Head-Fi.org › Forums › Equipment Forums › Computer Audio › Reasons I don't hear a difference between 128kbps, 256kbps, and ALAC??
New Posts  All Forums:Forum Nav:

Reasons I don't hear a difference between 128kbps, 256kbps, and ALAC?? - Page 2

post #16 of 43
Quote:
Originally Posted by Jaywalk3r View Post

I think you might be interpreting error in a manner different than intended. An estimate is virtually always wrong, but can often be very close to correct. Suppose you measure the length of a carrot with a ruler that is only graduated down to 1/8 inches. If one end of the carrot is lined up with the zero mark, the other end is extremely unlikely to lie exactly on a n/8 mark. The exact length remains unknown, but can be estimated with reasonable accuracy. The difference between the actual length and the estimated length is an error, but it might not be (probably isn't, at least in this example) consequential.

I guess I had always assumed that the loss occurred at the encoding end, where some of the signal was selectively removed to reduce the file size. I guess that was kind of naive smily_headphones1.gif
post #17 of 43
Quote:
Originally Posted by Jaywalk3r View Post


No, that's not correct.

That the quantity of data is reduced does not imply that sound is missing, i.e., no longer producing SPL. At least for audible frequencies, much of the "missing" data is estimated and replaced during playback. Compared to the original data, the estimations inevitably have errors, we should be careful not to assume that the error always results in the lossy passage being quieter than the original. It's just different. If we wish to determine whether or not that difference is audible, we must level match for our ABX tests.

 

Since no volume adjustment involved in encoding, shouldn't that the "error" to be the artifact of lossy codec?

 

As you've said, the replaced missing data is of estimated thus may be louder and quieter here and there, I don't really think there is any appropriate/practical way to do proper level matching...


Edited by kn19h7 - 1/22/13 at 4:06am
post #18 of 43
Quote:
Originally Posted by kn19h7 View Post

 

Since no volume adjustment involved in encoding, shouldn't that the "error" to be the artifact of lossy codec?

 

As you've said, the replaced missing data is of estimated thus may be louder and quieter here and there, I don't really think there is any appropriate/practical way to do proper level matching...

 

Of curse there is.  Use Foobar abx tool - and simply make sure you apply replay gain to the tags.  Easiest way to volume match.  Then simply make sure you tick the blind test box when performing the abx.  Run at least 15 iterations.

post #19 of 43
Quote:
Originally Posted by Brooko View Post

 

Of curse there is.  Use Foobar abx tool - and simply make sure you apply replay gain to the tags.  Easiest way to volume match.  Then simply make sure you tick the blind test box when performing the abx.  Run at least 15 iterations.

 

Replaygain apply a single gain value to an entire song... as my last post has mentioned, there may be quieter parts as well as louder parts in the song.


Edited by kn19h7 - 1/22/13 at 4:37am
post #20 of 43

That's the first time I've heard an argument against level-matching. I can't say I have enough knowledge of how replaygain is calculated to dispute the fact that it would make the compressed file sound louder overall(as much as it seems like it would conflict with the entire point of it) but...

 

If you're comparing a compressed file to a lossless file(or less compressed file), you should be looking for artifacts. You'd still be able to hear them just as well. If level matching makes you think a 128kbps file is better than a lossless file due to it sounding a tiny bit louder, then there isn't anything wrong with that 128kbps to your ears.

post #21 of 43

Quote:
Originally Posted by Jaywalk3r View Post


No, that's not correct.

That the quantity of data is reduced does not imply that sound is missing, i.e., no longer producing SPL. At least for audible frequencies, much of the "missing" data is estimated and replaced during playback. Compared to the original data, the estimations inevitably have errors, we should be careful not to assume that the error always results in the lossy passage being quieter than the original. It's just different. If we wish to determine whether or not that difference is audible, we must level match for our ABX tests.

At certain bitrates, lossy encoding can lowpass-filter the sound, so he isn't wrong. At 128 kbps for MP3, the cutoff frequency could be 17 kHz (this appears to be the case with LAME, see: http://wiki.hydrogenaudio.org/index.php?title=LAME#Recommended_settings_details).

post #22 of 43

My point is just that I think the correct way to blind test lossy/lossless or 128k/320k etc, shouldn't include replaygain, since no volume adjustment is involved in the transcoding process.

