Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[marek]

I am only entering JRiver’s world and this is my first post so I’ll start by saying “Hello” to this great community.

I have been searching for an answer on Diyaudio forum but so far no one has come up with a solution. It may be that what I am trying to achieve is just my wishful thinking so hopefully you guys will be able to put me back on track.

Basically, I would like JRiver to act as a crossover which sends the digital signal to two different (!) DACs. I do not want to use any extra crossover like nanoDIGI or DCX2496. It is only 2-way setup, so 4 channels, 2 for subs, 2 for mains, with cut off point at 200Hz, I intend to use DSP engine to add EQ to the subs, and be able to stream the signal up to 192Khz to 2 totally different DACs. I should be able to measure the delay/phase of each channel and adjust the settings to have perfect alignment.

Is it possible? Or I will always have timing issues with 2 different DAC and there is no way to correct it?

Thanks for help guys!

Marek

[Hilton]

Hi I wouldn't recommend that solution anyway. You will get slight timing variation in the different DAC clocks causing phase issues and god knows what else. (as you've probably already been told)
You may as well go back to a well constructed analog crossover or get a DAC that can do what you want from a single clock source.

There's plenty of USB DACs and sound cards with 4+ channels that could be adapted for this. You'd just need to do some re-routing with the MC DSP.

[jjazdk]

As for what I know, the only way to do what you want is to use ASIO4ALL.

This can present the two DACS as a single device to JRiver, which is paramount for what you want to achieve.

[marek]

Hi I wouldn't recommend that solution anyway. You will get slight timing variation in the different DAC clocks causing phase issues and god knows what else.

Thanks, it is exactly the timing issues that I am worried about.

[marek]

As for what I know, the only way to do what you want is to use ASIO4ALL.

This can present the two DACS as a single device to JRiver, which is paramount for what you want to achieve.

Thanks. So if I understand ASIO4ALL will make sure that regardless of amount of DSP for each section (subs or mains) the output from the PC will reach DACs at the same time. But will it force the DACs to work in slave mode? I am worried that I can still have the buffer in each DAC have different delay and this could possibly get worse with hi-res files. No idea how it would work.

[Hilton]

this >

still have the buffer in each DAC have different delay

will still be a problem unless you get expensive DACs with an external clock.

[mattkhan]

Thanks. So if I understand ASIO4ALL will make sure that regardless of amount of DSP for each section (subs or mains) the output from the PC will reach DACs at the same time. But will it force the DACs to work in slave mode? I am worried that I can still have the buffer in each DAC have different delay and this could possibly get worse with hi-res files. No idea how it would work.

You need a DAC that can sync to external word clock. If you don't have one then just buy a new DAC with sufficient channels. I don't know what your budget is but Focusrite so decent ones at the budget end, eg Scarlett 8i6. Lots of choices in the pro audio world though to suit all budgets.

[jjazdk]

this >  will still be a problem unless you get expensive DACs with an external clock.

The question remains, wether this timing issue is an issue at all, that depends on the specific DACs...

We don't even know which DACs the OP is using, or via type of interface they are connected to the PC.

[Hilton]

Here's some good advice in an old thread that still applies.

http://yabb.jriver.com/interact/index.php?topic=69858.0

[mattkhan]

The question remains, wether this timing issue is an issue at all, that depends on the specific DACs...

We don't even know which DACs the OP is using, or via type of interface they are connected to the PC.

if they are separate devices then they will drift over time. If you allocated one DAC to one 2 way speaker and the other to the other one then you might get results that are ok for the duration of a song (or you might not  ) but the likelihood is it will drift further and further over time.

[jjazdk]

I am not disagreeing that there are potential timing issues that can make the setup unusable, but as long as we do not know what equipment the OP is using, all we can do is try to make our best educated guesses.

With two USB soundcards, driven by two separate clocks, yes they will drift apart at an unknown speed. But perhaps that is not what the OP has.

if they are separate devices then they will drift over time. If you allocated one DAC to one 2 way speaker and the other to the other one then you might get results that are ok for the duration of a song (or you might not  ) but the likelihood is it will drift further and further over time.

[marek]

Fair point though I could ask if it is at all possible to avoid timing issues with two different DACs?
My main DAC is a semi diy construction based on es9018 chipset. It takes spdif and also usb via Amanero module. The second DAC can be anything really. I have a modified DCX2496 and DEQ2496 which I can use as DAC or a usb sabre DAC with es9023 chipset. I could even buy a new DAC but my main one must stay.

