SMSL G1 Review (Clock Generator)

[FERCSA]

Wondering if you'd use chain connection instead of star would the result be the same?

I know star connection is recommended and a very long chain can degrade the signal but syncing more DACs(4+) with this generator would be an interesting thing to try as well or at least knowing that the possibility is there.

 

[DWPress]

Great review and dive into clocks @NTTY. I'd be interested to see a pro audio reviewer take this on - SMSL made this as a dubious addition to home audio but the pro's need this sort of device to keep everything synced. My Okto DAC8-PRO has 4 digital AES inputs but all incoming clocks beyond input #1 are synced to the first to handle this sort of thing in home/pro environments.

 

[GPx86]

Could these performance measures vary/degrade over time and allow the performance of the CDP/CDT to drift? Is there a useful scenario where the clock generator could keep the CDP/CDT functioning _perfectly_ for a longer period of time?

 

[wwenze]

Great review and dive into clocks @NTTY. I'd be interested to see a pro audio reviewer take this on - SMSL made this as a dubious addition to home audio but the pro's need this sort of device to keep everything synced. My Okto DAC8-PRO has 4 digital AES inputs but all incoming clocks beyond input #1 are synced to the first to handle this sort of thing in home/pro environments.

Pro need a master clock because if you run for example your DAW on its internal clock and your have an SPDIF-in signal, you are going to desync and get a periodic noise. We solve this problem by getting the clock from the SPDIF connection, because the signal from that SPDIF connection will be synced to the clock from that SPDIF connection, but what if you have two SPDIF-in? (I guess one solution is to just ASRC everything, but you don't do that in production environment) That's why every equipment must share the same clock and send out their data synced to that clock. The consumer user doesn't see this issue because their audio chain is just a one-direction line often with ASRC thrown in somewhere, but pro audio goes in loops, maybe grids, where you cannot resample the original signal.

Now, this just means you need someone to be the master clock, and that can be and often is the main DAC or soundcard. You don't need to buy anything to keep everythng synced. So the value proposition becomes just how much you want a more accurate clock.

 

[antcollinet]

Could these performance measures vary/degrade over time and allow the performance of the CDP/CDT to drift? Is there a useful scenario where the clock generator could keep the CDP/CDT functioning _perfectly_ for a longer period of time?

Small amounts of clock drift don't matter - since all devices are receiving the same clock - everything drifts the same amount - everything stays in sync.

 

[DWPress]

Small amounts of clock drift don't matter - since all devices are receiving the same clock - everything drifts the same amount - everything stays in sync.

Unless you try to use multiple 2ch DACs (even same make and model) as an amalgamation for active multiway speakers as has been demonstrated repeatedly and why we have multichannel devices. Used in this way the drift does become apparent over time.

 

[Hayabusa]

I dont get the pitch measurements on the cd players with external dacs: if the the CD player is not suplied an external clock its digital output is the master clock, always...
So the pitch should not change at all.
Except if you use a fifo with reclocker, then you can push your new clock on the output of the CD player.

 

[AnalogSteph]

I dont get the pitch measurements on the cd players with external dacs: if the the CD player is not suplied an external clock its digital output is the master clock, always...
So the pitch should not change at all.

That's what puzzles me as well.

I can only surmise that the difference in input and output clock would eventually result in a buffer overflow and hence, a brief audio glitch. In case of the Revox there's about 5.5 excess samples per second if I'm not mistaken. If there is a large(ish) buffer it may take a good while before the glitches set in.

You generally can't take up such rate differences without an ASRC, and then the pitch would still remain constant as mentioned.

 

[wwenze]

You generally can't take up such rate differences without an ASRC, and then the pitch would still remain constant as mentioned.

We can surmise ASRC is already present... because that is usually what good DACs do.

 

[Martin]

The differing fonts on the displays of this stack bug the shit out of me...

Followed very closely by the slight color difference of the top unit vs. the bottom two.

Martin

 

[Russ_L]

10mhz. WTF.

Russ

 

[milosz]

This doesn't look useful for home music listeners, but I think in some prosumer home studios it could be useful.

 

[AnalogSteph]

We can surmise ASRC is already present... because that is usually what good DACs do.

If so, it's clearly not present where it matters.

The D200 appears to be a BD34352EKV job, so there's none in the DAC for starters.

10mhz. WTF.

10 MHz is pretty much the standard for lab reference oscillators.

Side note, If I wanted to determine the frequency inaccuracy of the unit under test, I would obtain another TCXO or OCXO that can be disciplined externally and sync it with a GPSDO. (Note, GPSDOs can take a good while before they're accurate.) Then use one to supply the local oscillator for a direct conversion receiver and feed the other into the receiver input, resulting in the frequency difference appearing on the output. (Said direct conversion receiver could be quite basic, I would think that not terribly much besides a 74HC4066 based sampling mixer plus an amplifier / limiter for the LO should be required. If there is a full-blown shortwave communications receiver or transceiver with a 10 MHz reference input already present, I'd gladly use that instead, obviously.)

 

[EXIF68]

The question is: Can you hear it?
I think not.

 

[solderdude]

Is the clock of the measurement device calibrated ?

Other than the PL200 none of them have an external clock input.
So it looks like we are simply looking at the ability of the DAC to not synchronize to the incoming clock derived from the SPDIF but probably has a buffered input (with large memory to pull this off) and reclocked the incoming data to that of the external clock. When the buffer is large enough and the deviation of both clocks is not too large this should work problem free without having to skip a sample value now and then.
Afterall... when the CDP does not have the external (10MHz) clock input the clocks from the CDP and and DAC are always different and the external clock has no actual benefit.
So it looks like the only thing the external clock does here is force the D200 to buffer the incoming SPDIF data (with its embedded clock) and tell the DAC to use the external clock instead of 'follow' the average speed of the derived clock (the one from the incoming SPDIF) which requires more buffering and prevents buffer overflow and thus missing samples once in a while.

An external (word)clock is essential in studios where all recording, playback and digital mixer stuff is synchronized in sample timing.

For home audio this accuracy nor an external wordclock isn't really needed.

 

[Mnyb]

Pro need a master clock because if you run for example your DAW on its internal clock and your have an SPDIF-in signal, you are going to desync and get a periodic noise. We solve this problem by getting the clock from the SPDIF connection, because the signal from that SPDIF connection will be synced to the clock from that SPDIF connection, but what if you have two SPDIF-in? (I guess one solution is to just ASRC everything, but you don't do that in production environment) That's why every equipment must share the same clock and send out their data synced to that clock. The consumer user doesn't see this issue because their audio chain is just a one-direction line often with ASRC thrown in somewhere, but pro audio goes in loops, maybe grids, where you cannot resample the original signal.

Now, this just means you need someone to be the master clock, and that can be and often is the main DAC or soundcard. You don't need to buy anything to keep everythng synced. So the value proposition becomes just how much you want a more accurate clock.

Good explanation +1 my old man was an video editor and general sound technician etc . Their studio had a clock generator even to the old video players etc even back into 80’s when they where partly analog , the frames of the analog video tape machines did not align without time sync you don’t want a line slowly drifting across your screen when fading between two video sources .

 

[Jimster480]

Very interesting review and deep dive. I haven't ever used an external clock generator, and I see that one is usually not needed. Which isn't hard to understand when looking at the performance of modern products. However it is definitely a cool product in the niche scenario when it is needed.

 

[BeeKay]

The differing fonts on the displays of this stack bug the shit out of me...

Followed very closely by the slight color difference of the top unit vs. the bottom two.

Martin

Not shure where to put the non removable MQA logo in your list

 

[NTTY]

I dont get the pitch measurements on the cd players with external dacs: if the the CD player is not suplied an external clock its digital output is the master clock, always...
So the pitch should not change at all.
Except if you use a fifo with reclocker, then you can push your new clock on the output of the CD player.

My understanding is that the PLL's output phase noise (of the DAC), that synchronizes and extracts the clock from the SPDIF, is determined by the reference clock's phase noise (scaled by the frequency multiplication factor, N). A cleaner reference clock results in a cleaner output clock signal from the PLL. That is why we get the same pitch out of all players.

The true advantage of that master clock is therefore in that it attenuates very well frequency offset. That’s related to the spec given in -dBc/Hz @ offset frequency. This is not a spec provided by the DAC vendors, so it’s difficult to compare, maybe by looking at the specs of their individual clocks. But I guess that’s the explanation.

 

[G|force]

This is a discontinued Apogee clock that I use to sync aggregate DACs, ADCs and mixer with my DAW. (not my pic)
It doesn't provide any sound quality benefits in and of itself.