Cheers. You wouldn't be the similarly named poster and mod at HifiHaven.org by any chance? If so I appreciated your extended comparison review of the Harmony and Cyan - thanks for that.
LAiV Harmony uDDC Re-clocker/DDC
- Thread starter catastrofe
- Start date
Sound Dragon
New Head-Fier
I am he. Glad it served you well.
I received and connected my LAiV ddc yesterday - my first impressions, straight out of the box: Absolutely stunning & beyond my expectations. Straight out of the box (as with my Harmony dac) it is superb, improving on every possible area of performance. Within seconds of playing the first (very familiar) track, even my wife noticed from the other end of the house how "tight" it was, with much more detail than ever before. It gives more attack, more detail - this "detail" being what leads to the big picture. We can't achieve the big picture without all the detail. Bass is fuller, deeper and tighter, treble is even more "sparkly" - brass instruments are more "brassy", cymbals are amazing and even more realistic. We can hear the speed of plucked strings - guitar, violins etc, and with all percussion instruments the attack is so realistic!
Pleased to see someone else (other than me!) using the TEAC VRDS 701 - I have the transport, big step up from my previous Audiolab 9000, which in turn was a step up from my Audiolab 6000. My 9000 is now for sale if anyone's interested!I bought my Denafrips gear back in February, but I researched it a few years. And how I'm also familiar with Alvin.
Anyway, had I been looking to get system now, I'd give the LAiV serious consideration as I love its looks, size and that the head amp is also a pre-amp.
The Denafrips Artemis is strictly a headphone amp so no outputs other than headphones. But man does it sound great with my Sennheiser HD800 S phones. It brings them alive, especially in the bass region. The amp is warm, but it brings this weight of notes to music that I love, especially with the Pontus feeding into it. This me....
Anyway, if I didn't have that, I'd seriously consider the LAIV setup. The only thing I don't like on it is the large display letters, nor the color. If they'd made the lettering blue and smaller, that setup in black would be killer. Wouldn't look so bad in silver either
1laraz
Head-Fier
Hi @GoldenSound , it is very interesting to know your expert opinion on how would the performance of internal word clocks in Holo May with a phase noise of -90 db@10Hz compare with some excellent external 10 Mhz clock with a phase noise of -140 db@10Hz that is connected to a DDC like the Audio-GD or LAiV uDDC reclocker?
As far as I understand there will be performance losses in form of additional jitter due to a longer path to the DAC (clock - BNC - DDC - I2S - DAC) as well as not perfect transition of 10Mhz signal to 44.1/48.
As far as I understand there will be performance losses in form of additional jitter due to a longer path to the DAC (clock - BNC - DDC - I2S - DAC) as well as not perfect transition of 10Mhz signal to 44.1/48.
Last edited:
Though well intended this question has the potential to derail this new thread with theoretical discussion and debate, so perhaps best to leave it/continue that discussion over in the Audio-Gd thread to leave space for new owners sharing their impressions here.
If you have a May, I'd strongly suggest just using USB and using the internal clocks, you get essentially perfect jitter performance that way as the May's clocking is extremely good. And whilst the uDDC does have internal galvanic isolation which can definitely be a benefit for a lot of DACs, the May has full isolation on the USB anyway so this isn't needed.
If you use an external DDC and leave PLL on you'll get basically identical performance to just using the internal clocks normally as the May has an extremely effective PLL. If you turn off the PLL the performance is dependent on the connected DDC as you're using the clock from that directly.
The uDDC has pretty excellent jitter performance, it's a very good DDC. And interestingly it is so far the ONLY device I've tested that actually sees an improvement in jitter performance when using an external 10Mhz clock, lowering jitter by about half. But it also seems that it may already be using a fractional PLL/clock synthesizer normally, which could explain why there is an improvement here, as this seems it could be a case of comparing a fractional PLL with the internal clock vs a fractional PLL with a better external clock.
This is a bit of a different situation to what you might have in most devices where you'd be comparing direct audio rate clocks vs a fractional PLL with an external clock.
In the majority of cases using an external 10Mhz clock you do indeed get worse jitter performance due to as you mentioned 10Mhz not being divisible by 48khz or 44.1khz. Gustard DACs for example have quite notably higher jitter when using the 10Mhz input.
This is the X26 Pro Jitter using normal USB input for example:
And then this is using a 10Mhz clock:
Quite a bit more!
The uDDC does seem to be doing some different stuff internally to most products though, it looks like it's using a 25.00Mhz clock internally so it's probably taking some quite different approaches especially since the device is heavily focused on synchronous inputs.
Measuring jitter directly, the uDDC has approx 2.5x the jitter of the Holo RED, and this drops to about 1.5x the jitter if you use the 10Mhz input.
We can use a clock divider, or in fact using the Cyan 2 as a clock divider since it has no PLL, we can visualise the jitter performance over I2S.
