Reply by Steve Underwood April 5, 20062006-04-05
Jerry Avins wrote:
> Randy Yates wrote: > >> gans1973@rediffmail.com writes: >> >> >>> hi, >>> I had certain querries about adding dither in audio signals. >>> 1) I presume one of the reasons to add it is to compensate for >>> quantisation error while >>> converting analog to digital. Is there any other reason for adding >>> it? >> >> >> >> No. > > > How about converting digital of one amplitude to another (gain change)?
Think for a while, and you can probably fill an entire page with good places to dither. Almost any form of filtering can make good use of it, depending on the circumstances.
>>> 2) Is it added only when the signal strength is zero, or would it be >>> advantageous to add it >>> always. If quantization error is the primary consideration, I >>> presume it would be >>> advantageous to add it always. But it seems that there are some >>> systems, >>> where its added only when the signal strength is zero for >>> sufficient period of time. >> >> >> >> Keep it on all the time. To go back and forth between dither/no dither >> is to modulate your noise floor, and that's not a good thing. > > > OT: It's snowing heavily in Central NJ right now. The grass is covered, > but not thee pavement (yet).
Its 30C here, and 90%+ humidity. :-) Steve
Reply by robert bristow-johnson April 5, 20062006-04-05
Jerry Avins wrote:

> The snow stopped after about a quarter inch on the lawn with blades of > grass poking through. In never stuck on pavement, and melted off the > cars first when the sun came out. It's all gone now.
this morning (at 6:30) i had to scrape my windshield.
> I'm glad to hear from you, Is the stress test over?
yeah, i posted a quikie in the other thread (where i first mentioned it). the results are "normal" (which was not the case two years ago for my first stress test) so i was not "taken upstairs" in a wheelchair. we'll let the cardiologist worry about the symptoms (unexplained occasional atypical upper chest pain and "tightness"), but at least i think i can jog, ride bike, or climb the Ver Monts without expecting i'll keel over. r b-j
Reply by Rune Allnor April 5, 20062006-04-05
Jerry Avins skrev:
> The snow stopped after about a quarter inch on the lawn with blades of > grass poking through. In never stuck on pavement, and melted off the > cars first when the sun came out. It's all gone now.
Heh, seems like the "reindeer calf spasm" is international... Over here, there is always a "backlash" of winter between late March and mid April. The oldtimers around here say it is connected to the reindeers giving birth; the exact nature of that connection being somewhat dodgy. Still, knowing that such a "spasm" will happen is crucial when tuning in for spring. The latest spasm I can remember, is snow falling on May 17th. I remember seing pictures of snow falling on June 10th, but that picture was taken some 40 years ago. Ah, I almost forgot: Today I saw my second White-throated Dipper (cinclus cinclus). I have seen one before, when I drove along some river in a remote valley. Today it was some 500m from home, diving off the ice shelf into some open water. Very fun, even though that particular critter doesn't migrate. Rune
Reply by Andor April 5, 20062006-04-05
Jerry Avins wrote:
...
> OT: It's snowing heavily in Central NJ right now. The grass is covered, > but not thee pavement (yet).
Yeah, it's odd. Yesterday was sunny and warm, today it is snowing large chunky flakes right into Zurich. My car was covered in snow when I wanted to go home from work this evening.
Reply by Jerry Avins April 5, 20062006-04-05
robert bristow-johnson wrote:
> Jerry Avins wrote: > ... > >>>Keep it on all the time. To go back and forth between dither/no dither >>>is to modulate your noise floor, and that's not a good thing. >> >>OT: It's snowing heavily in Central NJ right now. The grass is covered, >>but not thee pavement (yet). > > > it snowed here yesterday. (after a couple of days with highs in the > 60s or 70s.) in the mtns, 15-20 cm. just a dusting here by the lake. > > nice and sunny today though. > > oh, about dither, i might not add any before the A/D since most modern > A/D's are sigma-delta and most of them have a self-dithering property > to them (to make sure idle tones are suppressed). also modern audio > A/D's with "24 bits" of dynamic range, would likely get sufficiently > dithered just from the analog op-amp noise and such in the electronics > preceding the A/D. > > do not add digital dither to the signal coming out of the A/D, i don't > see how that necessarily gains you anything. however, anytime there is > a word width reduction (which will happen if you do *anything* to the > audio with DSP code, filtering, gain-change, SRC, anything), triangular > PDF dither of 2 LSB strength should be added to decouple first two > moments (the mean and variance) of the quantization error from the > signal. > > and any dither, when applied, should be on all the time.
The snow stopped after about a quarter inch on the lawn with blades of grass poking through. In never stuck on pavement, and melted off the cars first when the sun came out. It's all gone now. I'm glad to hear from you, Is the stress test over? Jerry -- Engineering is the art of making what you want from things you can get. �����������������������������������������������������������������������
Reply by robert bristow-johnson April 5, 20062006-04-05
Jerry Avins wrote:
...
> > > > Keep it on all the time. To go back and forth between dither/no dither > > is to modulate your noise floor, and that's not a good thing. > > OT: It's snowing heavily in Central NJ right now. The grass is covered, > but not thee pavement (yet).
it snowed here yesterday. (after a couple of days with highs in the 60s or 70s.) in the mtns, 15-20 cm. just a dusting here by the lake. nice and sunny today though. oh, about dither, i might not add any before the A/D since most modern A/D's are sigma-delta and most of them have a self-dithering property to them (to make sure idle tones are suppressed). also modern audio A/D's with "24 bits" of dynamic range, would likely get sufficiently dithered just from the analog op-amp noise and such in the electronics preceding the A/D. do not add digital dither to the signal coming out of the A/D, i don't see how that necessarily gains you anything. however, anytime there is a word width reduction (which will happen if you do *anything* to the audio with DSP code, filtering, gain-change, SRC, anything), triangular PDF dither of 2 LSB strength should be added to decouple first two moments (the mean and variance) of the quantization error from the signal. and any dither, when applied, should be on all the time. r b-j
Reply by April 5, 20062006-04-05
Doesnt applying dither during sample rate conversion or for adding more
number of
bits in an already digitized signal directly or indirectly based on
reducing quantization error??

