[regal]

All this talk here in the source compenents forums about 16 vs 24 bit and also Jittery sound how about talking about something that more audible .

Here are a few quotes:

" Altmann Micro Machines’ progenitor, Charles Altmann, is much more empathic about this issue. In his opinion, there is no music possible with sigma-delta DACs. In his experience R2R chips are the only way to achieve a listenable sound quality. "

" The PCM 1704 is the best sounding DAC chip ever made, sigma deltas don't even come close" - Erno Borbeley.

" The AD1862 is the best engineered DAC chip." - Jocko Homo


I guess my limited experience with sigma delta chips have been bad as well, what I have heard have been harsh. Even the instrumentation industry forced BB to bring back the R2R 1704 because of isses with the 179x series.

1990's DAC's are sky-rocketing in price on ebay, people are catching on.

 

[gyrodec]

You seem to be on a one-man campaign to abolish discussion of jitter. Every thread I've read that mentions it, you pop up and say it its all lies. Did jitter dump you for a another man or scratch your new car. Why the hatred for jitter?

 

[Pars]

I don't think Regal understands jitter, hence...

As for R2R DACs, they are quite a bit more expensive to fab, as the resistors (R2R) require precision laser trimming. That is the main reason why manufacturers moved away from them. I still use a TDA1541a, so I can't comment on sound specifically.

 

[22906]

Yesterday I read this amazing article on the subject of delta-sigma DACs:
Oversampling and bitstream methods in audio

I don't think he does justice to R2R architecture, he kind of just dismisses it because of the high cost of manufacture.

But it seems that delta-sigma DACs are a natural extension of the concepts of oversampling and digital filtering.

I use a NOS ladder DAC (USB -> PCM2707 -> TDA1543) myself and find the sound slightly clearer and more realistic sounding than my oversampling DAC (Pico).

As far as I can see the only problems for R2R DACs are quantization noise and building a good low-pass filter.

 

[regal]

Quote:

Originally Posted by Pars /img/forum/go_quote.gif I don't think Regal understands jitter, hence... .

I understand that measuring something at the pico-second level is not something I trust Stereophile magazine to pull off. But then I have probably 20 years engineering test experience on you, hence..


I wanted this thread to be about a significant source component feature. Thanks for the thread crap gyrodec.

 

[nikongod]

so we are putting DAC chips designed for economy into dacs designed (allegedly) for uncompromising sq at any price.

sweet.

 

[regal]

Quote:

Originally Posted by nikongod /img/forum/go_quote.gif so we are putting DAC chips designed for economy into dacs designed (allegedly) for uncompromising sq at any price. sweet.

This is exactly the issue I wanted to bring up before we got side-tracked with trillionth of a second neurosis.


These older nicer R2R based dac's are gaining a lot of popularity, what I have heard in direct comparison (rather limited) is surprising. I am curious if others have experience with this or if I am overgeneraizing?

 

[Pars]

Quote:

Originally Posted by regal /img/forum/go_quote.gif I understand that measuring something at the pico-second level is not something I trust Stereophile magazine to pull off. But then I have probably 20 years engineering test experience on you, hence.. <snip>

I'm sure you have regarding the test engineering experience. The statement I made was prompted by your statement in another thread about listening to wavelengths in the 10 E-6 or 10 E-9 range. No offense intended.

 

[gyrodec]

Regal

Sorry to "crap" your thread again, but you make staements that need addressing, like "side-tracked with trillionth of a second neurosis".

You were trying to bring down a jitter thread (Basic question about jitter ...) by questioning if the tests are officially certified in some way, and here by saying you wouldn't trust Stereophile to test pico seconds. I wouldn't trust Stereophile to measure pico seconds either, of course they couldn't do that accurately. BUT, if you had bothered to read the tests you are lampooning, you would have noticed that that is not how jitter is measured. A special test sinal is create, roughly 11Khz with the LSB toggled at 229Hz, and ordinary voltage measurements are made of the sidebands this creates at the output. The pico second jitter "measurement" is just the summing of the size of these sideband using some fancy maths. Nobody measures anything in the time domain. Its all voltage measurements.

This is why Pars and I were questioning this hatred you have for jitter, as you don't seem to be that well informed about it. If you take an extreme position, and spread that opinion around, you should expect to be tested on it.

Sorry if my first post was a little too sarcastic.

P.S. My Meridian 207 has a ladder DAC, and it sounds great, even though I'm sure its jitter performance is a bit poor my modern standards.

 

[regal]

Quote:

Originally Posted by Pars /img/forum/go_quote.gif I'm sure you have regarding the test engineering experience. The statement I made was prompted by your statement in another thread about listening to wavelengths in the 10 E-6 or 10 E-9 range. No offense intended.

No what I said was if people can barely hear a 10^4 range tone how can people hear a tone time shifted 10^9 to 12. Maybe I didn't expain it well.

 

[regal]

Quote:

Originally Posted by gyrodec /img/forum/go_quote.gif Regal Sorry to "crap" your thread again, but you make staements that need addressing, like "side-tracked with trillionth of a second neurosis". A special test sinal is create, roughly 11Khz with the LSB toggled at 229Hz, and ordinary voltage measurements are made of the sidebands this creates at the output. The pico second jitter "measurement" is just the summing of the size of these sideband using some fancy maths. Nobody measures anything in the time domain. Its all voltage measurements. .

Yes I know they don't measure in the time domain, but their results are in the time domain.
There are no standards, how do you calibrate this test? And guarantee a trillionth of a second accuracy? The sine wave would have to be very accurate, is it NIST traceable?

 

[gyrodec]

No, I'm sure there aren't real standards for this and that they aren't NIST traceable, but that doesn't mean that they are wrong.

The sine wave is perfect, as it is created directly in the digital domain. The numbers required for the signal are either fed directly to the DAC from a computer, or the test signal is burnt onto a CD. The sine wave and its LSB toggle were never analog, and so is "perfect" (of course nothing is, but that signal is real close to it for the purposes of the test). The accuracy of the test depends on only the accuracy of an AudioPrecision system to measure voltage. Assuming Stereophile had their AP system calibrated recently (that process would be to NIST standards), then the accuracy should be just fine, even to the 10 pico second level (educated guess, no proof for that). Repeatabilty between different testers will be lower than that level however, due to differences in the digital and anlog interconnects, local RFI etc.

 

[regal]

Good info, but wouldn't jitter from the burn (CD) or transfer to the DAC would corrupt the test. Without a standard somewhere you are at ground zero.

 

[gyrodec]

Good question, and one that occured to be while I was away from my desk. From the Stereophile jitter charts they mark some peaks as "data related" and others as not. The data related ones are all cause by the tricky test signal, the others are from sources like the CD burn, or sp-dif transfer for the DAC case, just as you noted. Now we hit the limit of my knowledge, because I don't know if they just use the data related ones for the jitter reading or not, but, if they did, that would skip these issues and make the test clean.

To me, a more interesting question, and one that plagues all hi-fi measurement is - Ok we have a figure 10ps jitter, 0.001% IMD, ... but does that really mean anything. I think jitter is quite important, but, even with that I'm not sure the Stereophile, or other published charts, are testing jitter in a way that correlates with sound quality, though I'm am sure they are accurately measuring some aspect of jitter performance.

 

[22906]

So, about those R2R DACs...