Which DACs or Renderers have external clock inputs?

[mansr]

41 minutes ago, One and a half said:

omg, you do?

I believe he does.

[plissken]

2 hours ago, One and a half said:

omg, you do?

 

Actually I do. I spent 7 years engineering suites that typically housed 3 to 5 million in equipment. From D1 decks ($50,000 to $75,000 a pop), Video consoles typically $150,000 to $250,000 a pop.

 

Cameras that were $100,000 a pop.

 

And we aren't even talking the post production suites.

 

If you are doing live broadcast you will have a need. Other than that if the manufacturer can't design a good local clock what makes you think they can design an external TBC loop input correctly?

[jabbr]

17 hours ago, Ralf11 said:

How do you discipline a crystal oscillator?

 

Whips?

 

There are two entirely distinct parameters that get mixed up when people try to decide "which clock is best?"

 

1) absolute long term frequency accuracy

2) (short term) frequency accuracy aka phase error

 

The phase stability/error is a fundamental property of the crystal itself, related to purity etc. Techniques to generate frequencies, stabilize frequencies with PLL etc are limited in their ability to improve phase error because of the need to "lock" onto a frequency range -- the phase error under the corner frequency is not improved.

 

PLL / VCXO can, however "tune" a crystal frequency and improve the long term stability. That's a type of discipline -- folks seem locked on "rubium" and "atomic" which are entirely irrelevant for close-in phase error. The very best close-in phase error is achieved when the crystal is right next to/immediately adjacent to the DAC latch.

[jabbr]

18 hours ago, GUTB said:

Clock quality ranking.

 

King:

1. Cesium Beam

2. Rubidium

 

Good:

3. OCXO

4. TCXO

 

Poor:

5. VCXO

 

Trash:

6. Non-XO timing schemes.

What parameter are you using to define "quality"?

 

a) personal opinion?

b) something you've read? what?

c) a measurement? what?

 

Cesium beam is of course not a crystal oscillator, so would also fit under "non-XO timing schemes". Best or trash?

I have extensively discussed using the close-in phase error aka the "linewidth" as the best way to measure clock quality. Apropos cesium: http://tf.boulder.nist.gov/general/pdf/101.pdf, http://geodesy.unr.edu/hanspeterplag/library/geodesy/time/met5_3_S10.pdf

[Ralf11]

4 hours ago, jabbr said:

 

There are two entirely distinct parameters that get mixed up when people try to decide "which clock is best?"

 

1) absolute long term frequency accuracy

2) (short term) frequency accuracy aka phase error

 

The phase stability/error is a fundamental property of the crystal itself, related to purity etc. Techniques to generate frequencies, stabilize frequencies with PLL etc are limited in their ability to improve phase error because of the need to "lock" onto a frequency range -- the phase error under the corner frequency is not improved.

 

PLL / VCXO can, however "tune" a crystal frequency and improve the long term stability. That's a type of discipline -- folks seem locked on "rubium" and "atomic" which are entirely irrelevant for close-in phase error. The very best close-in phase error is achieved when the crystal is right next to/immediately adjacent to the DAC latch.

 

Is absolute long term frequency accuracy important for SQ?  Or is it an issue mainly for instrumentation?

[mansr]

8 minutes ago, Ralf11 said:

Is absolute long term frequency accuracy important for SQ?

Not in the slightest.

[GUTB]

Atomic clocks (Rubidium and Cesium) have the absolute highest accuracy and stability. It is it what runs major public infrastructures and military systems. Simple XOs using electronic management methods don't even come close.

 

Now, when we're talking about audio, electronic circuits -- and our minds -- are excruciatingly receptive to noise. Those atomic clock physics packages are noisy, so a great deal of care is needed in implementing such clocks into an audio system. Antelope's 10M atomic clock is one example of a successful integration. Now, it's a very valid consideration wether that level of accuracy is simply overkill in audio or not, or if noise considerations out-weigh the gains of improved accuracy.

[mansr]

4 minutes ago, GUTB said:

Atomic clocks (Rubidium and Cesium) have the absolute highest accuracy and stability. It is it what runs major public infrastructures and military systems. Simple XOs using electronic management methods don't even come close.

Atomic clocks are simple XOs using electronic management methods.

[DomieMic65]

Two relatively affortable options

M2Tech EVO DAC Two Plus

&

EVO Clock Two

[jabbr]

16 hours ago, GUTB said:

Now, it's a very valid consideration wether that level of accuracy is simply overkill in audio or not, or if noise considerations out-weigh the gains of improved accuracy.

 

Its not a matter of accuracy "overkill" rather that absolute accuracy is the entirely wrong specification important to audio. 

 

Short term frequency fluctuation, aka phase noise, is far far far far far more important for audio recording and reproduction. In this setting noise considerations are paramount. 

[austinpop]

6 hours ago, DomieMic65 said:

EVO Clock Two

 

This is a meaningless spec:

• Phase noise:  -155dB (100kHz dev.)

