Depressed / confused about music management
Feb 27, 2013 at 2:35 PM Post #16 of 42
Quote:
In my testing, I can definitely hear the difference between 192k and 320k MP3 for example, and I haven't done enough testing yet to see if I can hear the difference between 320k MP3 and lossless - but it seems like I probably can.

 
Frauenhofer MP3 is not at all the same as AAC. Two totally different codecs. Do the comparison with AAC 256 VBR or AAC 320 CBR and lossless. Make sure you line level match. I guarantee you that you won't hear a difference, even with the best equipment made. Your guests won't hear a difference either, assuming they are human. If you never do the test, you'll never know. I've done it.
 
By the way, the difference between higher bitrate compressed audio and lossless has nothing to do with upper frequencies. Once you get over around 128 with AAC, the frequency response is identical. The difference is only artifacting... Once you get beyond the place where artifacting happens, the sound is identical.
 
There are definite advantages to smaller file sizes for long term storage. A smaller file is less likely to suffer from random data corruption. It's easier to back up and stream over wifi. There's more to it than just hard drive space.
 
Feb 27, 2013 at 2:47 PM Post #17 of 42
Quote:
 
Frauenhofer MP3 is not at all the same as AAC. Two totally different codecs. Do the comparison with AAC 256 VBR or AAC 320 CBR and lossless. Make sure you line level match. I guarantee you that you won't hear a difference, even with the best equipment made. If you never do the test, you'll never know. I've done it.
 
By the way, the difference between higher bitrate compressed audio and lossless has nothing to do with upper frequencies. Once you get over around 128 with AAC, the frequency response is identical. The difference is only artifacting... Once you get beyond the place where artifacting happens, the sound is identical.

I don't think that's true, but I'll have to do some testing to confirm now that I got a spectogram program to play around with. But I don't see why aac wouldn't at the very least put a 20kHz lowpass filter.
 
The most artifacting does happen in the high frequencies though(>16 or so kHz).
 
Feb 27, 2013 at 2:49 PM Post #18 of 42
I hear ya. (Get it?) But if I used you're logic, I'd have to go back to my vinyl collection (Which I foolishly no longer have.) and rip from there. Back in the day when CDs first came out all the talk amongst audiophiles was how much worse CDs sounded than records. :wink:
 
Feb 27, 2013 at 5:19 PM Post #19 of 42
CDs have better sound quality than records, and high bitrate AAC sounds just as good as CD.
 
Feb 27, 2013 at 5:35 PM Post #21 of 42
Quote:
CDs have better sound quality than records

They can have better SQ than records, yes....but not all of them do, because of the loudness war production techniques.
 
Feb 27, 2013 at 5:39 PM Post #23 of 42
According to a test I read, AAC128 rolls off high frequencies above 16kHz. But AAC 320 reproduces all of the frequencies in CDs, even the ones above 20kHz. The trade off, is a slightly higher noise floor (-90dB) above 18kHz, but below 16kHz, the noise floor drops back down to -110dB.
 
High bitrate AAC doesn't look precisely the same on the scope, but for human ears, it has a full frequency response and it's audibly transparent.
 
Feb 27, 2013 at 5:39 PM Post #24 of 42
Feb 27, 2013 at 5:43 PM Post #25 of 42
Quote:
Very much disagree. Analog has a much more "coherent" sound when it comes to acoustic instrumentation and voices.

 
Again, without doing the test, you can't know that...
 
Take the best sounding LP you own and digitize it and burn it to a CD, and as long as your sound card is capable of full CD quality, it will sound indistinguishable from the original LP. I did this using Lincoln Mayorga and Distinguished Colleagues Vol 2... one of the best sounding LPs ever made... Doug Sax mastered direct to disk by Sheffield Lab. On my best system it sounded exactly the same as the CD rip.
 
I'm referring to the capabilities of the format itself, Achmedisdead, not the mastering.
 
Feb 27, 2013 at 5:51 PM Post #26 of 42
Quote:
 
Again, without doing the test, you can't know that...
 
Take the best sounding LP you own and digitize it and burn it to a CD, and as long as your sound card is capable of full CD quality, it will sound indistinguishable from the original LP. I did this using Lincoln Mayorga and Distinguished Colleagues Vol 2... one of the best sounding LPs ever made... Doug Sax mastered direct to disk by Sheffield Lab. On my best system it sounded exactly the same as the CD rip.
 
I'm referring to the capabilities of the format itself, Achmedisdead, not the mastering.

Ah. I was talking about a little different test. Back in the day when CDs first came out, my friends and I frequently A/B'd the same album on CD and vinyl. There was near consensus that analog was better. I made the transition largely based upon convenience and not sound quality. Granted, CD quality may be much better than it was back then.
 
Feb 27, 2013 at 5:55 PM Post #27 of 42
Quote:
They can have better SQ than records, yes....but not all of them do, because of the loudness war production techniques.

Back in ancient times, the only thing digital had over analog when it came to SQ was dynamic range. Is that what you mean by "loudness war production techniques?"
 
Feb 27, 2013 at 5:59 PM Post #28 of 42
Quote:
According to a test I read, AAC128 rolls off high frequencies above 16kHz. But AAC 320 reproduces all of the frequencies in CDs, even the ones above 20kHz. The trade off, is a slightly higher noise floor (-90dB) above 18kHz, but below 16kHz, the noise floor drops back down to -110dB.
 
High bitrate AAC doesn't look precisely the same on the scope, but for human ears, it has a full frequency response and it's audibly transparent.

I followed up with testing this.
 
 
 
 
 

 
Top left: 320kbps aac
Top right: 256
Bottom left: 192
Bottom right: 128
 
You were correct about there being no explicit lowpass filters until 128. 
 
But on 256 and 192 while there is no hard lowpass filter, there might as well be(at least on most material, if something extends all the way up it's not going to destroy it here). There's only a tiny bit of data left there.
 
Feb 27, 2013 at 6:06 PM Post #29 of 42
Quote:
I followed up with testing this.
 
 
 
 
 

 
Top left: 320kbps aac
Top right: 256
Bottom left: 192
Bottom right: 128
 
You were correct about there being no explicit lowpass filters until 128. 
 
But on 256 and 192 while there is no hard lowpass filter, there might as well be. There's only a tiny bit of data left there.

Not much of a scientific/analytic type but if I read the chart correctly 320 is a DRAMATIC overall improvement, right? Or is that just a volume difference? Do you have the same kind of chart for lossless?
 
Feb 27, 2013 at 6:20 PM Post #30 of 42
Quote:
Not much of a scientific/analytic type but if I read the chart correctly 320 is a DRAMATIC overall improvement, right? Or is that just a volume difference? Do you have the same kind of chart for lossless?

If you look at the frequencies that are getting eaten by the compression, and the low volume they're getting played at relative to the volume of the other frequencies, no it is not dramatic. 2kHz of those frequencies aren't even audible by the vast majority of people, except for babies. And remember dark blue is a ~40dB difference from green.The difference is likely not even audible  between 256 and 320 for most people.
 
I find spectograms can look really ugly in comparison but in reality not have that much audible difference. If you check out the thread in my signature about transcoding, I provide some ugly spectograms as a result of high bitrate transcoding and have the corresponding files up for download to see what they sound like. 
 
Here's the lossless version though:
 
 

 
 
 
The only obvious difference you can see between it and 320kbps is that the top is nipped at a little bit.
 

Users who are viewing this thread

Back
Top