[Febs]

You need to use ReplayGain to ensure that the two files you are comparing are volume matched. Otherwise, the difference that you hear that enables you to ABX the two tracks may simply be a difference in volume.

 

[Cid]

Quote:

Originally Posted by Febs /img/forum/go_quote.gif You need to use ReplayGain to ensure that the two files you are comparing are volume matched. Otherwise, the difference that you hear that enables you to ABX the two tracks may simply be a difference in volume.

The volume can change just by changing codec? It's 1am here, I'll do more tests tomorrow with ReplayGain on. Thanks for letting me know.

 

[EnOYiN]

Mp3s do indeed change the volume. On average this is only 0.01 dB. An mp3 does have a higher peak in the dB. This is on average 0.15 dB. This is actually audible. You will need to use replaygain to get the exact copy of a track with a different encoder.

@ Altoids
I was not aware the you wanted me to ABX 320 CBR vs Lame V0. I will not be able to hear that. I do not know anyone who can. Most likely that is not audible with human ears.

However. This thread is not about comparing 320 vs v0. It is about the difference between lossless and lossy.

 

[Cid]

Well, I just did it with ReplayGain on and I got the same result. I don't see what the difference was with it on or off, but whatever. I'll post screen shots and logs later, I can't be bothered right now.

 

[EnOYiN]

Quote:

Originally Posted by Cid /img/forum/go_quote.gif Well, I just did it with ReplayGain on and I got the same result. I don't see what the difference was with it on or off, but whatever. I'll post screen shots and logs later, I can't be bothered right now.

Like I said before. The difference is 0.01 dB on average. Which is not audible at all. It COULD be more. It doesn't even have to be. As for the results, I believe you without them.

 

[AppA]

I've got a huge question, is it true that when you play a perfectly encoded 320kbps mp3 on a big PA system (for example on a big festival), that you get a 'muddier' and lesser sound then for example a CD?

greets,
appa

/edit:
hey first post

 

[EnOYiN]

Quote:

Originally Posted by AppA /img/forum/go_quote.gif I've got a huge question, is it true that when you play a perfectly encoded 320kbps mp3 on a big PA system (for example on a big festival), that you get a 'muddier' and lesser sound then for example a CD? greets, appa /edit: hey first post

Welcome to Head-Fi!

There is no such thing as a perfectly encoded mp3. That is why it is a lossy file format. If you want to get perfect encoding you will have to use a lossless encoder like FLAC or Ape.

How it sounds will have to with other things than just encoding. A DAC ( digital analog converter) is a part which is really important. There is a DAC in every CD player and in every soundcard. If the DAC in your computer is better than the DAC in your CD player, most likely the sound coming out of your computer will be of higher quality.

So, no not necessarily.

 

[AppA]

Quote:

Originally Posted by EnOYiN /img/forum/go_quote.gif Welcome to Head-Fi! There is no such thing as a perfectly encoded mp3. That is why it is a lossy file format. If you want to get perfect encoding you will have to use a lossless encoder like FLAC or Ape. How it sounds will have to with other things than just encoding. A DAC ( digital analog converter) is a part which is really important. There is a DAC in every CD player and in every soundcard. If the DAC in your computer is better than the DAC in your CD player, most likely the sound coming out of your computer will be of higher quality. So, no not necessarily.

Why thanks!
Ah, sorry. I meant 'encoded with minimal loss'.

So it could sound sharp and direct with the right DAC?
Also, are the PA DAC's different then the normal (Home listening) ones?

 

[EnOYiN]

Quote:

Originally Posted by AppA /img/forum/go_quote.gif Why thanks! Ah, sorry. I meant 'encoded with minimal loss'. So it could sound sharp and direct with the right DAC? Also, are the PA DAC's different then the normal (Home listening) ones?

Yes. You got it right. There are however home DACs that will be good enough for professional use. You can go here if you want to get some info:
http://www.headphone.com/

For professional DACs:
http://www.dcsltd.co.uk/about_dcs.html

You can also use Asio. (kernel streaming) I tried to make a thread about it. That is located here:
http://www.head-fi.org/forums/showthread.php?t=223586

 

[Joshatdot]

I did some tests afew years ago comparing different bit rates of mp3's. It's dependent on how you configure the mp3 codec. I did these tests on Cool Edit Pro 1.2a, I created a white noise wav file, and then encoded them to mp3 with different codecs and various bit rates.

http://www.cnw.com/~josh/MP3_Tests/mp3test01.htm

 

[ShadowVlican]

Quote:

Originally Posted by Joshatdot /img/forum/go_quote.gif I did some tests afew years ago comparing different bit rates of mp3's. It's dependent on how you configure the mp3 codec. I did these tests on Cool Edit Pro 1.2a, I created a white noise wav file, and then encoded them to mp3 with different codecs and various bit rates. http://www.cnw.com/~josh/MP3_Tests/mp3test01.htm

that shows nothing except their high cutoff point, which like u said can be manually configure with any MP3 encoder (yes, LAME, rated last on ur list, can be configured to a higher frequency cutoff if that's your fancy)

 

[Joshatdot]

Quote:

Originally Posted by ShadowVlican /img/forum/go_quote.gif that shows nothing except their high cutoff point, which like u said can be manually configure with any MP3 encoder (yes, LAME, rated last on ur list, can be configured to a higher frequency cutoff if that's your fancy)

Yeah, my tests are about 5 years old and codecs have improved alot since then.

 

[Jokieman]

Quote:

Originally Posted by Cid /img/forum/go_quote.gif Well, I just did it with ReplayGain on and I got the same result. I don't see what the difference was with it on or off, but whatever. I'll post screen shots and logs later, I can't be bothered right now.

There actually can be a big difference on some, if not many songs. On others there is very little difference. For some reason, part of the process of compressing a song file via an MP3 Codec raises the volume level, and on a lot of music that is mixed at the studo level with max dB in mind, it can actually cause the music to clip and distort.

Some people will mistake that clipping/distortion as a loss in SQ, yet if they go in and lower the dB levels back to a non-clipping state, it can become a lot harder to tell the difference between compressed and uncompressed.

The reason why is because this particular aspect of MP3 compression is not a permanent loss in quality because the clipping/distortion can be fixed via a program like MP3 Gain or Replaygain in Foobar.

A True loss in SQ comes from data being lost due to the compression. The clipping and distortion many people hear in compressed MP3 music is not that lost data rearing it's ugly head, but instead the volume level already being maxed out in the original file, being boosted over the top by the MP3 compression.

So any true testing needs to take this into account and the dB levels of the compressed file need to be lowered to match the dB level of the lossless file before doing the test.

 

[Rempert]

Quote:

Originally Posted by Jokieman /img/forum/go_quote.gif For some reason, part of the process of compressing a song file via an MP3 Codec raises the volume level, and on a lot of music that is mixed at the studo level with max dB in mind, it can actually cause the music to clip and distort. Some people will mistake that clipping/distortion as a loss in SQ, yet if they go in and lower the dB levels back to a non-clipping state, it can become a lot harder to tell the difference between compressed and uncompressed. The reason why is because this particular aspect of MP3 compression is not a permanent loss in quality because the clipping/distortion can be fixed via a program like MP3 Gain or Replaygain in Foobar.

How sure are you about this information? I don't know enough about the complex process of mpeg encoding to say that you are wrong, but I am quite skeptical. In order for an mp3 to clip at 100% volume, and yet sound perfectly fine at a lower volume, it would be necessary for the file to provide extra headroom beyond whatever is being defined as the 100% value. Since absolutely no sound information exists in this space in the original wave, why would we "round up" into that space, wasting bits on it that could be used to more accurately describe samples within the range where sounds actually exist? After all, the goal with lossy compression is more or less to describe each sample or group of samples in the least number of bits possible without creating a noticeable distinction from the original file. Going beyond the ceiling of digital audio is counterproductive on both points, requiring more bits and causing a clear audible difference.

 

[EnOYiN]

Quote:

Originally Posted by Rempert /img/forum/go_quote.gif How sure are you about this information? I don't know enough about the complex process of mpeg encoding to say that you are wrong, but I am quite skeptical. In order for an mp3 to clip at 100% volume, and yet sound perfectly fine at a lower volume, it would be necessary for the file to provide extra headroom beyond whatever is being defined as the 100% value. Since absolutely no sound information exists in this space in the original wave, why would we "round up" into that space, wasting bits on it that could be used to more accurately describe samples within the range where sounds actually exist? After all, the goal with lossy compression is more or less to describe each sample or group of samples in the least number of bits possible without creating a noticeable distinction from the original file. Going beyond the ceiling of digital audio is counterproductive on both points, requiring more bits and causing a clear audible difference.

It is true that mp3s tend to have a higher output volume than lossless files. Those are all at 0.9999999999 dB. Why this is I do not know. Maybe someone here can explain it since (like you already said) an mp3 would create something new which does not exist in a (good) source.

I would also like an explanation for this. I have been looking for it on the internet but I wasn't able to find it.