damage the file further but won't repair the damage already done. It'll technically be compressed, but not look like it.
 
The only surefire way to spot a compressed file is to open it in some sort of spectrum analyzer and look for a treble cutoff. Lossy files often cut frequencies off above a lower point because we probably won't hear it anyway.
 
Ultimately you shouldn't worry about it.
 
Jan 25, 2013 at 12:00 AM Post #3 of 15
I agree using a spectral analysis is the best way to view if the file is in a very lossy format or if it has been at one time and then converted back. 
 
Depending on your source there is a good deal of files labeled as FLAC that are lower quality formats converted in FLAC which is not good, and is  pretty apparent with artifacting and things like that when you look at the spectral analysis.
 
Jan 26, 2013 at 4:36 PM Post #4 of 15
cool thread i just posted this if anyone wants to compare these two tracks for example.  What are you looking for in a spectral analysis, higher peaks?   http://www.head-fi.org/t/648184/whos-up-for-a-track-quality-comparison-both-320kbps
 
Jan 27, 2013 at 12:39 AM Post #6 of 15
Quote:
cool thread i just posted this if anyone wants to compare these two tracks for example.  What are you looking for in a spectral analysis, higher peaks?   http://www.head-fi.org/t/648184/whos-up-for-a-track-quality-comparison-both-320kbps


We're talking about a spectral frequency analysis. It should look more like this (the pink part):
 
 

 
Audacity just shows you the volume peaks (same as the green part).
 
Jan 27, 2013 at 1:50 AM Post #7 of 15
Yeah i know that part, but how can you tell by looking at a spectrum what is good and what is bad.    Do you mind comparing the files I posted
here are two different ones i compared, i'm assuming the higher peaking one (top file) is -generally- better


 

 
Jan 27, 2013 at 2:43 AM Post #8 of 15
Quote:
Yeah i know that part, but how can you tell by looking at a spectrum what is good and what is bad.    Do you mind comparing the files I posted
here are two different ones i compared, i'm assuming the higher peaking one (top file) is -generally- better


 


Yep, that's correct. But it's pointless information when comparing CD's, this is just for lossy downloads/youtube rips. When the spectral analysis is that close, dynamic range is bound to have a bigger effect on your sound quality. From a technical point of view, track2 should sound better because of that.
 
Jan 27, 2013 at 3:02 AM Post #9 of 15
Well both are 320kbps so these both are considered compressed no?  So it depends on where-ever it was originally sourced from.  One of the files I converted from a flac file, the other not sure of it's source.  
Here's a different example where it goes slightly above the 0db mark (i think that's what it is called).  What does that mean.
How do you check dynamic range not sure what you mean by that.


 
Jan 27, 2013 at 3:39 AM Post #10 of 15
Look back at your first example here: http://www.head-fi.org/t/648184/whos-up-for-a-track-quality-comparison-both-320kbps
 
Track1 has more clipping, and the volume is pressed up against the edge more. If you were to reduce the volume on each track, you would see that track2 has larger peaks and valleys, aka dynamics, and would sound better as a result. Now, with the differences being this slight, it's hard to hear a difference at all and you shouldn't even worry about it. This video explains it much better than I can. http://www.youtube.com/watch?v=3Gmex_4hreQ
 
Jan 27, 2013 at 4:40 PM Post #11 of 15
Ahhh yes this makes sense, I get that it isn't technically compression now.   What about the spectrogram above- does that have missing information?  Or does it just have a more well-rounded wavelength?
 
Jan 27, 2013 at 5:01 PM Post #12 of 15
Quote:
Ahhh yes this makes sense, I get that it isn't technically compression now.   What about the spectrogram above- does that have missing information?  Or does it just have a more well-rounded wavelength?


I'm not sure what you mean by "well-rounded wavelength", but the one on top is indeed missing some of the high frequencies.
 
Jan 27, 2013 at 11:25 PM Post #14 of 15
Quote:
I'm not sure what you mean by "well-rounded wavelength", but the one on top is indeed missing some of the high frequencies.

I guess I just meant there's more data there on the bottom one than the top one.  Although, the bottom one does go too high in certain spots and peak off the chart so I wasn't sure which is worse.
 

Users who are viewing this thread

Back
Top