🚀 go-pugleaf

"Arny Krueger" <arnyk@pop3free.com> wrote in message
news:WKLO6.2$4N2.462963@newssvr16.news.prodigy.com...
>
> "Trevor" <trevor@home> wrote in message
> news:3b0b6499$0$25475$7f31c96c@news01.syd.optusnet.com.au...
>
> > "Arny Krueger" <arnyk@pop3free.com> wrote in message
> > news:96CO6.3473$EW5.538921099@newssvr16.news.prodigy.com...
>
> > > As far as your claim to have measured 97 dB SNR in a 16 bit
> > > environment, that appears to be better than that which is
> > > theoretically possible. Digital I/O can help, but not quite that
> > > much!
>
> > Hey Arny, I wouldn't complain about 96.3 dB being rounded to 97 dB
> instead  of 96.
>
> That's known in the trade as a rounding mistake. Let's put this in
> perspective. Mr. Dormer is way over the top. He's taken the
> unreasonable approach that you're accusing me of. He's said:

I have no beef with either of you. Lets stick to the technical points.

>
> Paul Dormer wrote:
>
> "So why don't you do something pro-active with your life and admit in
> the relevant locations on your website that your results with the
> Live! are flawed and/or antiquated, rather than allowing users of
> your website to be perpetually misinformed?  Why don't you finally
> correct the widespread rumour that you started, that the digital
> output of the
> Live! is worse than the analog output.. which it MOST CERTAINLY IS
> NOT!"

I have to agree about the digital output, if not the language.

>
> >It's SpectraLabs output after all, and none of YOUR measurements
> seem  to be calibrated or traceable to any National standard from
> what I can see?
>
> I also don't claim > 96 dB SNR for any 16 bit cards.

Quite right, he was sloppy.

> > > The cards I tested did what I say they did, pure and simple.
> Others
> > > have confirmed my results. I've confirmed my own results with
> other
> > > samples of the same model card.
>
> > This I have no argument with, I am sure such a cheap card does not
> have mega
> > strict quality control. Would you expect it to?
>
> IME different samples of the same card tend to be within a dB or so.
> This was not true in the past, but it has been true since the advent
> of sigma-delta converters, which are essentially all digital.

Something is going on though. I assume the numerous variations of the live
have some differences then. I get better results using standard drivers.
Paul gets better results using APS drivers. Your results are not that much
worse anyhow.

>
> But Dormer is not comparing apples and apples. He's obviously using a
> different card than I am, and no doubt different drivers. He's been
> yelling and screaming and asking for my head on a plate for a week,
> when nothing he does really lines up with what I did on 2/2/2000 -
> over a year ago.

Many things can affect the analog results. As for the digital I believe it's
the APS drivers.

> > > Mr. Dormer, if you've got a new hotshot card that's got analog
> I/O
> > > that's 4 dB better, that's not totally unbelievable. Good for
> you! I
> > > guess its the luck of the draw. But measuring 97 dB in a 16 bit
> > > environment takes some doing... especially if your test signal is
> > > properly dithered. Theoretical max is 96 dB unweighted, isn't it?
>
> > For the record I think your tests are a damn sight better than
> nothing,
> > which is pretty much what else is available. And I do read them and
> > appreciate having some data on cards that I have no access to. Do
> you really
> > need to make such ridiculous statements though?
>
> IMO it's not ridiculous.
>
> >1 dB measurement error is a hanging offence?
>
> It's not a hanging offense, its one of those things that slips by
> people who don't know what they are doing. It's like turning in a lab
> report claiming to have measured 1 degree below absolute zero.

True.

> >1 dB roll off in a card at 18 kHz makes it useless?
>
> Not at all. See my next answer.
>
> 83 dB  S/N ratio is inadequate for recording vinyl or cassette?
>
> Please read http://www.pcavtech.com/soundcards/compare/index.htm :
>
> "If the frequency response is within +/- 0.2 dB from 80  to 13 KHz
> and does not peak or dip excessively (say more than +/- 3 dB)  32
> Hz-17 KHz,  then the card is probably going not going to color the
> sound in such a way that it will bother you."
>
> "If the SNR is above about 75-80 dB, then noise is not going to be
> too much of an issue as long as you keep your recording levels up. In
> practice much of the audible benefit of SNR's > 80 dB is that you
> don't have to be so careful about setting levels because if you are a
> little on the low side, the noise from the card still won't intrude."
>
> > Give it a rest. You're destroying your own credibility.
>
> Trevor you are putting words in my mouth, and in my book that IS a
> hanging offense.

OK, I apologise if I mixed others statements with yours. I believe I only
answered statements as written, and did not direct deliberately at anybody
else.
However you did appear to join in the criticism of the SB, which IMHO is
still unwarranted at the price of the Value, especially considering your own
results.
Nobody has yet nominated a better card at a cheaper price. I have agreed
though, that the more expensive SB packages are a poor choice.

> You are also letting Dormer ride free when he makes the precise
> errors that you are falsely accusing me of. Virtually everything
> you've said about me applies to Dormer's wild claims. So why aren't
> you saying it to him?

I have no interest in attacking anyone. I just agree or disagree with any
statements as appropriate. Both of you have achieved essentially similar
results considering the measurement conditions and equipment used. I was
unaware the real problem is past history.
I think I will give it a rest now :-)

Trevor.