[jjazdk]

Fair point though I could ask if it is at all possible to avoid timing issues with two different DACs?
My main DAC is a semi diy construction based on es9018 chipset. It takes spdif and also usb via Amanero module. The second DAC can be anything really. I have a modified DCX2496 and DEQ2496 which I can use as DAC or a usb sabre DAC with es9023 chipset. I could even buy a new DAC but my main one must stay.

Yes, it is possible, IF you can syncronize the clocks.

The DEQ2496 has a wordclock input, perhaps you could get a wordclock output from your main DAC and use this for syncronization.

[marek]

Yes, it is possible, IF you can syncronize the clocks.

The DEQ2496 has a wordclock input, perhaps you could get a wordclock output from your main DAC and use this for syncronization.

That is a very interesting suggestion. Thanks. I need to investigate.

[mwillems]

You can also sync two DACs via SPDIF, ADAT or potentially I2S (you don't necessarily need wordclock).  So if you already have a DAC with an SPDIF input, you just need to make sure that your second DAC has an SPDIF output that's independently assignable from it's analog outputs (obviously).

The point is, however you link the DACs, you need to share clock information.  That means USB won't do it (since it doesn't carry a clock, and most DACs ignore USB timing information on purpose).  So the two DACs either a) need to be able to take an external clock signal separately (wordclock-style), or b) one DAC needs to be able to take a signal input that carries a clock (like SPDIF, ADAT, or I2S) and the other DAC needs to be able to send a signal that carries a clock.  The problem with option B is that it's not particularly scalable.  If you know you'll only ever want a bi-amped stereo pair, it's a good solution; if you ever want to move up to Tri-amp or a surround system with your bi-amped mains, you'll need a new solution and have to spend again.

A single multi-channel interface is still the best answer for most people.  But you can probably solve your immediate issue with the hardware you have on hand via SPDIF, and I can confirm from personal experience that daisy-chaining two DACs via SPDIF can work well enough for bi-amping.

[jjazdk]

Good input mwillems, which reminds me of my own solution.

I have a high-end (in hifi-terms) 2-channel DAC (SPDIF input) that I prefer to use for my tweeters and midrange, plus a 10-channel USB DAC (Focusrite Scarlett 18i20).

My solution to this is to use the SPDIF output from the Scarlett to feed my 2-channel DAC, thus I get the sonic benefit (that I believe in) from the 2-channel DAC and I avoid timing issues because the clock is common for both DACs (generated in the Scarlett DAC).

[marek]

Thank you guys for the suggestions. I understand spdif carries the clock but it also carries the signal. I think we are missing the crossover part in this setup, aren't we?

[mwillems]

Thank you guys for the suggestions. I understand spdif carries the clock but it also carries the signal. I think we are missing the crossover part in this setup, aren't we?

I'm not sure what you mean?  Once you have the sync sorted, you can just do an active crossover in JRiver the same way you do one with a multi-channel DAC.  Or is that what you're asking how to do?

If that's what you mean, the short answer is that you need to copy channels in JRiver's parametric EQ and then apply your crossover filters in the same place, set the delay etc.  Everything you need is in JRiver's PEQ blocks (assuming you have enough channels on your audio device(s).

[marek]

Sorry if I'm too confused mwillems.
I understand how your solution to sync 2 DACs via SPDIF would work when using a DAC with 2 inputs and kind of passthrough feature. But if I just resend the digital signal that the first DAC receives to the one that needs to be synced then in my view the second DAC would get the same information as the first one.
If I had an option to use one spdif input of my slave DAC purely for clock and other input to get the signal with applied DSP then I can see how this could work. I am not aware of a 2 channel DAC that can do it but I believe some exist.

In my setup my main DAC is based on ES9018 which does not even use master clock on I2S bus. The device itself has an internal clock which I could possibly use to send the signal to another DAC but it would require some tweaking as it doesn't have a digital output.

[mwillems]

Sorry if I'm too confused mwillems.
I understand how your solution to sync 2 DACs via SPDIF would work when using a DAC with 2 inputs and kind of passthrough feature. But if I just resend the digital signal that the first DAC receives to the one that needs to be synced then in my view the second DAC would get the same information as the first one.
If I had an option to use one spdif input of my slave DAC purely for clock and other input to get the signal with applied DSP then I can see how this could work. I am not aware of a 2 channel DAC that can do it but I believe some exist.

You may have missed part of my original post:

you just need to make sure that your second DAC has an SPDIF output that's independently assignable from it's analog outputs

 

So you need a DAC with at least two analog output channels and at least a single SPDIF output that can play something different than the analog outputs (is independently assignable).  Most such devices have a USB or ADAT input.

