Getting the best out of a Rega DAC (or any Dac)?

[Paul R]

It's important, but not as a distinguishing feature of a DAC. It's generally a done deal, and has been for years.

Paul

 

[Daniel.]

Id probably agree with that but leave room for doubt. I mean we stream music from around the world which in the end gets buffered and reclocked with no audible impact.
I think the differences between interconnect types is unrelated to jitter.

 

[darrenyeats]

Is it really worth trying to improve the performance of DACs from current levels from an audibility stand point? Whatever the answer to the question, manufacturers will continue to try to improve performance and jitter is a factor.

It's worth mentioning that jitter occurs in many places. Even if you eliminate jitter from a S/PDIF stream, the DAC chip still has jitter and noise caused from internal and external sources.

 

[adamdea]

Iv heard that a system using pll causes audible distortions unrelated to jitter, could this be a possible reason why optical could sound inferior to async usb.

And I've heard that putting Cds in the freezer makes them sound better. How exactly could a PLL cause distortions?

The PLL can itself generate noise. I understand that this was particularly the case with analog PLLs and it was argued that they could make the dacs sound worse (presumably by the noise either modulating the conversion clocks timing or by infectign th anlog stage) I'm not sure anyone (except perhaps retro-fi nos dacs maybe) use them now. I have the impression this is not really considered a problem with Digital PLLs but I dare say that everything in the world has some sort of effect which could in principle be measured.

I have heard an argument that the optical converters for toslink themselves generate noise too and that this is the "reason" which Optical connection "sounds" different

Why the inverted commas? because it certainly isn't my experience that optical connection sounds any different from coax etc.

Jitter becomes more important at higher frequencies so wouldn't the jitter approximation be different for oversampling dacs, especially ones with the highest oversampling frequency's..

It's a long time since i looked this up, but i think the answer is yes and no as regards the susceptiblity of oversampling dacs.
When you say more important at higher frequencies do you mean
a) the frequency of the jitter
b) the frequency of the music in the signal
c) the sampling rate at which the dac operates.

For c) I suggest you look it up- as I say , my recollection is that the answer is quite complicated.

For a) and b) standard modeal is to say that a particular frequency of jitter at conversion creatres sidebands s[pearated from the signal by the frequency of the jitter. so if your signal is a tone at 2 khz and the jitter is at 1khz, you will end up at dsidebands at 1 khz and 3 khz. This would potential be fairly audible if at a sufficent amplitude
i) because it's far enough apart from the 1khz tone not to be masked by it
ii) because 1 khz and 3khz are in the sensitive part of our hearing.

Now consider a 1khz tone with jitter at 20Khz. This will produce sidebands at 21khz (no chance of hearing) and 19 Khz (I think, due to reflection around zero, or if you prefer its just -19khz which (I think) is about the same thing)

Equally consider jitter at 1 hz. This will produce sidebands at 999hz and 1001 hz. You wont hear them because of masking it will just sound like a 1 khz tone.

But remember if the jitter is broadband (ie at all frequencies) then it will make sidebands at all frequencies and hence present as low level noise.

If you want to see more I suggest you read the following (but bear in mind that solutions have moved on in 20 years).


http://www.nanophon.com/audio/jitter92.pdf

http://www.nanophon.com/audio/antialia.pdf
http://www.nanophon.com/audio/diagnose.pdf
http://www.scalatech.co.uk/papers/aes93.pdf

Even better is a book from the late Julian Dunn "Measurement Techniques for Digital Audio v2" available for free download on
http://audioprecision.com/download/whitepapers

Yea its hard to say, best always to leave room for doubt and keep researching.

Absolutely. But beware of solutions to problems which might be illusory or overstated.

Also always bear in mind that only jitter on the conversion clock really matters. All previous jitter else is irrelevant if it does not cause the clock timing the conversion to be inaccurate. (although there could also be more indirect effects if noise gets to the analog circuits)

Also remember, before anyone tries to confuse you, that the rate at which the clock needs to convert to analog is not a secret- we know its 44.1Khz for red book. It does NOT need to recover this secret information from the transmitting device. This may have been the case with early dacs but not with anything (bar retro) made in the last 20 years. You will often read about a "recovered clock", but this is more of a way of matching the flow rate.
Since you are interested in the academic understanding of the field you may be interested in this quote from the leading Amercian text book on the field. I think you may find a similar one in Watkinson.

Pohlmann's Principles of Digital Audio (5 ed) p. 135
" Audiophiles have sometimes reported hearing differences between different kids of digital cables. That could be attributed to a D/A converter with a design that is inadequate to recover a uniformly stable clock from the input bitstream. But a well-designed D/A converter with a stable clock will be immune to variations in the upstream digital signal path, as long as the data values themselves are not altered.......As noted, if a receiving signal can recover a data signal without error, interface jitter is not a factor."

 

[adamdea]

I think the differences between interconnect types is unrelated to jitter.

I think you're right. The problem is that in the case of digital cables- what else is there? Assuming the data is correctyl transmitted and received (which it will be) what's left?
Ignoring toslink you have the possibiity of transmitting electrical noise from trasnport to dac. But how are two different S/PDIF cables going to do this differently? They're conductors aren't they, and the better the conductor then surely it can only transmit the noise better.

In the case of usb cables some of them cut off the 5v or maybe the ground. But that is not the case with most fo them and many of the fancy schmantzy ones have just the same geometry. So how come people still say they "sound" different? DO we have to postulate complex quantum schmantum effects? Do we possibly introduce the sneaking doubt that there is any difference?

No it's definitely jitter.