Dithering is a Mathematical Process - NOT a psychoacoustic process.

[scott wurcer]

One hopes Bob and Tom's work will suffice! Apparently mine is meaningless against the power of belief.

Shannon connecting information to entropy and the second law was sheer brilliance. His daughter was my neighbor in Cambridge for a while, she came by and said hi one day and I said you must be Claude's daughter the resemblance was so strong.

 

[DonH56]

One hopes Bob and Tom's work will suffice! Apparently mine is meaningless against the power of belief.

Always tilting at windmills?

I keep forgetting my resolve to stay out of these type of debates where clearly logical arguments and science/engineering references do not matter.

 

[j_j]

Always tilting at windmills?

I keep forgetting my resolve to stay out of these type of debates where clearly logical arguments and science/engineering references do not matter.

The alternative is having, well, having "young earth creation science" taught in place of evolution and plate tectonics.

 

[exaudio]

I know several others in this thread have already shared various graphics and plots illustrating the effects of dither, but I was inspired to try my hand at it, so here's one more for the pile.

I used SoX to generate dithered and non-dithered versions of a 1kHz sine wave sampled at 384kHz with a sample depth of 8 bits. Then I used Mathematica to plot the wave form and spectrum of each. Here's the result:

 

[j_j]

If human hearing would not have been equipped with "FFT spectrum analyser" in the inner ear there would have been no need to perform this operation on sound signals. With this caveat the operation is really mathematical and helps to preserve tonal components of a signal which were below LSB before quantization. Non-tonal components (noise-like) can not be saved for human hearing by this method.

Really now? Seriously? Now we're saying THAT?

Here's a narrowband noise signal, with the same plotting as the original sine signal.

Yes, one could do wider band noise, but I'm pretty sure that the statistical approach to detection will require endless explanations.

Oh, look, there's the noise signal in the output. Yes, it's noisier, but it's still detectable and this is purely by analytic methods.

 

[Serge Smirnoff]

Information theory has been around for a LONG time, I am not sure what you mean by it being applicable in the future?

Information theory is fundamental to discrete-time sampled systems.

Today information theory is extremely wide area of research. I meant the information in a sense as it is considered/viewed in semiotics. Audio information can be researched in syntactic, semantic and pragmatic aspects/layers. Syntactic layer deals with a signal, semantic - with perception, pragmatic - with a meaning of sound/music for individuals and society. I always had this in view when developed df-metric. Df level is pure syntactic parameter in this sense. And actually it is indifferent to the nature of a signal - a signal can represent change in sound pressure levels, prices on financial markets or, say, temperature in some system. Where comparison of waveforms has sense it can be helpful. Connection with semantic/psychoacoustic level of audio information in df-metric is realized by means of art.signatures. Such broad understanding of audio information is the primary source of my broader view of some terms accepted in traditional audio science, such as distortion, psychoacoustic processing and probably many others.

 

[j_j]

Today information theory is extremely wide area of research...

Claude Shannon's paper was published in when?

http://people.math.harvard.edu/~ctm/home/text/others/shannon/entropy/entropy.pdf

1948.

Information is a real thing. You can measure it. Dithering preserves information that is otherwise lost. PLEASE stop trying to redefine words to suit your argument.

Dithering is linearization. It is not psychoacoustic in nature.

 

[tw99]

@j_j There are some people you'll never convince that the world is round(-ish) !

The good thing is that this thread now exists to explain the truth to those people who want to understand, which is probably 99% of the readership.

 

[xr100]

Today information theory is extremely wide area of research. I meant the information in a sense as it is considered/viewed in semiotics. Audio information can be researched in syntactic, semantic and pragmatic aspects/layers. Syntactic layer deals with a signal, semantic - with perception, pragmatic - with a meaning of sound/music for individuals and society.

I had to Google the above terms and it seems that their origin is in structuralism/post-structuralism/linguistics?

Leaving aside their validity, I am all for broader and cross-disciplinary views but it is clear that you are mixing categories and getting into an epistemological mess here.

Probably, they could be used to look at, say, a "bitcrushed" sound in a recording (as discussed in the other thread.)

What is being discussed here, though, is the application of dither to linearise quantisation, as set out in the ("narrowly" (!) defined) discipline of information theory. No "broader" perspective, valid or otherwise, is needed.

And, I thought the objective of the "DF Metric" was to identify audio equipment and/or processes which yield perceptual transparency?

 

[xr100]

Apparently mine is meaningless against the power of belief.

You can lead a horse to water...

 

[DonH56]

You can lead a horse to water...

Smirnoff...

May explain some things...

 

[mansr]

Interesting. My first reaction is that that's going to have some high level linearity oddities. Maybe not too bad at 200mV levels.

Easy enough to check.

Here's such a CD player (Philips CD150) playing a 1 kHz tone at -6 dBFS without emphasis:

And here's the same 1kHz tone with the emphasis flag set (same digital data):

Both look rather nasty by today's standards. The de-emphasis filter lowers both noise and harmonics, yielding better figures in this measurement.

 

[j_j]

Easy enough to check.

Here's such a CD player (Philips CD150) playing a 1 kHz tone at -6 dBFS without emphasis:
View attachment 48759

And here's the same 1kHz tone with the emphasis flag set (same digital data):
View attachment 48760

Both look rather nasty by today's standards. The de-emphasis filter lowers both noise and harmonics, yielding better figures in this measurement.

Looks like there are other problems beyond the jfet there.

 

[j_j]

Smirnoff...

View attachment 48748

May explain some things...

Gan bay!

 

[mansr]

Looks like there are other problems beyond the jfet there.

Well, there's the TDA1540.

 

[SIY]

Here. Well, nearby, anyway.

 

[restorer-john]

Both look rather nasty by today's standards.

It's Philips first 14 bit D/A converter (2xTDA-1540P/S lowest grade epoxy packs) strapped onto their first 4x OS filter in the cheapest CD player they could bring to market in 1985.

Also compare apples with apples. It's 16/44 not 24/44 which all ASRs DACs get tested with- a format that doesn't really represent the majority of files or physical content does it?

And it beat its spec of 0.005% by a decent margin. And what is your A/D's contribution BTW?

Don't get me wrong, the Philips CD-150 was a horrible cheap nasty thing, but let's put its performance into perspective- it's actually quite good.

 

[xr100]

Both look rather nasty by today's standards. The de-emphasis filter lowers both noise and harmonics, yielding better figures in this measurement.

Frequency response? Or no suitable test tracks?

 

[restorer-john]

Frequency response? Or no suitable test tracks?

It's not a full review @xr100 !

You can use white noise or a multitone to give a pretty decent FR.

 

[xr100]

It's Philips first 14 bit D/A converter (2xTDA-1540S lowest grade epoxy packs)

Can you elaborate, please?

IIRC, the early players weren't too hot in terms of having properly engineered/specified power supplies/isolation between sections. Performance could depend on how hard the transport has to work...