The No.1 Website for Pro Audio
 Search This Thread  Search This Forum  Search Reviews  Search Gear Database  Search Gear for sale  Search Gearslutz Go Advanced
Experts please explain Equalisers (HW)
Old 12th September 2007
  #1
Lives for gear
 
tamasdragon's Avatar
 

Experts please explain

I recently purchased a protools le with digi003. Only one thing what is not clear for me. The dither (pow-r) is limited only to 16 and 20bit. What if I would like to save and dither 24bit files?
Please teach me.
Thanks Tamas Dragon
Old 12th September 2007
  #2
Lives for gear
 

If you are recording at 24 bits (and you definately should) then there is no need for bit depth reduction in order to save 24 bit files. Therefore there is no reason to add dither noise.

Dithering shouldn't be confused with sample rate conversion, or bit depth reduction/truncating/appending. (Although it seems to be a fairly common mis-use of the word to apply it to the whole bunch of things).

Dither - Wikipedia, the free encyclopedia
Old 12th September 2007
  #3
Lives for gear
 
tamasdragon's Avatar
 

Thanks. The confusion is in my own head of course. But I always saw softwares says dither to 24 bit, and now in the pow-r , there is no 24bits dither.
But you're right if I think it over. Thank you.
Tamas Dragon
Old 12th September 2007
  #4
Mastering
 

Quote:
Originally Posted by tamasdragon View Post
Thanks. The confusion is in my own head of course. But I always saw softwares says dither to 24 bit, and now in the pow-r , there is no 24bits dither.
But you're right if I think it over. Thank you.
Tamas Dragon
In theory, you should dither down to 24 bits when working in any native program, as there is a potential loss (truncation) when going from 32 bit float (for example) down to 24. But the audibility of this loss is so little that in practicality we can argue about the theory all day long but at the end of the day, the loss is very minor, usually inaudible. However, like chicken soup, "it couldn't hurt", and if you have some fairly aggressive and agregious processes, 2 or 3 in a row, each one truncated to 24 bits, you may notice the beginning of some harshness in the sound which could have been prevented if you had dithered each one to 24. If working in a Native DAW entirely, then this is a moot point, because each process will communicate with the next at 32 bits float, (or 64, these days!) but if you are feeding an external processor, then it couldn't hurt, and it might help, to dither down to 24.

The Weiss processors used to have 24 bit dither built in, but when they converted to powr dither they forgot and they currently now truncate internally when going from 40 bit float down to 24 to feed the outside world. It remains to be seen (heard) how much improvement (if any) there will be when they restore the 24 bit dither to the Weiss. I plan on doing some blind comparisons... Daniel will get around to restoring the dither some day soon.
Old 12th September 2007
  #5
Dear Bob,how are you!

Do you recommend to use the digidesign dither on Protools HD Accell when you work on 24 bit?
Old 13th September 2007
  #6
Mastering
 

Quote:
Originally Posted by AMIEL View Post
Dear Bob,how are you!

Do you recommend to use the digidesign dither on Protools HD Accell when you work on 24 bit?
I have never bothered to turn on the digidesign dither because I don't like to have dithering going on behind the scenes unless I've examined it thoroughly and know how it's working. Since I haven't done that with the digi dither I generall, when mixing in Pro Tools, put a plugin I'm confident in, like the Waves L2 with the threshold at 0 dB, with its dither, on the master fader. Then I'm comfortable I've got just what I want where I want it, no more no less.

If I wanted to learn more about Pro Tools (which I really don't) :-) Then I would investigate the Digi dither and see how well it works... so I don't currently have a good answer for you!
Old 13th September 2007
  #7
Lives for gear
 
masteringhouse's Avatar
 

Verified Member
Quote:
Originally Posted by AMIEL View Post
Do you recommend to use the digidesign dither on Protools HD Accell when you work on 24 bit?
Try using the dithered mixer instead of the default (truncation). I've found that it sounds a bit more "organic".
Old 18th September 2007
  #8
Lives for gear
 
Alécio Costa's Avatar
 

Verified Member
Power Level1 is my main dither choice nowadays.
Although hard to judge, Level3 seems to add some slight top end, which in many cases in not a goal.
Old 18th September 2007
  #9
Lives for gear
 
Nordenstam's Avatar
 

Verified Member
Hello!