You'd then do your crossover filtering in JRiver and output 4 different channels audio from JRiver to the four output DAC; two channels would come out of the analog outs, and two channels would come out of the SPDIF output, which would go to your "main DAC" with the SPDIF input.  You'd then have four channels of analog output synced.  You won't be able to use your Main DAC as the clock "master" unless it has a digital output, it will have to be the clock "slave" of another DAC with a digital output.

It would look like this:

Code: [Select]

JRiver  -->USB?--> 4+ Channel Audio Device --> 2 channels analog
                                     |
                                     --> SPDIF --> Other DAC


An example:  I have a Steinberg UR824.  It has 8 analog outputs, and 2 SPDIF outputs (each representing two channels) for 12 total channels out (it will do more channels with ADAT, but let's bracket that out it's not relevant).  I can assign different audio in JRiver to all 12 output channels.  But to take advantage of those extra 4 SPDIF channels I'd need to have DACs with SPDIF inputs on the other end.  Those DACs will play whatever I assign to their channels in JRiver, so they don't need to be the same as my 8 analog channels.

If you're existing devices just have SPDIF outputs as a "passthrough" of their digital input, that won't help you any.  This is why I pointed out above that a single multichannel interface is usually much easier to manage.

[marek]

Mwillems, thank you!!! I indeed missed the important part of your first message but now everything is very clear. I know my options and limitations of my setup and if I decide to fight with two different DACs I will report on my results here.

BTW, your rePhase guide is fantastic. Big thumbs up for that too!

[pschelbert]

I had the same prblem. Either you use a multichannel audiointerface you find such with 12 to 26 channels. Lynx, RME or others. Look for pro audio not HiFi. They are relatively cheap. I use a RME UFX with 12 channels DAC. I can serve 4 speakers with each 3-way. To extend that the only viable solution I found is using Adat interface of the UFX. There are 2 Adat with each 8 channels at 44k, 2x4 channels at 88 or 96k 2x2 channels at 176 or 192 channels. With these outputs you enter the second DAC. Go with the same vendor for second DAC. For example RME with Motu does not work chaining via Adat. If you choose two UFX you get 16 or 8 or 4 additional channels out. Timing I do not see a problem. However you can measure and compensate.

[jjazdk]

That is not a very good advice if you read the original post, he want to use his existing DAC which do not feature ADAT input.

A relatively cheap solution would be to invest in a Focusrite Scarlett 6i6 http://us.focusrite.com/usb-audio-interfaces/scarlett-6i6 , as this features line-outs (4 channels) as well as SPDIF out (2 channels).

This would be usable with his existing DAC and ensure correct timing.

I had the same prblem. Either you use a multichannek audiointerface you fet such with 12 to 26 channels. Lynx, RME or others. Look fo pro audio ni HiFi. They are relatively cheap. I use a RME UFX with 12 channels DAC. I can 4 speakers with each 3way. To extend that the only viable solution i found is using adat interface of the UFX. There are 2 adat with each 8 channels at 44k, 2x4 channels at 88 or 96k 2x2 channels at 176 or 192 channels. With these outputs you enter the second DAC.

[marek]

I was able to find some info that may sadly put an end to my attempt. Because ES9018 only reclocks the signal the only way to synchronize the clocks would be to use another DAC with the same chip. DACs based on this chip are getting cheaper but I would probably need to fiddle with soldering on the PCB instead of using a "normal" cable. I am also not sure how I could possibly send the same signal to two DACs as SPDIF from one of them is not an option.

[marek]

Looks like I will use the same DACs. Is ASIO4ALL the only way to achieve synchronization on the PC output?

[jjazdk]

Looks like I will use the same DACs. Is ASIO4ALL the only way to achieve synchronization on the PC output?

ASIO4ALL is not to achieve a synchronized output from the PC, but to present both DAC's as one single soundcard to JRiver.

That said, using the same ESS chip for both DACs does not ensure synchronized output, what is important is that they share the same clock.

Why do you insist on a DAC with this particular ESS chip?

[marek]

ASIO4ALL is not to achieve a synchronized output from the PC, but to present both DAC's as one single soundcard to JRiver.

That said, using the same ESS chip for both DACs does not ensure synchronized output, what is important is that they share the same clock.

Why do you insist on a DAC with this particular ESS chip?

I'm not sure what advantage will I get from having 1 sound card in JRiver. I should be able to play HP to one card and LP to another. Correct me if I'm wrong please.

