試左十幾隻dac 終於選擇MYtek 192

karsonkwan · 發表於 2013-5-16 22:08

chhanthony 發表於 2013-5-16 22:05
佢係唔係講個Beta firmware?

佢冇講firmware 問題，問我而家咩野setup，做乜要加word clock，諗住用咩野牌子。。。好多野吹

karsonkwan · 發表於 2013-5-16 22:55

Mytek 話用Clock 只會令聲音更差

Dear Karson,

I was asking about your target system, about the word clock source device.
In a few words: External clock source for Mytek Stereo192-DSD DAC will make its sound worse, not better.

Michal Jurewicz, founder of Mytek states:
"Frankly there is no benefit of using external clock with Mytek. The DAC has a very good 10ps internal generator 1/2 inch from DAC chip and this low jitter figure is impossible to achieve using external clocks and all the cables involved.

Also some clock vendors offer "atomic clock" as I suppose "atomic" sounds very marketable, but all this is also somewhat dubious. In fact this clock:

https://en.wikipedia.org/wiki/Atomic_clock

is really useful for long term accuracy of time measurement, does not really have any advantages when clocking audio data."

He adds:
"Designers of ADC and DACs are in business of providing most accurate digital conversion as oppose to accurate time measurement over long term.

If one is concerned with their song length that has to be, let's say, exactly 3 mins 00 sec 00 ns long with 0.001%, then they should consider atomic clock.  Everybody else concerned with sound quality should rather be looking at digital converters with built in low jitter oscillators. PPM spec means nothing in terms of clock usefulness for audio conversion.

Oven cooked crystal or atomic oscillators make a great marketing pitch but they are intended for scientific time measurement and have no impact  audio quality.  It's the short term periodic (phase) jitter that matters most and it's not atomic source or thermal stability that assures this, but a crystal designed specifically for low phase jitter.  These designs are different, they are low noise, not low ppm.

In our (Mytek) designs) we have used oscillators with down to 1ps jitter specs. Such specs can only be achieved by placing the xco very close (few mms) to converter chipset and by using special buffering and pcb layout techniques, dedicated power supplies etc. 1psec crystal is expensive, can cost $50 as compared to a $2 for 50psec but this is the only approach that makes a difference IMO. A digital synthesis system can also be designed to have  low periodic jitter.

There is no way that the most stable external clock box can provide psec jitter figures when feeding another box through dubious few ft long wordclock connection and associated buffer circuitry , grounds and power supplies on both ends.

It's unfortunate the myths about clocks have perpetuated so deeply despite easily demonstrable evidence (for example see: http://www.soundonsound.com/sos/jun10/articles/masterclocks.htm)

When someone tells me their system sounds "better" with external clock I challenge them to run a sinewave distortion FFT. In most cases this test shows more distortion on external clock (as it should be expected when high end ADC/DACs are involved).

Yet these distortion are often perceived as audible "improvement".

When somebody tells me "I connected the "Galactic" External Clock Machine and  all started to sound great, vocal came to front etc etc.",  I know they are talking about increased midrange grunge coloration caused by more jitter. I'd rather use a nice sounding analog gear to color the sound, doing the same with clock is possible but probably misguided.

Having said that there is some badly designed equipment out there that can benefit from external clock but this should be an exception, not the rule, they are just badly designed. External clock  may help a Digi box where a 50Mhz clock is sent over a foot of ribbon, but won't help our Mytek nor Prism nor Lavry.

FWIW Mytek makes a good standalone clock oscillator. We recommend using this clock as a central source to sync a large system, not as an external silver bullet box that would "improve" the sound through some psychoacoustic magic that cannot be scientifically explained.

My 2C

Michal at Mytek New York".

I hope this clears things up a bit.

Best regards
Adam Bielewicz