24bit vs 16bit, the myth exploded!

[U2Bono269]

Why do you care so much? I didn't know I had to agree with you in order to participate.

I think I've gotten what I can out of this thread and it's time for me to move on to other topics. Thanks to those who helped me!


Sent from my iPad using Tapatalk

 

[lamode]

  A MP3 or AAC release would have different masters,

 
Why do you think so? I would be surprised if they bothered to remaster for MP3 at all compared to CD. What would be the point?
 
Quote:

16 bit has a noise level -96dB while 24bit give about 24dB more so the noise is roughly -120dB down, which is limited by physics to get much better. An extremely silent studio will be about 20dB SPL you would playing the music at an ear splitting116dB SPL before you to start to hear the noise floor of the 16bit file.

 
24dB has a noise level around -144dB, which is a 48dB difference.

 

[bigshot]

Why do you care so much? I didn't know I had to agree with you in order to participate.

 
It's not that we have to agree... it's that if you can prove that you can successfully and consistently discern a difference between say MP3 LAME 320 VBR or AAC 256 VBR, you will have done something no other human being has been able to do before under controlled testing. That is VERY interesting to those of us who are interested in compressed audio codecs and how to use them. Either your hearing is beyond normal human's ability, or perhaps your comparison wasn't very well controlled and expectation bias crept in making your test results invalid. We're asking the questions to determine which is the case.

 

[RRod]

   
24dB has a noise level around -144dB, which is a 48dB difference.

 
I think he means that the actual effective bits out of an ADC/DAC never gets to 24 in practice. 20—21 is the best I've seen actually tested results for.

 

[RRod]

   
It's not that we have to agree... it's that if you can prove that you can successfully and consistently discern a difference between say MP3 LAME 320 VBR or AAC 256 VBR, you will have done something no other human being has been able to do before under controlled testing. That is VERY interesting to those of us who are interested in compressed audio codecs and how to use them. Either your hearing is beyond normal human's ability, or perhaps your comparison wasn't very well controlled and expectation bias crept in making your test results invalid. We're asking the questions to determine which is the case.

 
I thought he already admitted he can't tell a blind difference. Isn't his main question whether 24bits is a "mark" signifying better mastering? Hopefully we've argued convincingly otherwise, but perhaps an example of a CD with superior mastering to an HD release might help answer his question better.

 

[lamode]

 
Thank you. Has anyone done an ABX in this sort of scenario?

 
There are tests comparing a 16/44 AD-DA loop to a live feed, and no-one could hear the difference. Given that the live feed is the original, it is by definition better than any possible digitized version, regardless of resolution, this makes the question about HD audio moot.

 

[lamode]

 To my ears lossless sounds better.

 
If that were really true, you could easily pass an ABX test.

 

[U2Bono269]

To be fair I only did the mp3/flac test once. I think I got 8 out of 10. But when I listen to stuff I like the sound of the flac better. Sounds fuller to me. I can't tell 100% of the time. For example for fighters stuff sounds the same. But for the most part it sounds different to me. I never said I could tell 100% of the time.

My question was indeed about the mastering of 24 bit files. I think it's been answered pretty well. I don't expect to hear any differences or better mastering with 24 bits.


Sent from my iPad using Tapatalk

 

[dprimary]

  A MP3 or AAC release would have different masters,

 
Why do you think so? I would be surprised if they bothered to remaster for MP3 at all compared to CD. What would be the point?
 
Quote:

16 bit has a noise level -96dB while 24bit give about 24dB more so the noise is roughly -120dB down, which is limited by physics to get much better. An extremely silent studio will be about 20dB SPL you would playing the music at an ear splitting116dB SPL before you to start to hear the noise floor of the 16bit file.

 
24dB has a noise level around -144dB, which is a 48dB difference.

Read the mastering for iTunes part of the page
http://www.justmastering.com/article-masteredforitunes.php
 
the point would be to deliver the best sound you can for each format. Believe it or not huge amounts of effort went into getting what quality they could into high speed duplicated cassettes in the 80's
Even if I did not use a certified  mastered for iTunes mastering engineer I would still check the master myself before sending to apple.
 
A -144dB signal to noise ratio converter cannot be made. Most good 24bit converter are around -120dB as I said before physics limits what we can build. The thermal noise of one resistor - around -135dB there is a bit more then one resistor in a DAC. An Audio Precision APx555 has a THD+N of -120dB which is about as good as it gets.

 

 

[lamode]

  Read the mastering for iTunes part of the page
http://www.justmastering.com/article-masteredforitunes.php
 

 
There was zero information on that page, just marketing fluff.
 
And there is a difference between transcoding and creating a new master.

 

[Greenears]

There are tests comparing a 16/44 AD-DA loop to a live feed, and no-one could hear the difference. Given that the live feed is the original, it is by definition better than any possible digitized version, regardless of resolution, this makes the question about HD audio moot.

I an aware of tests with 16 bit loop off a recorded feed of various formats, but not live. That would indeed settle it. Link please to the full testing methodology.

 

[bigshot]

I don't expect to hear any differences or better mastering with 24 bits.

 
You might hear an improvement, but it has nothing to do with the file being 24 bit. Formats don't matter. The quality of the recording, mixing and mastering does. That depends on the skill, budget and most of all taste of the engineers working on the record.

 

[lamode]

I an aware of tests with 16 bit loop off a recorded feed of various formats, but not live. That would indeed settle it. Link please to the full testing methodology.

 
I tried googling to find the study I saw a while back but could only find references to the study (without links), and not the study itself. If I find it I will post it (thought I have a feeling I already have )

 

[dprimary]

   
There was zero information on that page, just marketing fluff.
 
And there is a difference between transcoding and creating a new master.

 
It gets into some of the requirements that apple has for iTunes. They do expect the mastering engineer to confirm it will encode properly.  
Every mastering studio tells me the same thing iTunes is a separate master. If you send in a CD master it will likely clip and apple will reject it. Apple does provide all the tools to do this, anyone can download them.
 
Different production masters is not new it was done this way for decades. Just for about 10 years the only format was CD so multiple production masters was forgotten about. You always have wanted to maximize the quality of every format. Ok some are in it for the glamor. Then they need to eat.

 

[bigshot]

When you transcode from Redbook to AAC, the volume bumps up a hair. So you normalize down to 90% and then transcode. It isn't an audible difference. Just a slightly lower volume level.