That's correct about the same clock for this particular chip. That is why I may need to look for other chips as ES9018 uses the clock differently to other DACs. My main DAC based on this chip sounds ridiculously well and that was the reason why I tried to use it.

[jjazdk]

I'm not sure what advantage will I get from having 1 sound card in JRiver. I should be able to play HP to one card and LP to another. Correct me if I'm wrong please.

Unless you want to fiddle around with enabling multiple zones at once (each zone on its own soundcard), it is MUCH easier to have one soundcard presented in JRiver and play everything out of one device (one device seen from JRiver point of view that is).

That's correct about the same clock for this particular chip. That is why I may need to look for other chips as ES9018 uses the clock differently to other DACs. My main DAC based on this chip sounds ridiculously well and that was the reason why I tried to use it.

Well, then I strongly suggest you buy something like the Scarlett 6i6 and hook your ES9018 based DAC on the SPDIF output from the Scarlett DAC. This will solve your issue for less than 250USD and even give you a few extra outputs for future use 

[mwillems]

I'm not sure what advantage will I get from having 1 sound card in JRiver. I should be able to play HP to one card and LP to another. Correct me if I'm wrong please.

That's correct about the same clock for this particular chip. That is why I may need to look for other chips as ES9018 uses the clock differently to other DACs. My main DAC based on this chip sounds ridiculously well and that was the reason why I tried to use it.

The advantage of a single sound card is that it will actually work.  Two DACs won't work unless they share the same clock source, and we might be talking past each other a bit.

It doesn't matter that the two chips do their internal clocking the same way; Two chips with the same adaptive reclocking mechanism will fail to sync because they will both be reclocking the signal separately.  I've tested this by trying to sync two of the same model of USB DACs.  It did not work.

For sync to happen, the devices need to be connected to each other in such a way that they share the same reference clock at the same time.  But based on your description, your DAC can only take a clocked signal via SPDIF.  So you'll either need to daisy chain it in the way I described above, or find a multi-channel solution.

ASIO4ALL is not to achieve a synchronized output from the PC, but to present both DAC's as one single soundcard to JRiver.

That said, using the same ESS chip for both DACs does not ensure synchronized output, what is important is that they share the same clock.

Why do you insist on a DAC with this particular ESS chip?

+1 ASIO4All will not usually get you synchronized playback, and it's quite buggy to boot.

[marek]

It doesn't matter that the two chips do their internal clocking the same way; Two chips with the same adaptive reclocking mechanism will fail to sync because they will both be reclocking the signal separately.  I've tested this by trying to sync two of the same model of USB DACs.  It did not work.

For sync to happen, the devices need to be connected to each other in such a way that they share the same reference clock at the same time.  But based on your description, your DAC can only take a clocked signal via SPDIF.  So you'll either need to daisy chain it in the way I described above, or find a multi-channel solution.

My DAC actually can NOT take any clocked signal via SPDIF because it uses the clock for upsampling and reclocking only. It is theoretically possible to provide the same clocked signal to two DACs but I could not find any example that anyone succeeded. That is why I am not going to proceed with trying to make it work for this particular chip. But for others it should be relatively straightforward as long as DACs can work in slave mode and accept synchronous signal. Asynchronous DACs as you mentioned may still have timing issues.

[marek]

Unless you want to fiddle around with enabling multiple zones at once (each zone on its own soundcard), it is MUCH easier to have one soundcard presented in JRiver and play everything out of one device (one device seen from JRiver point of view that is).

Well, then I strongly suggest you buy something like the Scarlett 6i6 and hook your ES9018 based DAC on the SPDIF output from the Scarlett DAC. This will solve your issue for less than 250USD and even give you a few extra outputs for future use 

Thanks for the hint regarding zones. I don't want to complicate software side and will try to get the DACs as one soundcard.

Re Scarlett, it will still not work with ES9018. This chip does not accept any clock signal over I2S/SPDIF.

[jjazdk]

My DAC actually can NOT take any clocked signal via SPDIF because it uses the clock for upsampling and reclocking only. It is theoretically possible to provide the same clocked signal to two DACs but I could not find any example that anyone succeeded. That is why I am not going to proceed with trying to make it work for this particular chip. But for others it should be relatively straightforward as long as DACs can work in slave mode and accept synchronous signal. Asynchronous DACs as you mentioned may still have timing issues.

I think you are misunderstanding the operation of your ES9018 based DAC.

If it takes a SPDIF signal as input (I assume it does, as you have not mentioned it to be a USB DAC), it WILL work if you use the SPDIF output from another converter (be it a Scarlett or something else).

