8xNOSDAC

[cviller]

>>>But why does it sound nicer with higher clocks then?

easy answer - more samples at audible (up to ~20khz) range .
much easer to recostruct original signal without too coplicated brickwall filter .

regarding Niquist -Shannon theorem - problems moved higher (if jitter is under control)...

I don' think that is the explanation. There is a already a 8x upsampling filter in pcm1798, so the ear's low pass filter should well below frequency used in DA conversion.

cviller, try your HTPC/DAC combo with this free program:
http://audio.rightmark.org/products/rmaa.shtml
at the different rates (settable in the program) and see what's going on... hopefully

Looks kinda like the program Ulas posted in another thread - but more advanced. I think I'll try it tonight, if it will run on my Linux box.

I think my findings are related to better ratio between jitter and signal at high frequencies or some effects of interpolating done in the computer. It might just be effects similar to enlarging a photo digitally and applying a smoothing filter - looks horrible, but maybe the ear is pleased.
I should try to test it with a better 44.1k source than a radio stream...

Too bad Ulas schem is missing. How is it different than a upsampler like df1704? Is it the "linear" interpolation which is better than digital interpolation?

 

[ssmith]

Re: Re: Re: Re: 8xNOSDAC

DDDAC users collectively think this dac sounds great

It sounds fantastic.

I have the USB>I2S and 2 dac towers. Haven't got the right rload resistors to hand, so am only using one tower for the time being. So still another tower to add, still mods to do from stock parts (onboard caps, rload resistors), but it still sounds excellent as is.

My only comparison is a modded M-Audio Transit feeding SPDIF to a Benchmark DAC1. Doede's design beats that. Not really a knock out, just subtle all-round improvements and much easier on the ears for extended listening.