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
Dithering to 24 bit question (64 bit plugins/DAW) Dynamics Plugins
Old 6th June 2016
  #61
Gear Maniac
 
Dina Mastering's Avatar
 

Quote:
Originally Posted by Jon Hodgson View Post
Typically you're not really adding noise to the final product, but rather changing its colour.

Quantization introduces errors, which depending on the signal being quantized can be more or less correlated, the more correlated it is, the more like distortion it sounds (if it's loud enough to hear at all), the less correlated (i.e more random) it is, the more it is just like white noise.

While the implementation of a dithered quantizer involves injecting some noise, this should be viewed in the context of a scenario where noise is being introduced anyway, the net result is that it's more useful to view the complete process as an smoothing out of what you would otherwise get.

That surprises me not at all.

Either something was wrong, or cognitive biases came into play. The worst case scenario for dither is silence, where it would result in about a half bit's worth of broadband noise... in other words white noise at -141dB or so. No way that's going to make a 16 bit quantization (which itself will introduce noise at about -93dB) sound worse.
Interesting, thanks. I'll probably make more tests with different files when I have the time.

To me it was quite clearer that the "24-dithered to 16" sounded subtly worse than the "24-undithered to 16", than any difference in between the 24 dithered and undithered. Repeated blind AB. That's why I'll keep not dithereing the 24 files for now until possible further tests. But that's just me with my setup (and actually with that particular mix for now).
Old 6th June 2016
  #62
Lives for gear
 
JP__'s Avatar
 

Verified Member
Its a common phenomenon that even subtle distortion added to an audio file is often declared as "better sounding". It can sound more upfront and intimidate e.g. while a cleaner file sounds more set back and maybe somewhat boring.
But after all its just an artefact which as a particular effect is better done with a dedicated process than with a somewhat random error.

In a world full of little errors its not that easy to learn what "clean" really sounds like.
Old 6th June 2016
  #63
Gear Guru
 
UnderTow's Avatar
 

Verified Member
Quote:
Originally Posted by Bob Olhsson View Post
Lack of dithering is actually far more audible to people having hearing damage!
I guess this is for the same reason that people with hearing damage can be more sensitive to MP3s or other lossy formats artefacts (masking issues. Or rather lack of masking) but it is the first time I hear anyone say this about quantization distortion. Do you have any references?

Alistair
Old 6th June 2016
  #64
Gear Maniac
 
Dina Mastering's Avatar
 

Quote:
Originally Posted by JP__ View Post
Its a common phenomenon that even subtle distortion added to an audio file is often declared as "better sounding". It can sound more upfront and intimidate e.g. while a cleaner file sounds more set back and maybe somewhat boring.
But after all its just an artefact which as a particular effect is better done with a dedicated process than with a somewhat random error.

In a world full of little errors its not that easy to learn what "clean" really sounds like.
Totally agree, nicely put.

I count on that when mastering. What is people doing when clipping, limiting, or even passing thru analog gear to improve something if not distorting?
Old 6th June 2016
  #65
Lives for gear
 
JP__'s Avatar
 

Verified Member
It is. But should we add even some more then by accident? Theres distortion and distortion and distortion. I prefer to select it carefully.
Adding dither (even "flat") will for sure do not that much harm than leave it off in my experiences.

But I would simply trust my ears before theory here. If you like what you hear, leave it like that. For me it was just an ear opener to get rid of a lot of artefacts. But its also a kind of learning experience (especially for us "digital kids"?)
Old 6th June 2016
  #66
Gear Maniac
 
Dina Mastering's Avatar
 

FWIW I heard the files again after what you said about the distorion. The '16bit from 24 undithered' sounded closer to the 'direct 16 dithered render' to me than the '16bit from 24 dithered'.

So I guess is safe to say that it's not that it sounded more distorted and "better" than the original, it was the other one that sounded a touch duller.

Again, the difference is so marginal that's actually not worth caring too much about as reading this thread may suggest.