Quote:
Originally Posted by bob katz View Post
In theory, you should dither down to 24 bits when working in any native program, as there is a potential loss (truncation) when going from 32 bit float (for example) down to 24. But the audibility of this loss is so little that in practicality we can argue about the theory all day long but at the end of the day, the loss is very minor, usually inaudible.
This is something that's been puzzling me for a while. Have seen several hardware L2 users report that they prefer to turn the dither off, even at 24 bits. I've found the same to generally be true of the DBX Quantum - it sounds better with 24 bit dither turned off.

This is quite the opposite of what theory suggests should be the result. I don't even believe I can hear stuff at something like -30dB SPL..

Any idea what's going on to produce these listening results?


Regards,

Andreas Nordenstam
Old 18th September 2007
  #10
Lives for gear
 
MattGray's Avatar
 

Verified Member
Quote:
Originally Posted by bob katz View Post
The Weiss processors used to have 24 bit dither built in, but when they converted to powr dither they forgot and they currently now truncate internally when going from 40 bit float down to 24 to feed the outside world. It remains to be seen (heard) how much improvement (if any) there will be when they restore the 24 bit dither to the Weiss. I plan on doing some blind comparisons... Daniel will get around to restoring the dither some day soon.
Hi Bob, I checked with Daniel about this in May & this was his response...

Quote:
Hi Matt, I was wrong - I just checked all EQ1 variants, DS1-MK2 and DNA1 for the 24 bit level dithering. All of them do proper dithering even with the 24 bit setting. Somehow I was under the impression that we haven't taken care of that yet.


The version numbers this information is valid for are:

unit DSP OS Scales Coeff.
----------------------------------------
EQ1-MK2 4.2 7.1 4.1 -
EQ1-DYN 1.3 1.3 4.1 -
EQ1-LP 2.5 5.1 4.2 -
DS1-MK2 2.1 2.1 1.3 1.2
DNA1 3.0 3.0 3.0 -
Rest easy that it's being taken care of correctly if you have the latest OS

Matt
Old 18th September 2007
  #11
Mastering
 

Quote:
Originally Posted by MattGray View Post
Hi Bob, I checked with Daniel about this in May & this was his response...



Rest easy that it's being taken care of correctly if you have the latest OS

Matt
I'm not resting easy. Not according to my measurements. It's real easy to confirm that the Weiss is not dithering at the 24th bit level. The bitscope has to show activity in the LSB, and it does not. There are other tests, such as in my book on page 182 at the top, which was the Weiss BEFORE Daniel added POWr but not anymore....

Daniel and I are still in communication about this, however slowly :-).

BK
Old 18th September 2007
  #12
Mastering
 

Quote:
Originally Posted by Lupo View Post
Hello!



This is something that's been puzzling me for a while. Have seen several hardware L2 users report that they prefer to turn the dither off, even at 24 bits. I've found the same to generally be true of the DBX Quantum - it sounds better with 24 bit dither turned off.

This is quite the opposite of what theory suggests should be the result. I don't even believe I can hear stuff at something like -30dB SPL..

Any idea what's going on to produce these listening results?

If you can hear "quite obviously" the sonic difference between dither on or off at the 24th bit setting of the L2 it is clearly an illusion. Even the most discerning listeners require considerable time and an incredible monitoring system to hear that difference.

As to which is better.... well, at that level it's hard to even decide, but if you like it a little colder and smaller, or you think the sound is more "open", then you might hear a difference, after straining for about 15 minutes and I can't even guarantee that. It's truly an academic question, you're talking about dither at -141 dBFS level, which is 20 or so dB BELOW the noise floor of any DAC!!!!

That's why the slight argument I'm having with Daniel is really insignificant.

BK
Old 18th September 2007
  #13
Lives for gear
 
Cellotron's Avatar
 

Verified Member
Quote:
Originally Posted by bob katz View Post
If you can hear "quite obviously" the sonic difference between dither on or off at the 24th bit setting of the L2 it is clearly an illusion. Even the most discerning listeners require considerable time and an incredible monitoring system to hear that difference.
In fact you probably are hearing a bigger difference just by turning your head!
Why We Believe

Best regards,
Steve Berson
Old 21st September 2007
  #14
Lives for gear
 
Nordenstam's Avatar
 

Verified Member
Quote:
Originally Posted by bob katz View Post
If you can hear "quite obviously" the sonic difference between dither on or off at the 24th bit setting of the L2 it is clearly an illusion. Even the most discerning listeners require considerable time and an incredible monitoring system to hear that difference.