This is the uDDC (USB input) jitter feeding Cyan 2 via I2S:
If we add the 10Mhz clock to the uDDC, high frequency jitter components are removed, but we see a reasonable increase to low frequency phase noise (note the wider bottom of the 'stem' of the fundamental signal)
The uDDC does have quite a bit more jitter if using the synchronous inputs though, so stick to USB input if you can.
If you use the Cyan 2 USB input and internal clocking though you can get lower jitter and phase noise than with any DDC:
I need to do more testing on the uDDC though, only just got it in. I've not yet been able to determine if it's 100% bitperfect as I don't have a way to check this with it only having I2S output. (Managing a synchronous clock from the input AND a separate 10Mhz reference can be tricky without employing ASRC)
The fact that it has AES/SPDIF and even I2S input though is pretty cool, hardly anything else on the market that has that as an option. So as a device to clean up/reclock not so ideal digital sources like maybe some CD transports, TVs, or other stuff with only SPDIF out, it's fantastic.
On many modern DACs though (not just the Holo stuff), USB is often going to be the best option as shown above. This is typically because as you touched on in your message, even if you have a theoretically PERFECT clock source, a 10Mhz clock has to be used in conjunction with a PLL + local clock anyway, so it's not 'replacing' the clock signal in the same way you can with an actual audio rate clock via a dedicated word clock or over I2S. And also even if the clock itself is perfect, the inherent fact that you've got to go via multiple connectors, an external cable etc, means the clock transmission line is never going to be 100% perfect.
The ideal situation would be to have a good clock as physically close to the DAC itself on the PCB as possible in most cases.
If you have a May, I'd strongly suggest just using USB and using the internal clocks, you get essentially perfect jitter performance that way as the May's clocking is extremely good. And whilst the uDDC does have internal galvanic isolation which can definitely be a benefit for a lot of DACs, the May has full isolation on the USB anyway so this isn't needed.
If you use an external DDC and leave PLL on you'll get basically identical performance to just using the internal clocks normally as the May has an extremely effective PLL. If you turn off the PLL the performance is dependent on the connected DDC as you're using the clock from that directly.
The uDDC has pretty excellent jitter performance, it's a very good DDC. And interestingly it is so far the ONLY device I've tested that actually sees an improvement in jitter performance when using an external 10Mhz clock, lowering jitter by about half. But it also seems that it may already be using a fractional PLL/clock synthesizer normally, which could explain why there is an improvement here, as this seems it could be a case of comparing a fractional PLL with the internal clock vs a fractional PLL with a better external clock.
This is a bit of a different situation to what you might have in most devices where you'd be comparing direct audio rate clocks vs a fractional PLL with an external clock.
In the majority of cases using an external 10Mhz clock you do indeed get worse jitter performance due to as you mentioned 10Mhz not being divisible by 48khz or 44.1khz. Gustard DACs for example have quite notably higher jitter when using the 10Mhz input.
This is the X26 Pro Jitter using normal USB input for example:
And then this is using a 10Mhz clock:
Quite a bit more!
The uDDC does seem to be doing some different stuff internally to most products though, it looks like it's using a 25.00Mhz clock internally so it's probably taking some quite different approaches especially since the device is heavily focused on synchronous inputs.
Measuring jitter directly, the uDDC has approx 2.5x the jitter of the Holo RED, and this drops to about 1.5x the jitter if you use the 10Mhz input.
We can use a clock divider, or in fact using the Cyan 2 as a clock divider since it has no PLL, we can visualise the jitter performance over I2S.
This is the uDDC (USB input) jitter feeding Cyan 2 via I2S:
If we add the 10Mhz clock to the uDDC, high frequency jitter components are removed, but we see a reasonable increase to low frequency phase noise (note the wider bottom of the 'stem' of the fundamental signal)
The uDDC does have quite a bit more jitter if using the synchronous inputs though, so stick to USB input if you can.
If you use the Cyan 2 USB input and internal clocking though you can get lower jitter and phase noise than with any DDC:
I need to do more testing on the uDDC though, only just got it in. I've not yet been able to determine if it's 100% bitperfect as I don't have a way to check this with it only having I2S output. (Managing a synchronous clock from the input AND a separate 10Mhz reference can be tricky without employing ASRC)
The fact that it has AES/SPDIF and even I2S input though is pretty cool, hardly anything else on the market that has that as an option. So as a device to clean up/reclock not so ideal digital sources like maybe some CD transports, TVs, or other stuff with only SPDIF out, it's fantastic.
On many modern DACs though (not just the Holo stuff), USB is often going to be the best option as shown above. This is typically because as you touched on in your message, even if you have a theoretically PERFECT clock source, a 10Mhz clock has to be used in conjunction with a PLL + local clock anyway, so it's not 'replacing' the clock signal in the same way you can with an actual audio rate clock via a dedicated word clock or over I2S. And also even if the clock itself is perfect, the inherent fact that you've got to go via multiple connectors, an external cable etc, means the clock transmission line is never going to be 100% perfect.
The ideal situation would be to have a good clock as physically close to the DAC itself on the PCB as possible in most cases.
