Re: Test report MBP built-in audio device
Re: Test report MBP built-in audio device
- Subject: Re: Test report MBP built-in audio device
- From: Richard Dobson <email@hidden>
- Date: Tue, 26 Aug 2008 15:24:06 +0100
Mikael Hakman wrote:
..
Distortion caused by a device (and possibly its driver etc.) itself, on
the other hand, varies with a number of external factors. From where did
you get that "a DAC cannot have stabilization time that lasts longer
than one sample period"? This is untrue. I say that DACs take time to
stabilize.
You can say it; but that doesn't make it true. You have to show it with
evidence (from a reliable source, such as an AES paper) that can be
replicated. Otherwise it is just a snake-oil argument. I say you have
too much oxygen in your cables.
Can you cite an independent source for your claim that it is quite
common for DAC to exhibit higher distortion at the very beginning of
a signal?
I cannot cite independent source, because they don't publish this data,
probably because they don't know how to measure it (in the best case) or
don't want to (in the worst case) or simply because they don't do it.
Or because it is a delusion. Disortion on the scale you describe would
be visually as well as aurally obvious. If it were a real effect,
believe me, ~everyone~ would have documented it by now. Can't you add
an imput/output waveform plot to your report?
..
You are saying this without any proof. There is nothing wrong with my
analysis and the proof is that it gives correct answers if the signal
does not pass through a DAC. If my computations were causing distortion
variation, then this would be present also when not passing the signal
though a DAC, wouldn't it? The effect is there only if there is a DAC in
the path.
For all we know, you have an earth loop, or are picking up RF
interference. You are making a circular argument - you say your method
is correct because it produces correct results - but you are also saying
you are the only person capable of producing those results.
You have difficult to accept my results because you never before seen
such an analysis. Well, somebody has to be the first, in this case I'm.
Windowing or not has nothing to do with it. You are free to give me a
spec for a windowing function of your choice, including all its
parameters and I will repeat measurements using your function.
I propose the standard Hann Window. Frequency-domain coeffs:
Fw_t = 0.5 F_t - 0.25[ F_{t-1}+F_{t+1}]
It isn't MacBook Pro distortion that varies but in there used codec.
This is normal.
In fact, this is worst case when you start the test signal after
silence, because the FT sees a repeating waveform that is very
complex - and this totally explains the high distortion you're
seeing in your custom tools.
Worst case!? Well, about as worst as saxophone starting to play a
note from silence or near silence. About as worst as an artist
playing a slow classic composition on acoustic guitar and letting
every note decay to low levels! What are you talking about?
Because your signal is (in effect) a sinusoid multiplied by a step
function (see my other post). The step function (as does any transient)
results in a multitude of harmonics (possibly variant in frequency).
Inside the FFT, it is equivalent to an infinite number of very short
attacks, with no steady-state. Think of it as like switching on the amp
while a sound is already playing. You are analysing the click (or the
speaker thump), not the signal.
...
Who said that a DAC is a perfect linear time-invariant system? On the
contrary, all tests aim at assessing how far from this ideal a
particular device is.
Of course they are not "perfect", but were the codec as bad as you are
claiming, the machines would be unusable; worse than the
cheap'n-cheerful musical card I was given a few months ago. Even cheap
24bit codecs are linear over at least 22bits these days (with the
errors almost entirely at the bottom - where, in the absence of liquid
cooling, brownian noise starts to get in the way).
So to be clear, are you claiming the codec is seriously time-variant, or
seriously non-linear, or both?
If Fourier analysis had trouble with attack time then it would be having
it even without DAC in the path, wouldn't it?
Because what? I send pure simple sine wave
We have no evidence yet of how pure your sinewave is. Have you measured
it independently? Can you prove that all harmonics and noise are more
than (say, generously) 100dB down? Are you generating it with analog
means (what instrument), or digitally? How, exactly? Simple linear
interpolation of a wavetable? Direct sine computation? Double precision?
Are you really expecting the startup of a note to show the same spectrum
as the steady-state part of the note? What do you expect the spectrum
of this to be (ASCII art, needs mono-spaced font):
________________ (envelope)
| /\ /\
______________|/ \ / \
0------------------------------
^ \ / \
(discontinuity)| \/ \ etc (pretend this is a sine!)
t=0 t=1024
?
Richard Dobson
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Coreaudio-api mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden