Can I run two Presonus iTwo boxes at the same time using ASIO??

[scook]

Just now, Terry Kelley said:

So the question is if ASIO will enable two of the same brand/model interfaces.

If not explicitly mentioned in the documentation, it is extremely unlikely the driver supports it.

If not documented, contact Presonus support for confirmation.

 

[John Vere]

I'm glad you brought this topic up because I had never done this and tried WASAPI Exclusive. I always use ASIO and just turn off the interface I'm not using. I have to keep my Scarlett alive as it has SPDIF and the Motu doesn't. Funny how WASAPI mode is calling the digital ins and outs analogue. But this gives me 8 analogue inputs if I need them without dusting off the Tascam. 

There are lots of small mixer around I used to have a little Yamaha too. I have the Mix 8 for the studio and then a Mackie Pro effects 4 that I used for my solo gigs. 

For it's size the Mix 8 does a lot. Especially useful is the Tape input which you can send to your headphones or to the Main outputs or both. 

If I want to do a loopback test all I need to do is push the Mains button if I have the mixer running back to inputs 3/4. 

I have used it in many configurations. I have a Lexicon MX 200 hooked up and when I was tracking clients I would use it to give them a different cue mix than I was hearing. As I said my Scarlett could do all that but was always confusing because of software. 

[Terry Kelley]

2 minutes ago, scook said:

If not explicitly mentioned in the documentation, it is extremely unlikely the driver supports it.

If not documented, contact Presonus support for confirmation.

 

I did and initially got a pondering "it should" but they passed me off to an "expert" who will call me back.

I was thinking about the master clock for each. Without having any knowledge about how they work, I assume the interface sends a clock to Cakewalk and Cakewalk follows that.  With a second interface, would it be sync'd via some signal over USB or could it free run if the clock speed was high enough (and drift would be minimal)?

Someone has to lead but that doesn't mean someone else couldn't be told to follow.

[Terry Kelley]

6 minutes ago, John Vere said:

I'm glad you brought this topic up because I had never done this and tried WASAPI Exclusive. I always use ASIO and just turn off the interface I'm not using. I have to keep my Scarlett alive as it has SPDIF and the Motu doesn't. Funny how WASAPI mode is calling the digital ins and outs analogue. But this gives me 8 analogue inputs if I need them without dusting off the Tascam. 

There are lots of small mixer around I used to have a little Yamaha too. I have the Mix 8 for the studio and then a Mackie Pro effects 4 that I used for my solo gigs. 

For it's size the Mix 8 does a lot. Especially useful is the Tape input which you can send to your headphones or to the Main outputs or both. 

If I want to do a loopback test all I need to do is push the Mains button if I have the mixer running back to inputs 3/4. 

I have used it in many configurations. I have a Lexicon MX 200 hooked up and when I was tracking clients I would use it to give them a different cue mix than I was hearing. As I said my Scarlett could do all that but was always confusing because of software. 

I have a bunch of mixers going as far back as Tascam 2A's but don't want to get one out and use up the space. This isn't the 1900's man!

I sent a PM on the Mackie Mix 8.

[John Vere]

Ya agreed, The SPDIF is for my Yamaha 01v digital mixer and it is just to big to keep set up, even though it blows the Mix 8 out of the water with options, the mix 8 is super small. I will someday build a under the desk slide out tray for the Yamaha. But the other issue it has is then I'm stuck with 44.1. 

[Noel Borthwick]

1 hour ago, Terry Kelley said:

But as you see in both ASIO and WASAPI Shared I can only use one or the other interface. 

But in WASAPI Exclusive I can easily use both.

Yes in WASAPI shared we disallow multiple interfaces because recording from multiple devices would be very complex to synchronize. 
In WASAPI shared mode the device returns variable size buffers so handling them and synchronizing them across multiple devices would be difficult to do. Its more deterministic and fixed size in WASAPI exclusive so we allow it there.

WASAPI shared mode is quite complex because you are really reading and writing data that has is processed by the Windows audio engine additionally, unlike exclusive mode where the app is directly communicating with the driver (similar to ASIO). This is how audio output from Cakewalk output can be mixed with other application streams in Windows  (via the Windows audio engine)
 

[msmcleod]

1 hour ago, Terry Kelley said:

I did and initially got a pondering "it should" but they passed me off to an "expert" who will call me back.

I was thinking about the master clock for each. Without having any knowledge about how they work, I assume the interface sends a clock to Cakewalk and Cakewalk follows that.  With a second interface, would it be sync'd via some signal over USB or could it free run if the clock speed was high enough (and drift would be minimal)?

Someone has to lead but that doesn't mean someone else couldn't be told to follow.

This is an important point, that azslow3 touched upon in an earlier post.

Even if you're able to record from multiple devices (either through ASIO if the drivers support it, or through WASAPI), the device's wordclock's still need to be synchronised to be sample accurate.

It'll still work if they're not, but they won't be sample accurate and you may find recordings from one device drifts slightly compared to another.  This is true for both recording and playback.  If you're recording from multiple devices and outputting to a single device, this is less of a problem as you can always manually adjust the recording within Cakewalk.... but outputting to more than one device really does need to be sample accurate, else you'll get all sorts of random phase issues.

One device needs to be the wordclock master, and all other devices are then wordclock slaves.  The master can be one of your interfaces, or it can be a dedicated wordclock generator.

Then the wordclock needs to be sent to all the devices.

There are three common ways of doing this (there are more, such as MADI, AES, mLAN etc.., but these are the 3 most common):

1.   Via a BNC wordclock connector  - one of the devices will need a BNC wordclock out, and the others will need a BNC wordclock in.  BNC "T" connectors can be used to daisy chain the signal to multiple slave devices.  A 50ohm terminator is usually required after the last connection (unless your device has one built-in).

2.  Via ADAT  - any device with an ADAT input can use the incoming ADAT connection as it's wordclock source from either the master, or any of the slaves sending an ADAT signal.  It doesn't need to be sending audio - it just needs to be connected and sync'd.

3. Via SPDIF  - any device with an SPDIF input can use the incoming SPDIF connection as it's wordclock source from either the master, or any of the slaves sending an SPDIF signal. Again, it doesn't need to be sending audio - it just needs to be connected and sync'd. 

If you're already using ADAT, then consider using that as your wordclock sync by setting your slave device's wordclock to ADAT.  If not, consider connecting the SPDIF out of your master interface to the SPDIF input of your slave (and tell the slave it's syncing to SPDIF). It's a quick and easy way to get them synchronised.
 

Also,  although it's not recommended, you can mix and match how you distribute the wordclock , as long as there's only one master.

At one point I was using my Focusrite Scarlett 18i20 as a wordclock master, with a Fostex VC-8's & ADA8000 slaved via BNC, my TC Helicon VoicePrism sync'd via SPDIF, and an mLAN network slaved via the ADA8000's ADAT out to one i88x, where another i88x and 01X were then sync'd to the first i88x via mLAN.  So four different wordclock distribution methods over 7 devices, but only one (the 18i20) was the master.
 

[Terry Kelley]

So for playback-only to multiple interfaces, do the word clocks matter? All of this is about playback mixdown to quad or 5.1. While it might be nice to record from both interfaces at the same time, I doubt I would. I just want multiple outputs.

Edited July 16, 2021 by Terry Kelley

[msmcleod]

19 minutes ago, Terry Kelley said:

So for playback-only to multiple interfaces, do the world clocks matter? All of this is about playback mixdown to quad or 5.1. While it might be nice to record from both interfaces at the same time, I doubt I would. I just want multiple outputs.

Actually it matters more for playback than recording.

Recording isn't an issue, as you can adjust the recording by nudging the waveform to align it.

Playback is more of an issue.  If the wordclocks aren't in sync, the devices won't be playing at exactly the same time for the same sample time.  This can cause phase issues or a chorus/flange effect if the wordclocks drift significantly.   More subtle wordclock drifting will mess with your stereo imaging, making your mix seem "wider" in places or even lop sided. 

You might get away with it if you carefully choose which instruments go to which output - i.e. make sure stereo pairs, or doubled parts always go to the same device.   But you can't rely on it.

To avoid any of these issues, I'd seriously recommend getting a multiple out audio interface.  You can get them pretty cheap nowadays off eBay, or even refurbished from the manufacturer (Focusrite quite often sells refurbished or older generation interfaces off cheap).  
 

[John Vere]

This was good information , something I actually know about but always forget. Syncing is important and this is why I use ASIO so that it is correct.  Dang audio is complicated!  Anyways I guess it's back to a 4x4 interface and possibly a little mixer to do duty as the second volume control. 

I guess I got away with 2 interfaces for live recording because live recordings have a ton of latency anyway due to leakage. 

One could do a latency test easily buy splitting a signal to both interfaces and comparing the 2 tracks. I might just do that! 

I wonder if Cakewalk would always use the interface that is at the top of the device list for the clock? 

[msmcleod]

13 minutes ago, John Vere said:

This was good information , something I actually know about but always forget. Syncing is important and this is why I use ASIO so that it is correct.  Dang audio is complicated!  Anyways I guess it's back to a 4x4 interface and possibly a little mixer to do duty as the second volume control. 
 

ASIO doesn't guarantee two interfaces are in sync with each other.  If I bought another RME device and didn't sync the wordclocks, they'd still go out of sync even though they share the same driver.   They need to be physically linked to the same wordclock, via BNC, ADAT, SPDIF or MADI etc.

WASAPI and WDM are no different.  I used to use WDM to record from several interfaces at once, and this was fine as the wordclocks were physically sync'd via ADAT.
 

16 minutes ago, John Vere said:

I guess I got away with 2 interfaces for live recording because live recordings have a ton of latency anyway due to leakage. 

One could do a latency test easily buy splitting a signal to both interfaces and comparing the 2 tracks. I might just do that! 

This very true - especially if you're recording live instruments.   And what you suggest is the perfect way of checking how out of sync your interfaces are.
 

18 minutes ago, John Vere said:

I wonder if Cakewalk would always use the interface that is at the top of the device list for the clock? 

I think by default it does, but you can obviously change it.  You should set it to whichever interface is acting as the master for wordclock:



 

[Terry Kelley]

1 hour ago, msmcleod said:

Actually it matters more for playback than recording.

Recording isn't an issue, as you can adjust the recording by nudging the waveform to align it.

Playback is more of an issue.  If the wordclocks aren't in sync, the devices won't be playing at exactly the same time for the same sample time.  This can cause phase issues or a chorus/flange effect if the wordclocks drift significantly.   More subtle wordclock drifting will mess with your stereo imaging, making your mix seem "wider" in places or even lop sided. 

You might get away with it if you carefully choose which instruments go to which output - i.e. make sure stereo pairs, or doubled parts always go to the same device.   But you can't rely on it.

To avoid any of these issues, I'd seriously recommend getting a multiple out audio interface.  You can get them pretty cheap nowadays off eBay, or even refurbished from the manufacturer (Focusrite quite often sells refurbished or older generation interfaces off cheap).  
 

I can see where this could effect the image of anything not exclusively in the front or rear pair. Where is the clock generated? I missed that.

[msmcleod]

1 minute ago, Terry Kelley said:

I can see where this could effect the image of anything not exclusively in the front or rear pair. Where is the clock generated? I missed that.

Most audio interfaces have an internal clock, and this is where they derive their timing from.

If your interface has an ADAT input or an SPDIF input, it will normally have the option to slave it's clock to those inputs instead, which override the internal clock. Both ADAT and SPDIF have a clock signal which is continuously sent along with the audio  - it's there regardless of whether there is audio playing through it or not.

BNC is a dedicated clock input / output - but the theory is the same.

So if you connect the SPDIF output of device A to the SPDIF input of device B, then tell device B to slave to SPDIF, they'll both be using device A's clock.

