Quantcast
Will dithering be obsolete in the next few years? - Gearspace.com
The No.1 Website for Pro Audio
Will dithering be obsolete in the next few years?
Old 22nd January 2013
  #1
Gear Addict
 
jwleeman's Avatar
 
🎧 10 years
Will dithering be obsolete in the next few years?

Given that optical media may be a thing of the past pretty soon (for music at least), would that imply that dithering would go out of style too? Here's hopin'!
Old 22nd January 2013
  #2
Gear Addict
 
Nerine's Avatar
 
🎧 10 years
Without sounding harsh, just adding a dither plugin to a master buss is not a particularly hard or taxing thing to do.
I hope physical forms of music don't die out. I'd hate not to be able to buy a CD.
I'd hate to not be able to listen to a record.
Cassettes can do one though. They sound pretty toilet.
Old 22nd January 2013
  #3
Moderator
 
psycho_monkey's Avatar
 
🎧 10 years
Quote:
Originally Posted by jwleeman ➡️
Given that optical media may be a thing of the past pretty soon (for music at least), would that imply that dithering would go out of style too? Here's hopin'!
what's wrong with dithering? properly done, it's not audible...that's the point of it!

Mp3s are still generally 16bit streams as well....
Old 22nd January 2013
  #4
Lives for gear
 
🎧 10 years
People really don't understand dithering.

It's got nothing to do with optical media, it's got to do with quantizers, which are unavoidable in digital audio. Whether you're quantizing to 16 bit fixed, 24 bit fixed, 32 bit float, whatever, you will be quantizing to something, probably multiple times... e.g. when you sample the analogue signal that's quantization, when you apply a gain change that's a multiplication followed by a quantization.

When you're talking about fixed point, dithered quantizers give a more reliably accurate result than undithered, in other words, they're better, simple as that .It may seem counter intuitive since you're adding noise, but what you're doing is exchanging a tiny increase in the average noise level (taken over a typical track) for relatively large distortion in certain signal conditions.

Increasing the word length reduces noise and distortion, so the benefits of dithering become less audible, but conversely they also require less noise to be added for dithering, so there's no benefit to dropping it.

Quite frankly, I would say that if in a proper blind A/B test you find yourself preferring the undithered quantizer to the dithered one then one of two things is happening
a) You've got some settings wrong or there's a problem with the dithering
b) You just happen to like the distortion that occurs in the places where dithering has the most effect.

Saying that, at 24 bits I would not be surprised if in many cases you couldn't even tell the difference in a double blind A/B/X test.

I've yet to encounter a proven dithering scheme for floating point, so it's a moot point when you're working in that format.
Old 22nd January 2013 | Show parent
  #5
Deleted 6ccb844
Guest
Quote:
Originally Posted by Jon Hodgson ➡️
People really don't understand dithering.

It's got nothing to do with optical media, it's got to do with quantizers, which are unavoidable in digital audio. Whether you're quantizing to 16 bit fixed, 24 bit fixed, 32 bit float, whatever, you will be quantizing to something, probably multiple times... e.g. when you sample the analogue signal that's quantization, when you apply a gain change that's a multiplication followed by a quantization.

When you're talking about fixed point, dithered quantizers give a more reliably accurate result than undithered, in other words, they're better, simple as that .It may seem counter intuitive since you're adding noise, but what you're doing is exchanging a tiny increase in the average noise level (taken over a typical track) for relatively large distortion in certain signal conditions.

Increasing the word length reduces noise and distortion, so the benefits of dithering become less audible, but conversely they also require less noise to be added for dithering, so there's no benefit to dropping it.

Quite frankly, I would say that if in a proper blind A/B test you find yourself preferring the undithered quantizer to the dithered one then one of two things is happening
a) You've got some settings wrong or there's a problem with the dithering
b) You just happen to like the distortion that occurs in the places where dithering has the most effect.

Saying that, at 24 bits I would not be surprised if in many cases you couldn't even tell the difference in a double blind A/B/X test.

I've yet to encounter a proven dithering scheme for floating point, so it's a moot point when you're working in that format.
+1

Simply put, analogue = noise, digital = none = Quantisation error. Each bit decreases audible noise floor "Granulation noise" by 6dB.. SNR is a subset of dynamic range. Noise shaping and / or dither applied to create analogue noise because it sounds better than the crap digital comes out with (Distortion).

No dithering is going no where for a while..

A) Because the actual recording ability of even modern converters is not 140dB even A-weighted.
B) We still use .Mp3
C) There is a lot of cheap crap AD / DA out there for personal media systems.
📝 Reply
Topic:
Post Reply

Welcome to the Gearspace 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…

Forum Jump
Forum Jump