Thank you so much for your invaluable message, it contains a lot of important information. Allow me to ask you a silly question.
You wrote that "May has full isolation on the USB anyway".
My question is: does this mean that for May there is no difference between:
- The source of the digital signal (whether it is a laptop or, for example, Holo Red)
- The quality of the USB cable (budget or very expensive).
Thank you.
That's correct yep!
With USB there is no Jitter or any audio clock signal from the source at all, so that's not a factor with any dac running USB (unless it's an older one running UAC1.0). It's just dependent on the DACs internal clocking and USB implementation.
But noise can be a factor. Some DACs have filtering or some partial isolation, like using transformer based galvanic isolation, though this can still allow high frequency noise to pass through so isn't always a silver bullet.
The May uses full laser diode isolation on all lines, so no noise is able to pass through at all. In fact the connector doesn't even share a ground with the DAC or chassis (you may notice the usb connection intentionally has a gap around the connector to prevent it touching the chassis).
There is literally no electrical connection between the DAC and the USB source.
This means that it doesn't matter at all what the source is doing or if it's noisy/quiet. As long as the data is all intact the DAC performance is 100% identical
Last edited:
sajunky
Headphoneus Supremus
Question is: How low is a low frequency phase noise? I thought you don't measure phase noise vs. frequency with APxxx analyzer. Correct me if I am wrong...If we add the 10Mhz clock to the uDDC, high frequency jitter components are removed, but we see a reasonable increase to low frequency phase noise (note the wider bottom of the 'stem' of the fundamental signal)
Now related and critical for understanding what is measured: You said that a wider bottom of the 'stem' shows increased low frequency phase noise. Where you got such information? Source, please.
What I see on the plot is distribution of energy around fundamental signal in result of modulation. Wider stem represent larger deviation of base frequency, but there is no hint of the frequency of modulation.
Last edited:
Thanks. I mainly use mine in transport mode - Optical to DAC. Just use the built in DAC/Preamp section to drive my active speakers. Had I not got the deal I got on it, I would have just got the Transport, but the seller offered the higher end for the price of the Transport. Turns out I'm glad I did as I now don't need a pre-amp for my speakers.
I have a DDC feeding it and my Pontus - I2S to the Pontus, Optical to the 701.
pjw241142
100+ Head-Fier
Wow that's really technical above. Feel a bit thick TBH. What I distilled from above is focusing on the uDDC is best utilised as follows:
- USB input is best (I'm using a DMP-A8) vs I2S / Coax etc to feed the uDDC
- The (assumed) 25MhZ clock / FGPA chips etc set's it apart re jitter etc (to be explained how it works)
So my question to Laiv ( @wengfai ), there is a statement above as follows from @GoldenSound:
"The ideal situation would be to have a good clock as physically close to the DAC itself on the PCB as possible in most cases."
Are there demonstrable advantages of using the Laiv uDDC to Laiv Harmony DAC vs other DACs to achieve the above. I assume there must be and that a Chain consisting of:
- Source to (USB out):
- uDDC (USB in) / I2S out to
- Harmony DAC (I2S in) / XLR out to
- HP2A
Has inherent advantages as they are engineered together.
Re the Source, I note that the Laiv Net2 is due to drop in Q325
- USB input is best (I'm using a DMP-A8) vs I2S / Coax etc to feed the uDDC
- The (assumed) 25MhZ clock / FGPA chips etc set's it apart re jitter etc (to be explained how it works)
So my question to Laiv ( @wengfai ), there is a statement above as follows from @GoldenSound:
"The ideal situation would be to have a good clock as physically close to the DAC itself on the PCB as possible in most cases."
Are there demonstrable advantages of using the Laiv uDDC to Laiv Harmony DAC vs other DACs to achieve the above. I assume there must be and that a Chain consisting of:
- Source to (USB out):
- uDDC (USB in) / I2S out to
- Harmony DAC (I2S in) / XLR out to
- HP2A
Has inherent advantages as they are engineered together.
Re the Source, I note that the Laiv Net2 is due to drop in Q325
I think subjective listening sessions will be deciding here.
Lots of perfectly measuring DAC's sound terrible, and many bad measuring sounds great.. Take it as example
pjw241142
100+ Head-Fier
Page 4 has arrived on the 6Moons review of the uDDC. I can't work out whether it's a mickey take or whether it should have been:
1. At the bottom of Page 3 or
2. Dropped at the same time as Page 3.
Either way Page 4 adds very little to the sum total of humanity. Just saying
1. At the bottom of Page 3 or
2. Dropped at the same time as Page 3.
Either way Page 4 adds very little to the sum total of humanity. Just saying
Let me got study that page ..dictionary and thesaurus in handPage 4 has arrived on the 6Moons review of the uDDC. I can't work out whether it's a mickey take or whether it should have been:
1. At the bottom of Page 3 or
2. Dropped at the same time as Page 3.
Either way Page 4 adds very little to the sum total of humanity. Just saying
Users who are viewing this thread
