[sdiy] Re: Limits of perception...

[cheater cheater]

But Tom, of course the point isn't if someone can hear supersonics. No
one can hear 100 kHz, it is physically impossible with the shape of
ears and mechanical properties of our ear's inner parts.

The point is that there is a situation where what goes on at 100 kHz
will become audible by influencing something else.. I can easily bring
up situations where this happens, but I'm sure you can too. This sort
of situation tends to happen with *everything* that you can measure
with any tool you can build. At some point it stops mattering in
*general* but because of the infinite possible complexity of
environments and setups that hardware is used in, there will quite
often be *particular* situations where the difference is quite
audible; and if you do not take those into consideration, then you are
not thinking about the best possible way of solving a problem. The
great thing about our chat here is that we can come up with the most
perfect solution here, and consider when a specific improvement will
or will not be audible. Blanket-bombing something with 'it will never
matter' without actually looking into the question at hand disables
this. This argumentation holds not only for audio engineering, but for
any discipline pursued by man, be it science, logic, mathematics, the
humanities or cooking.

Whether or not you use the added possibilities that come from very
specific considerations here is your choice - nobody's forcing you
and, frankly, some of the ideas posted here are too much for the day
to day hardware used by Joe Blow. But it's still cool to be able to
talk about this sort of thing, and I think it's worthwhile to question
'urban myths' and 'general knowledge'; sometimes it doesn't yield
anything new, but sometimes it can give a beginning to very useful
ideas. For example look at the SSM based exponential converter that we
have come up with here a few months back. I find it is not only
worthwhile but also very important to think about how to achieve ideal
solutions to problems; they can be a benchmark of what we're doing.

Fact is that no numbers have been brought up here beyond flat
argumentation that 'it does not matter'. But not having numbers you
cannot decide about engineering details like this. Sadly, being in a
hotel right now I cannot tell you what the numbers *are* and cannot
even build the hardware in question to listen if there's an audible
difference. But it works for you too :)

D.

On Mon, Feb 22, 2010 at 20:54, Tom Wiltshire <tom at electricdruid.net> wrote:
>
> On 22 Feb 2010, at 18:01, Nick Zampiello wrote:
>>
>>  do you think you can't tell the difference between 16 and 24 bit ( for
>> audio)?
>> do you think 44.1k sounds just like 48k?  96k?
>
> I don't know, but I'm damn sure that there comes a point when you *can't*
> tell. And that's what I'm getting at - above a certain limit, extra
> 'quality' is well beyond what human beings can perceive.
> My original contention was that that point is actually much sooner than most
> people think. Increasingly, I think digital technology is arriving at that
> point.
>
>> > it seems to me that the 'perception' answer is all about what people are
>> > tuned in to mentally.
>
> I disagree. You can be as tuned in as you like, but you still can't hear
> 100KHz. I agree that *exactly* where the limit lies might depend in some
> degree to how tuned in you are (ear training, genetic luck, etc) but that's
> a debate over 20Khz vs 25. There's a clear limit whether we're talking about
> pitch perception, bit depth, convertor accuracy, whatever. Human beings
> vary, so individual results vary, but that's science, and it doesn't stop
> there being a scientific fact about what the limits are.
>
> Like Greg said, if it rocks you, it works.
>
> Ok, I promise I'll shut up now.
>
> T.
>
> _______________________________________________
> Synth-diy mailing list
> Synth-diy at dropmix.xs4all.nl
> http://dropmix.xs4all.nl/mailman/listinfo/synth-diy
>