Re: [fluid-dev] Purpose of dither?

[Z F]

--- Edenyard <address@hidden> wrote:

>       Can somebody please re-assure me that I'm not dreaming all of this: 
> I've tried pinching myself and I SEEM to be awake, but....
> 
> Mihail Zenkov <address@hidden> wrote, very reasonably:
> 
> > :) Where are you read this? I six years do amplifiers (mainly on
> > tubes)
> > and acoustics. All amplification strongly linear. Otherwise, we
> have
> > big THD and IMD.
> 
>       Absolutely right, Mihail.

I agree too.

>       Furthermore (just for the record), tube (or 'valve') amplifiers will
> 
> generally have worse linearity than solid-state ones because they
> need a 
> large, lossy, iron-cored transformer at the output to match the high 
> impedance of the valve output circuit to the low impedance of the 
> speaker coils. In actual fact, tube amplifiers are much more likely
> to 
> provide nolinearity (the so-called 'soft clipping') for
> high-amplitude 
> output signals than solid-state circuits, due to (among other things)
> 
> saturation of the output transformer core.

That is why they are not used at full amplification!! Only thier linear
piece is used, but it is still usually (as I recall) longer than
in semiconductor ones.. And, as I said, the trick was used to
"increase" the dynamic range of consumer electronics. No one says that
it should be done like this in high quality audio. Plus, you were not
carefull to undestand that I am not talking about instantaneous output
versus input which should be linear, I am talking about average. So it
is a relatively slow process which would adjust gain based on the
average input which would cause only slight distorision. Only rapid
jump in the amplitude would clip. And this is not the automatic gain
control either. But regardless, we are dicussing a different problem.

My point was that there is something wrong in the setup which Mihail
uses and not with
the synthesizer itself. The fix was proposed to modify the synthesizer
which I think is not correct. This is where we are arguing. Can you
make 
a comment on that? My argument was the he is connecting a relatively
low
quality source (16bit) to a high quality output and is not satisfied
with the 
results. I understand his frastration, but I do not think that adding
dither is a solution to his problem. Solution would be to add bits to
DAC. what is your opinion?

> 
> > when I was little I used to fix analog
> > consumer audio/video equivment, but I do not do that anymore. 
> 
>       Well - with all due respect, it's probably just as well that you 
> stopped when you did. The world doesn't need non-linear audio
> amplifiers.

I agree with you on both issues. As to the first one, I can add that I
found that many (not all) engineers have no respect to mathematics
and I do, so I stopped.

And no, you are not dreaming and this is not a nightmare.

Good night.

ZF

__________________________________________________
Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around 
http://mail.yahoo.com