But in my case I still think I choose what I see as the "minor possible final damage" for my setup, working mix, etc. And yes, to each it's own / ymmv
Old 7th June 2016
  #67
Lives for gear
 

Verified Member
Quote:
Originally Posted by Dina Mastering View Post
Interesting, thanks. I'll probably make more tests with different files when I have the time.

To me it was quite clearer that the "24-dithered to 16" sounded subtly worse than the "24-undithered to 16", than any difference in between the 24 dithered and undithered. Repeated blind AB. That's why I'll keep not dithereing the 24 files for now until possible further tests. But that's just me with my setup (and actually with that particular mix for now).
Personally I'd be quite shocked if you heard any difference in the 24 bit dithered versus undithered at ordinary listening levels with no further processing.

What dither does is change the nature of the error signal, but that error signal has a peak level of +/- 1 lsb, which puts it at -138dB FS.

Now, unless our sources are all digitally generated, we're going to have a signal noise floor that is way higher than that, and anyway, even if our signal is totally pure bear in mind that the difference between the pain threshold and the threshold of hearing is about 120dB.

In other words, if your levels are set such that a full scale signal actually HURTS, then the error in a 24 bit quantization is going to be 18dB below what you can hear on its own (let alone with the masking caused by the actual audio content).

So, unless you then do something that brings that error component up in level such as it is audible, you're not going to hear if it's the "wrong" one (undithered quantizer) or the "right" one (dithered).
Old 7th June 2016
  #68
Here for the gear
Quote:
Originally Posted by Dina Mastering View Post
FWIW I heard the files again after what you said about the distorion. The '16bit from 24 undithered' sounded closer to the 'direct 16 dithered render' to me than the '16bit from 24 dithered'.

So I guess is safe to say that it's not that it sounded more distorted and "better" than the original, it was the other one that sounded a touch duller.

Again, the difference is so marginal that's actually not worth caring too much about as reading this thread may suggest.

But in my case I still think I choose what I see as the "minor possible final damage" for my setup, working mix, etc. And yes, to each it's own / ymmv
Pls permit me to ask this question...
Does dithering affect the sizÉ of Þhe dithered file?
Old 7th June 2016
  #69
Lives for gear
 

Verified Member
Quote:
Originally Posted by DONCOLLIO View Post
Pls permit me to ask this question...
Does dithering affect the sizÉ of Þhe dithered file?
No
Old 7th June 2016
  #70
Lives for gear
 
JP__'s Avatar
 

Verified Member
Quote:
Originally Posted by Dina Mastering View Post
FWIW I heard the files again after what you said about the distorion. The '16bit from 24 undithered' sounded closer to the 'direct 16 dithered render' to me than the '16bit from 24 dithered'.

So I guess is safe to say that it's not that it sounded more distorted and "better" than the original, it was the other one that sounded a touch duller.

Again, the difference is so marginal that's actually not worth caring too much about as reading this thread may suggest.

But in my case I still think I choose what I see as the "minor possible final damage" for my setup, working mix, etc. And yes, to each it's own / ymmv
"Sounding duller" could be another phenomenon of lacking these artefacts as the distortion is mostly obvious in the upper mid/lower high area.
How many tests do you have done? In my experience this is highly music depended too. E.g. I have found that some rock stuff with high gained gtrs could be very sensitive here.

Yes, of course the diff is often quite subtle, but I never found an undithered version better soundwise.
Old 7th June 2016
  #71
Gear Maniac
 
Dina Mastering's Avatar
 

Quote:
Originally Posted by Jon Hodgson View Post
.
Good, but because the possible side effect was audible in my test at normal listening level, I care much about it than what will be heard at painful levels.

Quote:
Originally Posted by JP__ View Post
.
I have only tested once with one mix. I'll do more test for sure. I find this an interesting subject and I learnt some theory here and the thread made me do this test I liked to do, and more to come. What I'm really curious to test now is how much different dithering the case will be or not with a 16files-16project-16render(64 bit daw). But anything will have to wait to the next week (short trip starting today).

