ya know what's cool about that paper, Vicki, is getting real OCD ("you can't
spell 'analysis' without 'anal'") with offsetting the dither so that the end
result doesn't really give a rat's ass whether it's round-up, round-down, or
round-to-nearest-even.
i cannot out-geek these guys. in a contest of paying attention to detail, i
will lose and it will hurt.
> On 01/10/2022 10:00 PM robert bristow-johnson <[email protected]>
> wrote:
>
>
> 12 ACKNOWLEDGMENTS
> ...
> We are particularly grateful to Morten Lindberg, Bob Ludwig, Mick Sawaguchi,
> Peter McGrath, and John Atkinson who participated in testing their recordings
> on workstations.
>
> Thanks also for many stimulating discussions on the dither topic over the
> years with Stanley Lipshitz, John Vanderkooy, Vicki Melchior, Rhonda Wilson,
> Wieslaw Woszczyk, and, of course, our co-conspirator, the late Michael Gerzon.
>
and i would give my left testicle to be on that list, Eric.
i think it was 1991 (whatever was the last year that AES was in the Hilton on
6th Ave), that i got to go on an exclusive walk with Michael Gerzon in Central
Park for about an hour (it was real close to the AES Convention). it was like,
uh, what they say is a "grace gift". total serendipity.
i also got to meet Jon Dattorro and the Eventide people. maybe i met you,
Vicki, then. i dunno. i remember meeting Rhonda and i had met Stanley earlier
when i was at Northwestern in the mid 80s.
--
r b-j . _ . _ . _ . _ [email protected]
"Imagination is more important than knowledge."
.
.
.
> > On 01/10/2022 7:42 PM Zhiguang Zhang <[email protected]> wrote:
> >
> >
> > sorry to beat a dead horse, but from this 2019 paper's synopsis:
> >
> > https://www.aes.org/e-lib/browse.cfm?elib=20457
> >
> > "Although dither methods for ameliorating quantization error have been well
> > understood in the literature for some time, these insights are not always
> > applied in practice. It is rare for an audio performance to be captured,
> > produced, and played back with a flawless chain."
> >
> > no, we aren't in the dark ages of concert-going as the only way to
> > experience music, and haven't been for quite some time. audio performances
> > are never experienced in a flawless chain, and even at the event you have
> > the potential of hearing the quantization error of a Behringer/Midas FOH
> > board's DAC
> >
> >
> > On Mon, Jan 10, 2022 at 6:51 PM Zhiguang Zhang <[email protected]> wrote:
> > > Seeing that the president of MQA came from Warner, I'm sure he's well
> > > aware of what goes on in the studio. It might just be a matter of your
> > > average modern consumer's distribution / consumption habits and various
> > > demographics
> > >
> > >
> > > On Mon, Jan 10, 2022, 6:47 PM vicki melchior <[email protected]>
> > > wrote:
> > > > Your confusion is different than Eric’s though. I was aiming to address
> > > > his previous post about "well i'm not here to talk about whether or not
> > > > i can discriminate dither from music”
> > > >
> > > >
> > > > I agree with your comments. Subtractive dither is used in the MQA codec
> > > > because Peter Craven is trying to salvage every last ounce of SNR, but
> > > > it's not difficult to handle in a streamed file, and in a situation
> > > > where the original file’s specs are measured during encoding.
> > > >
> > > > Vicki
> > > >
> > > >
> > > > > On Jan 10, 2022, at 5:53 PM, robert bristow-johnson
> > > > > <[email protected]> wrote:
> > > > >
> > > > >
> > > > > > On 01/10/2022 1:37 PM vicki melchior <[email protected]>
> > > > > > wrote:
> > > > > >
> > > > > >
> > > > > > Okay. In the context of subtractive dither, I understand the
> > > > > > confusion.
> > > > >
> > > > > I don't. I'm just confused.
> > > > >
> > > > >
> > > > > > Yes, the purpose of subtractive dither is to reduce the total
> > > > > > dither noise remaining in the output, typically after adding dither
> > > > > > during bit reduction. Clearly, that will increase the SNR
> > > > >
> > > > > But only by 4.77 dB. But, hey, it's 4.77 dB. So if they can
> > > > > standardize how the dither is generated from the LSBs of the
> > > > > quantized signal, and, if noise shaping is done, what the transfer
> > > > > function is from dither to quantized output, then why not do this
> > > > > subtractive dither thing?
> > > > >
> > > > > The worst that can happen is that the receiver does not decode those
> > > > > LSBs and subtract the dither. Then you're no worse off than if it was
> > > > > just additive dither and you don't recover those 4.77 dB SNR.
> > > > >
> > > > >
> > > > > > and allow you to hear more of the signal.
> > > > >
> > > > > only really makes a difference if the audio is living down by the
> > > > > noise floor (which it could be if it's a CD and classical or
> > > > > uncompressed acoustic music with a large dynamic range).
> > > > >
> > > > > but 4.77 dB is 4.77 dB. that's something. 96 dB dynamic range is
> > > > > better than 91 dB.
> > > > >
> > > > > --
> > > > >
> > > > > r b-j . _ . _ . _ . _ [email protected]
> > > > >
> > > > > "Imagination is more important than knowledge."
> > > > >
> > > > > .
> > > > > .
> > > > > .
> > > >
>
> --
>
> r b-j . _ . _ . _ . _ [email protected]
>
> "Imagination is more important than knowledge."
>
> .
> .
> .