.. It's truly an academic question, you're talking about dither at -141 dBFS level, which is 20 or so dB BELOW the noise floor of any DAC!!!!
It's by no means an obvious difference, but sometimes I feel something happens. It's something I've also have seen several other people mention on the forums, that they have a preference for turning dither off at 24 bit. The human ability to hear far into the noise floor keeps amazing me and it may be something there.
An idea that came to mind now is; the truncation distortion may be easier masked by other loud noises, with the small spikes of energy being inaudible due to their extremely microscopic level, as opposed to the noise floor of dither that fills up the entire spectrum - perhaps making it easier to detect on some subtle material. That's just a quick idea though. Much likely as wrong as mentioning anectodal findings in the first place..

It's definitely bordering on religion. That's why I also wrote "I don't even believe I can hear stuff at something like -30dB SPL..". It's quite odd and quite likely belief - as Steve mentioned.


Thanks,

Andreas
Old 21st September 2007
  #15
Moderator
 
jayfrigo's Avatar
 

Verified Member
Quote:
Originally Posted by Lupo View Post
An idea that came to mind now is; the truncation distortion may be easier masked by other loud noises, with the small spikes of energy being inaudible due to their extremely microscopic level, as opposed to the noise floor of dither that fills up the entire spectrum - perhaps making it easier to detect on some subtle material. That's just a quick idea though. Much likely as wrong as mentioning anectodal findings in the first place.
If I had to venture a guess, it might be the opposite of that. Human hearing is good at looking for patterns, and we tend to be able to hear correlated things better than wideband, steady state things like dither noise (which is why we can percieve signal below the LSB in the dither noise). Not only are the peaks of the truncation distortion higher, but they are related to the signal, and I'd expect them to be more audible than the dither. However, subtle truncation distortion can sometimes be perceived as a slight brightening, or more subjectively like greater detail, when it is actually an addition of distortion. We all know that sometimes subtle (or even not-so-subtle) distortion is actually preferred by some listeners upon first impression. We love to add distortion of tubes, transformers, analog summing etc. Some people may simply like the sound of truncation distortion in their L2. I'd rather get my pleasing distortions elsewhere.
Old 21st September 2007
  #16
Mastering
 

Quote:
Originally Posted by jayfrigo View Post
If I had to venture a guess, it might be the opposite of that. Human hearing is good at looking for patterns, and we tend to be able to hear correlated things better than wideband, steady state things like dither noise (which is why we can percieve signal below the LSB in the dither noise). Not only are the peaks of the truncation distortion higher, but they are related to the signal, and I'd expect them to be more audible than the dither. However, subtle truncation distortion can sometimes be perceived as a slight brightening, or more subjectively like greater detail, when it is actually an addition of distortion. We all know that sometimes subtle (or even not-so-subtle) distortion is actually preferred by some listeners upon first impression. We love to add distortion of tubes, transformers, analog summing etc. Some people may simply like the sound of truncation distortion in their L2. I'd rather get my pleasing distortions elsewhere.
Jay is right. You are likely hearing the results of subtle truncation distortion (which may brighten the sound) rather than the result of the dither itself. The truncation distortion produces products which are much higher in amplitude and at more nasty frequencies. Nevertheless, I reiterate that it is subtle. The distortion from a single truncation at 24 bits is still likely only 6 to 10 dB above the -144 mark and very hard to hear under the noise floor of a DAC whose noise is -120 dBFS. I've done the listening again and again, and most of the apparent differences with 24 bit dither are just illusion. 24 bits is VERY far down!
Post Reply

Welcome to the Gearslutz Pro Audio Community!

Registration benefits include:
  • The ability to reply to and create new discussions
  • Access to members-only giveaways & competitions
  • Interact with VIP industry experts in our guest Q&As
  • Access to members-only sub forum discussions
  • Access to members-only Chat Room
  • Get INSTANT ACCESS to the world's best private pro audio Classifieds for only USD $20/year
  • Promote your eBay auctions and Reverb.com listings for free
  • Remove this message!
You need an account to post a reply. Create a username and password below and an account will be created and your post entered.


 
 
Slide to join now Processing…
Thread Tools
Search this Thread
Search this Thread:

Advanced Search
Similar Threads
Thread
Thread Starter / Forum
Replies
Trailerman / Music Computers
12
mike kable / Music Computers
12
kjcoral / High end
4
Frequency / So much gear, so little time
9
XHipHop / Music Computers
8

Forum Jump
Forum Jump