More meaningful is the phase noise at a 1Hz or 10Hz deviation. That is where the best-sounding reference clocks truly stand out.

 

Does EVO publish that (phase noise at varying deviations)? I don't see it.

[DomieMic65]

1 hour ago, austinpop said:

 

This is a meaningless spec:

• Phase noise:  -155dB (100kHz dev.)

More meaningful is the phase noise at a 1Hz or 10Hz deviation. That is where the best-sounding reference clocks truly stand out.

 

Does EVO publish that (phase noise at varying deviations)? I don't see it.

i do not know but m2tech normaly responds quickly to any question...

 

[One and a half]

On 20/11/2017 at 12:41 AM, jabbr said:

 

Its not a matter of accuracy "overkill" rather that absolute accuracy is the entirely wrong specification important to audio. 

 

Short term frequency fluctuation, aka phase noise, is far far far far far more important for audio recording and reproduction. In this setting noise considerations are paramount. 

Ok I take this to be accurate. To maintain the low level of phase noise should the clock be stable? IOW does the specified phase noise spec a guaranteed value when the clock has stabilised, mainly reference to OXCO types. Does stability take overnight / days  to settle? 

[esldude]

25 minutes ago, One and a half said:

Ok I take this to be accurate. To maintain the low level of phase noise should the clock be stable? IOW does the specified phase noise spec a guaranteed value when the clock has stabilised, mainly reference to OXCO types. Does stability take overnight / days  to settle? 

Takes a few minutes at most.  

[mansr]

Lab equipment datasheets sometimes call for a warmup period of 15 minutes or so, occasionally as much as an hour, before the specified tolerances are met. There's of course a lot more than an oscillator involved in that.

[The Computer Audiophile]

On 11/18/2017 at 3:47 PM, GUTB said:

Atomic clocks (Rubidium and Cesium) have the absolute highest accuracy and stability. It is it what runs major public infrastructures and military systems. Simple XOs using electronic management methods don't even come close.

 

Now, when we're talking about audio, electronic circuits -- and our minds -- are excruciatingly receptive to noise. Those atomic clock physics packages are noisy, so a great deal of care is needed in implementing such clocks into an audio system. Antelope's 10M atomic clock is one example of a successful integration. Now, it's a very valid consideration wether that level of accuracy is simply overkill in audio or not, or if noise considerations out-weigh the gains of improved accuracy.

Atomic clocks are great if you need accuracy over days, weeks, months, etc...

 

In audio when we talk about nano, pico, and femto seconds an atomic clock is not that great. 

[Summit]

6 hours ago, The Computer Audiophile said:

Atomic clocks are great if you need accuracy over days, weeks, months, etc...

 

In audio when we talk about nano, pico, and femto seconds an atomic clock is not that great. 

Which ones have you tried?

http://www.theabsolutesound.com/articles/tas-180-esoteric-g-orb-rubidium-master-clock-generator-1/

http://www.thinksrs.com/products/PERF10.htm

[The Computer Audiophile]

I’ve tried several over the years. 

[Dev]

 

[GUTB]

Okay guys...looks like the absolute best is actually active hydrogen masers. They have ludicrously large stability numbers which I have no idea mean in the real world...like 10^14.

[rickca]

2 minutes ago, GUTB said:

hydrogen masers

It's a frickin' maser!

[jabbr]

8 hours ago, Dev said:

 

Excellent

[GUTB]

The TASCAM DA-3000 has a clock in...and does native DSD decoding via SDIF-3 inputs. Which brings to...

 

https://www.ebay.com/itm/Digital-interface-XMOS-USB-or-Amanero-USB-Module-Supports-DSD-R98/122209728498?hash=item1c744477f2:m:m18Pi4yhWbR-wdMSZAoXRow

[One and a half]

11 minutes ago, GUTB said:

The TASCAM DA-3000 has a clock in...and does native DSD decoding via SDIF-3 inputs. Which brings to...

 

https://www.ebay.com/itm/Digital-interface-XMOS-USB-or-Amanero-USB-Module-Supports-DSD-R98/122209728498?hash=item1c744477f2:m:m18Pi4yhWbR-wdMSZAoXRow

I have the Tascam DA-3000, and it works very well with MC-3+USB. 

 

On the ebay item, SDIF surely must be dead by now, unless there's a rack of Sony devices still about using the protocol. IIRC it only went as far as 48kHz for DAT purposes, or did it accept DSD64.

[GUTB]

3 minutes ago, One and a half said:

I have the Tascam DA-3000, and it works very well with MC-3+USB. 

 

On the ebay item, SDIF surely must be dead by now, unless there's a rack of Sony devices still about using the protocol. IIRC it only went as far as 48kHz for DAT purposes, or did it accept DSD64.

 

Sony allowed for DSD to be streamed un-encrypted across SDIF-3 links. If used both L and R lines together, you get DSD128 if I recall. The DA-3000 specifically skips SRC and does native DSD decoding if input through the SDIF-3 inputs.