Your DAC upsamples and reclocks, but since it is fed from a SPDIF stream it must not run away from the clock from the SPDIF, otherwise you would get drop outs. Asynchronous upsampling still has to follow the source clock, otherwise it cannot play from a SPDIF input.

[marek]

Your DAC upsamples and reclocks, but since it is fed from a SPDIF stream it must not run away from the clock from the SPDIF, otherwise you would get drop outs. Asynchronous upsampling still has to follow the source clock, otherwise it cannot play from a SPDIF input.

I wish it worked the way you say it but please see the white paper from ESS:

"The SPDIF interface is more complex than the DSD and I 2 S since it must first derive the embedded clock in the bi-phase encoded data. In fact, experience with many forms of SPDIF decoder suggest that most fail in the presence of high jitter due to the lack of robustness in the clock recovery process. To avoid this potential problem the Sabre SPDIF interface avoids having to extract the clock at all: decoding is done using a method that does not require an explicit measure of the clock frequency. Specifically, the digital input is first corrected for 50% duty cycle by means of a discrete digital delay line that is able to delay either the positive edge or the negative edge of the signal such that after this delay line the signal is at 50% duty cycle. Thereafter an assessment is made of the width of each pulse based on its relation to recently seen pulse widths and a decision circuit assigns each a width of 1, 2, or 3 units. A state machine then operates on the assigned widths in succession; this state machine is searching for the block boundaries and the bit states. The state machine makes no attempt to re-time or otherwise decode the clock – it simply “time stamps” the event and passes it to the downstream processor. Using this method the SPDIF interface is able to accommodate 50nS of random jitter and 200nS of sinusoidal jitter in the incoming data."

[jjazdk]

If you read what it does, it is obvious that it must adhere to the source clock. Otherwise it would either run out of data or have to throw data away.

Yes, it reclocks, but the new clock MUST follow an average of the source clock, it is simply not possible for it to drift away long term.

[JAVA Alive]

Hi all,
 Any progress from any one on this question ?
 Anyone implemented this solution and has a feed back ?

 Regarding delay in Dacs, I think digital links are not equal and what we call "reclock" can be done in many different ways and have different impacts on delay.
 A very common reclock function (for spdif for example) is done with PLLs. By construction, PLLs will not generate variable delays as they follow initial clock speed. They can generate a very small delay for jitter correction but at a very small scale with no impact on phase respect.
 Asynchronous USB links are probably different. i don't exacltly know how they work but as far as I know they must have a kind of memory buffer in order to operate asynch. And it could effectivly generate variable delay in playback which is a huge issue for phase respect.

 

[jan.didden]

Hi,hoping you don't mind to resurrect this thread.
I'm also thinking about using two different DACs in a multiway setup.
I read this thread (and another related one) and there seems to be a possible issue with the two DAC having different timings and drifting apart.
But is that the case? Presumably, both DACs are connected to the same PC on two different USB ports, but they should receive the same timing and would sync to the same USB embedded clock.
So there should be no issue with different timings.

Am I correct?

Jan

[eve]

Hi,hoping you don't mind to resurrect this thread.
I'm also thinking about using two different DACs in a multiway setup.
I read this thread (and another related one) and there seems to be a possible issue with the two DAC having different timings and drifting apart.
But is that the case? Presumably, both DACs are connected to the same PC on two different USB ports, but they should receive the same timing and would sync to the same USB embedded clock.
So there should be no issue with different timings.

Am I correct?

Jan

That's unfortunately wrong.
Assuming we're talking about standard USB D/A. Each USB DAC is getting an asynchronous stream of audio data, and using it's own internal clock.

[jan.didden]

OK. Bummer.

But I went back over the discussion and I think I can use my RME ADI-2 Pro to provide 4 analog output channels (1 to 4) while outputting channels 5 and 6 over S/PDIF to another 2 channel DAC (I only need 6 channels). That should work and remain synched, right?
I'll figure that out and report back.
Thanks guys,

Jan

[eve]

OK. Bummer.

But I went back over the discussion and I think I can use my RME ADI-2 Pro to provide 4 analog output channels (1 to 4) while outputting channels 5 and 6 over S/PDIF to another 2 channel DAC (I only need 6 channels). That should work and remain synched, right?
I'll figure that out and report back.
Thanks guys,

Jan

SPDIF out on it's own will use the PLL clock recovery of the connected DAC. It's not the exact same 'clock'. The differences may be imperceptible or at least, not serious enough.