Regards to all.
Old 7th June 2016
  #72
Gear Guru
 
UnderTow's Avatar
 

Verified Member
Quote:
Originally Posted by Dina Mastering View Post
I don't work with numbers, I work with sound.
But your DAW works with numbers. Not Voodoo.

Quote:
That's exaclty the oposite idea to the one that allowed humanity to evolve.
Humanity did not evolve by ignoring facts.

Quote:
I respect people's reasons to do that and never said I don't or what people should do for sure, but If somebody doesn't respect my decision to do not, specially when I said it's based on my own conclusions from my own practical test with my own setup (so "what works for me") that's certainly not my problem. But hey, best wishes anyway.
To be clear, I am not suggesting that you shouldn't do whatever you want to do but you wrote that adding dither caused more damage than not dithering. That is incorrect and that is not up to personal opinion. That is maths. That was my point.

Alistair
Old 7th June 2016
  #73
Gear Maniac
 
Dina Mastering's Avatar
 

Quote:
Originally Posted by UnderTow View Post
.
I never talked about dithering in general. I always dither when reducing wordlenght (considering file / project / output bit depth) like most people do. I was talking about the 24 / 24 / 24 (64bit daw) case in particular (wich I understood was the main topic of this thread). And always said in my own setup / what works for me / ymmv.

Have a nice week everybody!
Old 7th June 2016
  #74
Lives for gear
 

Verified Member
Quote:
Originally Posted by Dina Mastering View Post
Good, but because the possible side effect was audible in my test at normal listening level, I care much about it than what will be heard at painful levels.
Either something is wrong with that dither, or it was cognitive errors.. to be honest, any test that isn't double blind is useless in this situation (unless the dither is fubar).

How do I say that so confidently? well let's think about it for a moment..

The difference between the dithered quantize and the undithered one (assuming no noise shaping is also turned on), is that before quantizing a noise signal with an RMS of about -141dB is added to the signal.

The only effect it has is to raise or lower individual samples by less than one EIGHT MILLIONTH of full scale.. and therefore to change a component of the signal (the error introduced by quantizing) at that level.

Now even if you didn't like what it was doing (hey, some people like quantization distortion, that's why you get bitcrushers), it's still, as I already said, below your audible threshold.

And if your next step is a conversion to 16 bit.. that difference will NOT increase to be audible, it will either be lost completely (without dither), or with a dithered quantize to 16 bits be preserved to some degree (at the same level) under the new noise floor which is 48dB higher... so still irrelevant.
Old 7th June 2016
  #75
Lives for gear
 

not a huge deal, but always dither from 64 or 32bit to 24. not doing so sounds slightly worse. i've heard the effects (which have exactly nothing to do with listening level) on a number of different systems, in different rooms, over many years. not really a difficult listening test. going to 16bit will kill the mini benefit though.
Old 7th June 2016
  #76
Lives for gear
 

Verified Member
Quote:
Originally Posted by Timesaver800W View Post
not a huge deal, but always dither from 64 or 32bit to 24. not doing so sounds slightly worse. i've heard the effects (which have exactly nothing to do with listening level) on a number of different systems, in different rooms, over many years. not really a difficult listening test. going to 16bit will kill the mini benefit though.
Detecting a difference which if were isolated would actually be smaller than the amount needed to even move the sensory hairs in your ear isn't a difficult listening test?
Old 7th June 2016
  #77
Lives for gear
 

obviously, it depends on monitoring and skill level. you don't isolate the effect though, it's an integral part of the signal. the da acts as a distortion multiplier. but you certainly need fresh tweeters.
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
plus6vu / Mastering forum
41
JAY55 / Mastering forum
20
hackenslash / Mastering forum
4
sound_forward / So much gear, so little time
4

Forum Jump
Forum Jump