     Secondly, if the ADC is say 40 bits (or some v. high number of
bits) would
it still make sense to dither?? At 40 bits, the normal noise in a
system may
override the quantization error, and the system noise may itself act as
dither??


ganesh

Carlos Moreno wrote:
> Steve Underwood wrote: > > >>> I had certain querries about adding dither in audio signals. > >>> 1) I presume one of the reasons to add it is to compensate for > >>> quantisation error while > >>> converting analog to digital. Is there any other reason for adding > >>> it? > >> > >> No. > > > > Eh? Wouldn't you dither when changing the number of bits of a signal > > which is already digital? > > Or when changing the sampling rate to a higher (presumably oversampled) > rate for the purpose of simplifying the D/A conversion filter. (right?) > > Carlos > --
Reply by Randy Yates April 5, 20062006-04-05
Jerry Avins <jya@ieee.org> writes:

> Randy Yates wrote: >> gans1973@rediffmail.com writes: >> >>>hi, >>> I had certain querries about adding dither in audio signals. >>>1) I presume one of the reasons to add it is to compensate for >>>quantisation error while >>> converting analog to digital. Is there any other reason for adding >>>it? >> No. > > How about converting digital of one amplitude to another (gain change)?
See my other post.
>>>2) Is it added only when the signal strength is zero, or would it be >>>advantageous to add it >>> always. If quantization error is the primary consideration, I >>>presume it would be >>> advantageous to add it always. But it seems that there are some >>>systems, >>> where its added only when the signal strength is zero for >>>sufficient period of time. >> Keep it on all the time. To go back and forth between dither/no >> dither >> is to modulate your noise floor, and that's not a good thing. > > OT: It's snowing heavily in Central NJ right now. The grass is > covered, but not thee pavement (yet).
It's a lovely Spring day, in the 70s, here in central NC. -- % Randy Yates % "My Shangri-la has gone away, fading like %% Fuquay-Varina, NC % the Beatles on 'Hey Jude'" %%% 919-577-9882 % %%%% <yates@ieee.org> % 'Shangri-La', *A New World Record*, ELO http://home.earthlink.net/~yatescr
Reply by Jerry Avins April 5, 20062006-04-05
Randy Yates wrote:
> gans1973@rediffmail.com writes: > > >>hi, >> I had certain querries about adding dither in audio signals. >>1) I presume one of the reasons to add it is to compensate for >>quantisation error while >> converting analog to digital. Is there any other reason for adding >>it? > > > No.
How about converting digital of one amplitude to another (gain change)?
>>2) Is it added only when the signal strength is zero, or would it be >>advantageous to add it >> always. If quantization error is the primary consideration, I >>presume it would be >> advantageous to add it always. But it seems that there are some >>systems, >> where its added only when the signal strength is zero for >>sufficient period of time. > > > Keep it on all the time. To go back and forth between dither/no dither > is to modulate your noise floor, and that's not a good thing.
OT: It's snowing heavily in Central NJ right now. The grass is covered, but not thee pavement (yet). Jerry -- Engineering is the art of making what you want from things you can get. &#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;
Reply by Randy Yates April 5, 20062006-04-05
Steve Underwood <steveu@dis.org> writes:

> Randy Yates wrote: >> gans1973@rediffmail.com writes: >> >>>hi, >>> I had certain querries about adding dither in audio signals. >>>1) I presume one of the reasons to add it is to compensate for >>>quantisation error while >>> converting analog to digital. Is there any other reason for adding >>>it? >> No. > > Eh? Wouldn't you dither when changing the number of bits of a signal > which is already digital? Wouldn't you dither within an IIR to > mitigate the awfulness of those things?
I took the OP's question to be "When you add it during A/D, is the only reason to compensate for quantization error?" and not "Are there any other situations in which you use dither?" -- % Randy Yates % "...the answer lies within your soul %% Fuquay-Varina, NC % 'cause no one knows which side %%% 919-577-9882 % the coin will fall." %%%% <yates@ieee.org> % 'Big Wheels', *Out of the Blue*, ELO http://home.earthlink.net/~yatescr