The best analogy I can think of is,  say you've got two guitarists playing the same part.  They've both got an Alesis SR-16 drum machine set to 120BPM as their metronome, and they're wearing headphones so they can't hear each other.

They both press play at the same time and start recording.

Because the drum machines are going to be very slightly out (or one started very slightly later than the other), the timing will drift.   This is what playing/recording on two separate devices is like if they're not sync'd.

Now say, they get only one person to press play on both machines at the same time (or they've wired up a single footswitch to both devices) .  It's more likely that they'll start at the same time, but the two drum machines are never gonna be exactly 120BPM... they'll start to drift. One might be closer to 120.001 BPM, and the other 119.999BPM.  By the end of the song, they've drifted apart. This is what it's like using a shared ASIO driver without wordclock sync.

So realising this isn't working, they decide to share a single drum machine and both play to that.  Now they're both in sync for the whole song, because they're both using the same "clock".

[John Vere]

The test was a total failure. I could select both interfaces in preferences under WASAPI exclusive. I could also select either interface in the 2 tracks I set up but when you hit record it tells you the old pop up that the device in use does not support the currant driver mode. So testing individually it turns out the Motu does not support WASAPI exclusive. The Focusrite works fine under WASAPI exclusive. So once again every manufacture has a little bit different coding in its driver. 

 

 

 

 

 

Edited July 16, 2021 by John Vere

[Terry Kelley]

Ok, so in Cakewalk/Preferences/Audio/Drivers you can select the Playback or Record Timing Master. Cakewalk is listening to the interfaces clock? I always thought Cakewalk specified and generated the clocks for the interface and these settings determined where the clock was going and coming from. I also didn't consider that the time bases probably aren't high enough that dividing it down would effective divide cut the error.

Very interesting insight. Thanks for taking the time to go over this.

Terry

[Terry Kelley]

10 minutes ago, John Vere said:

The test was a total failure. I could select both interfaces in preferences under WASAPI exclusive. I could also select either interface in the 2 tracks I set up but when you hit record it tells you the old pop up that the device in use does not support the currant driver mode. So testing individually it turns out the Motu does not support WASAPI exclusive. The Focusrite works fine under WASAPI exclusive. So once again every manufacture has a little bit different coding in its driver. 

 

 

When I select WASAPI Exclusive with my iTwo and UCA--222, Cakewalk complains that my format (44.1k/16b) isn't compatible with one interface (the UCA-222) and switches to 48k/16b. I can then assign four tracks to each of the four inputs and they do record.

Edited July 16, 2021 by Terry Kelley

[John Vere]

So in WDM mode things are worse. The Motu just does not like anything other than ASIO and WASAPI shared. Look at the gibberish I get in WDM mode..

[Terry Kelley]

To MSMcLeod:

When running these interfaces without a common clock the setting of the Mixing Latency Buffer Size affects the sync of the two interfaces doesn't it ...

If I set it short it will start to cutout/jump or if too long sync starts to drift. There is a sweet spot but it's a guess.

 

[Xoo]

My older MOTU doesn't like Wasapi exvlusive either unless I disable a setting in its control panel (Use winrt or similar?).  See if that's there for you and if it makes a difference.

[John Vere]

@Kevin PerryIt's not worth the bother, I was just curious  and love a challenge. I probably could have gotten it to work but its summer and the swimming hole is close by..  I have my Tascam us1641 which has 14 analog in and 4 outs. It works great but the driver is 2014 and I do not expect it to survive to much longer. Tascam was writing good drivers for a short while and I see they have returned to the lazy method again. They are selling large expensive interfaces that use "class Compliant drivers" ,, no ASIO??    But I still have a system with W7 on it so I have that as my mobile live recording rig. If I keep it in a cool dry place I could get 10 more years out of it.  

I love the Motu so far and it's very happy with ASIO. It would have been cool to switch on the 6i6 and grab some extra channels but seems much easier to just plug the Tascam in and bingo, 14 channels of ASIO. 

Edited July 16, 2021 by John Vere