 

I am not against lossy codecs, harddisks are expensive to me.

post #23 of 43
You should use Replaygain when ABXing because if the only difference is a slight change in volume, that's not very interesting, and doesn't say anything about the codec's ability to reach transparency.
post #24 of 43
Quote:
Originally Posted by Brooko View Post

Of curse there is.  Use Foobar abx tool - and simply make sure you apply replay gain to the tags.  Easiest way to volume match.  Then simply make sure you tick the blind test box when performing the abx.  Run at least 15 iterations.
Quote:
Originally Posted by skamp View Post

You should use Replaygain when ABXing because if the only difference is a slight change in volume, that's not very interesting, and doesn't say anything about the codec's ability to reach transparency.

As I understand ABX best practices for testing, to eliminate the bias introduced by differences in volume, you need leveling at the 0.1 db level because even a 1db change in volume level can influence the results. Replaygain doesn't have the level of discrimination. So sure, maybe you might reduce the likelihood of the volume level bias with replaygain, but you can't guarantee it's eliminated.
post #25 of 43
Quote:
Originally Posted by cel4145 View Post

Replaygain doesn't have the level of discrimination.

[citation needed]
post #26 of 43
Quote:
Originally Posted by skamp View Post

[citation needed]

Doah! Why was I thinking replaygain was only 1 db measurements.

I need to stop reading head-fi and responding when busy at work.
post #27 of 43
Quote:
Originally Posted by skamp View Post

You should use Replaygain when ABXing because if the only difference is a slight change in volume, that's not very interesting, and doesn't say anything about the codec's ability to reach transparency.

 

Replaygain is based on psychoascoustics analysis, the difference in measured perceived loudness(s) is basically a form of numerical measure of the codec's artifacts.

 

Speaking of codec transparency, if the only thing I've done to the lossless file was to encode a lossy file from it, shouldn't the lossy codec be responsible for any perceived differences?

 

hmm.. I would take how applying gain to the song impacts the difficulty of a blind test as another topic...

post #28 of 43
It would be stupid to dismiss a lossy codec solely on the ground that it incurs slight variations in volume, if that can be fixed with Replaygain.

Also, I actually apply replaygain while transcoding, so that it works on devices/players that don't support Replaygain, like the iPod Classic's original firmware. Are those files audibly different from their source? Obviously they are, since I've applied large negative gains to them! Does it mean that they're not transparent? Absolutely not, it's just a matter of adjusting the volume.
Edited by skamp - 1/22/13 at 9:11am
post #29 of 43
Quote:
Originally Posted by kn19h7 View Post

My point is just that I think the correct way to blind test lossy/lossless or 128k/320k etc, shouldn't include replaygain, since no volume adjustment is involved in the transcoding process.

 

That is not true, especially at low bit rates the level is often attenuated by a small amount (~0.5 dB). Such difference is already enough for a positive ABX result, even if the files would contain otherwise identical sound. For a fair comparison, the louder file needs to be attenuated on playback to match the levels; with 24-bit output resolution being available on any decent DAC and even onboard HDA codecs, the effect of this on sound quality is negligible compared to that of lossy compression.


Edited by stv014 - 1/22/13 at 9:18am
post #30 of 43
Quote:
Originally Posted by kn19h7 View Post

Quote:
Originally Posted by Jaywalk3r View Post

No, that's not correct.


That the quantity of data is reduced does not imply that sound is missing, i.e., no longer producing SPL. At least for audible frequencies, much of the "missing" data is estimated and replaced during playback. Compared to the original data, the estimations inevitably have errors, we should be careful not to assume that the error always results in the lossy passage being quieter than the original. It's just different. If we wish to determine whether or not that difference is audible, we must level match for our ABX tests.

Since no volume adjustment involved in encoding, shouldn't that the "error" to be the artifact of lossy codec?

As you've said, the replaced missing data is of estimated thus may be louder and quieter here and there, I don't really think there is any appropriate/practical way to do proper level matching...

Proper level matching matches the total volume, not the volume for each of many intervals of the audible frequency range. The whole point of the ABX test is to determine if the tracks are blindly audibly distinguishable after that level matching. It doesn't matter if one track is slightly louder at frequency A and slightly quieter at frequency B, since we don't care which track is preferred. We want to know if those tiny differences are audible.
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Computer Audio
Head-Fi.org › Forums › Equipment Forums › Computer Audio › Reasons I don't hear a difference between 128kbps, 256kbps, and ALAC??