SAA7210 to PCM56, a Dutch-American connection

Status
Not open for further replies.
Jump to Latest
#103
who records a sqaurewave from a generator onto a CD. Then play it back again. Should be a sqaurewave without ringing, right? Closest to the original.
Click to expand...

Was not stated by me Nixie

You are right about the filters from crossover and after dac, total different application, good from you to correct me and point that out.

The best solution would be after all this discussing: (thats where diyaudio is good for) : Add a switchable filter after the dacs, choose for yourself at that very moment. Sometimes i like Nonos, sometime dig os (and sometimes the 14 bit tda1540!) also a strange perceptive thing.
 
#104
weissi said:
IS it so hard to understand THAT a square wave with a period of, let's say 10 kHz, ist not compareable with a sine of 10kHz Signal.
Click to expand...
Again, you cannot hear a square wave. It is physically impossible. Your year will remove all harmonics above around 20 kHz and will leave the same rounded wave with lots of ringing that a perfect filter does -- with the caveat being whatever small difference the limited support of the ear's filter creates.

I am looking at this ultra-high end DAC with ovenized 24-bit discrete R2R ladded, Lavry DA924, and his approach looks best -- limited oversampling, very steep filter (7 pole)
 
#105
IS it so hard to understand THAT a square wave with a period of, let's say 10 kHz, ist not compareable with a sine of 10kHz Signal.
No way.
Click to expand...

Hi weissi

With your comment and looking at Bernards dig. OS vs nonos square waveforms, i must conclude that the nonos waveform resembles most with a sinus. And music resembles the most with sinuses.
 
#106
tubee said:
With your comment and looking at Bernards dig. OS vs nonos square waveforms, i must conclude that the nonos waveform resembles most with a sinus. And music resembles the most with sinuses.
Click to expand...
Oh come on guys, you're not even trying. A non-versampling DAC produces as much energy in images as in the actual audio band. This is very bad for the analog stages after the DAC. To remove them from a non-oversampled DAC you would need an extremely steep analog filter, which cannot be passive (means you'd have a bunch of opamps). There are graphs for the imagination-challenged in this easy to understand white paper: http://www.lavryengineering.com/white_papers/sample.pdf
If you can't make sense of this, then you've no business arguing about DACs and filter design.

Also it's worth checking Lavry's posts at this forum (would be nice if a good digital audio engineer like him came to diyaudio... moderators hint hint nudge nudge)
 
#107
Oh come on guys, you're not even trying. A non-versampling DAC produces as much energy in images as in the actual audio band. This is very bad for the analog stages after the DAC.
Click to expand...

Nixie, this could be the reason why tube amps are often chosen, and tube outputs in dacs.

I must say the rising nonos waveform resembles more a sinus, but don't look at the falling edge....

Will read the posted wiki's carefully first.
 
#109
'think there's a reason why most ppl do not like opamps as I/V converters.... tubes are not as sensible as solid state devices when it comes to HF.

Those pictures posted from Bernhard tell nothing unless the measurement conditions are not cleared up, I think it's not that easy to compare Oversampling DAC'S to non-os by measurements with an Oscilloscope... You'll need a good FFT (not a software FFT based on a PC-audiocard...)
Anyway, non-os doesn't sound as bad as it looks in theory.

cheers
 
#110
there's a reason why most ppl do not like opamps as I/V converters.... tubes are not as sensible as solid state devices when it comes to HF
Click to expand...


It is becoming better these days, in the last few years a few new opamps have seen the light, and they can sound rather well. But its not only the opamp itself, the implementation of it too. When the opamp is used as filter it has to eat all HF rubbisch, and that's not the best way imo, for sure when a 4V/uS opamp is used. That's why i chose Pedja's AD844 stage for experimenting, opamp is used as a common base circuit, like done often in HF.
 
#111
weissi said:

Those pictures posted from Bernhard tell nothing unless the measurement conditions are not cleared up, I think it's not that easy to compare Oversampling DAC'S to non-os by measurements with an Oscilloscope... You'll need a good FFT (not a software FFT based on a PC-audiocard...)
Click to expand...

My pictures show what you can read about the os filters in various discussions: They produce pre-echos which are unnatural.

By the way this "Oscilloscope" does 32000 points FFT with 16 bit ADC resolution 😉 on request.

IMHO it does not make much sense to show a FFT of the squarewave, which is not really a square but SMPT code from test CD.
I have no square on CD.
 
#118
Re: another sidetrack

Cobra2 said:
about filters, how do this perform? Does it have a (type)name?

Arne K
Click to expand...

They're called GIC filters from Generalized Immittance Convertor. They simulate inductance the electronic way. Kind of.

More theory on them here:

http://focus.ti.com/lit/an/sbaa001/sbaa001.pdf

I like them because the filter/opamps are not in series with the signal path but they're parallel with it !
 
Status
Not open for further replies.
Top Bottom