Print

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

[Amfibius]

I have done a search and looked at old topics, but that has not helped me. In particular, I found this. It seems as if that guy gave up and got rid of his RME and got a Lynx instead.

What I want to do: use REW to generate a sine wave sweep and send it through JRiver's convolver via ASIO mode. This will require that I set up loopback in RME Totalmix.

What I need to do in detail:

1. Check that RME Totalmix is multiclient capable. A bit of googling says yes.

2. Set 48kHz in REW, Totalmix, and JRiver. Done.

3. Check that my config file is correctly loaded into JRiver's convolver.
---> Result: yes, I can play back music and hear it on my speakers.

4. Check that loopback is working.
- In REW, set output to ADAT3 and input to ADAT3
- In Totalmix, click on the "loopback" button in Hardware Outputs, then slide up the volume in ADAT3 "Software Playback" and "Hardware Outputs".
- Back in REW, press on "check levels". I can see the output bar, and I can see the input bar. The input bar disappears if I disable loopback.
- Use REW to do a measurement using loopback from 1Hz to 24kHz. Result: I measure a perfectly flat line.
---> To me, this confirms that loopback is working in Totalmix, and it is correctly sending output to ADAT3.

5. Find out what output channel ADAT3/4 sends to. According to Totalmix's Matrix view, this is OUT 13/14.

6. Open ASIO Live in JRiver
- Tools-Options-Audio-Advanced: "Set the active zone for WDM/ASIO driver live playback" TICKED.
- Configure ASIO: correct device chosen, 48kHz, 2 channels, Channel offset 13 (I also tried every number from 0 to 20!!!)
- In REW: play pink noise
---> RESULT: I don't hear any pink noise. JRiver says "ASIO Opening..." and that's it. Nothing happens.

I am pretty confident that I have loopback working correctly, given that flicking the loopback switch on and off makes the input bar appear/disappear in REW. So I do not think this is an error in RME setup.

Given the opinion on the other thread from pschelbert and mattkhan about MC having issues with some multiclient ASIO software and JimH saying that a problem was found and will be fixed, I am wondering if it actually has been fixed. For the record, that thread was about MC 28. I am on MC 31.

[mattkhan]

which RME interface do you have?

if you only want to measure your system then you can use the offline sweep method instead as that takes loopback out of the equation (as, depending on the system, that can also lead to instability in the audio stream that isn't present for simple playback)

it's basically https://wiki.jriver.com/index.php/Verifying_DSP_Studio#Generate_a_Measurement_Sweep though the specific options in rew have changed a bit since I wrote that (i.e. the REW UI has changed quite a bit)

[Amfibius]

Hi Matt. I have a RME Fireface UC.

I am aware of the offline sweep method. However, there are many reasons why I want to send REW through JRiver "live":

1. I want to measure the latency of the entire playback chain and do timing measurements with an electrical loopback timing reference. That requires both ASIO and WASAPI loopback through JRiver. I have already figured out how to do WASAPI loopback.

2. I have suspicions about the variable latency of WASAPI loopback (I have done about a dozen measurements and observed timing differences between all of them!). WASAPI loopback requires FIVE separate software in the pipeline to get it to work - REW, VAC or ASIO4ALL, JRiver, Totalmix, and Windows itself. If I am making timing measurements, I don't trust that it will all work together, and indeed my timing measurements have poor repeatability. I want to see if timing measurements with ASIO are more repeatable.

3. It should work, and if it doesn't, it needs to be brought to the attention of developers. I don't think we should resort to workarounds because there is a bug in the software. I don't know whether it's JRiver's fault or RME's fault, but your comments (and JimH's comments) in the thread that I linked to suggest that there may be a problem with JRiver that was supposedly addressed years ago.

[mattkhan]

The linked thread has someone saying it was fixed between 28.0.44 and 28.0.50 but nothing in the release notes seems related that I can see (https://wiki.jriver.com/index.php/Release_Notes_MC28) so that's a bit confusing.

Given it works in multi client mode with other software and given MC is the one initiating the connection, it really needs MC developers to look at it. I don't know why it's unreliable as i definitely remember it not working at times then working at others. Just to debug, have you tried a reboot and then play nothing else before you attempt this? Have you tried playing a sweep via some other software just to remove rew from the equation?

[Amfibius]

The linked thread has someone saying it was fixed between 28.0.44 and 28.0.50 but nothing in the release notes seems related that I can see (https://wiki.jriver.com/index.php/Release_Notes_MC28) so that's a bit confusing.

Given it works in multi client mode with other software and given MC is the one initiating the connection, it really needs MC developers to look at it. I don't know why it's unreliable as i definitely remember it not working at times then working at others. Just to debug, have you tried a reboot.and then play nothing else before you attempt this? Have you tried playing a sweep via some other software just to remove rew from the equation?

The answer to all your questions is "yes", except for the last one. I can try sending a sweep from Acourate through loopback and into JRiver, but I am extremely sceptical that the result would be any different. I have tried two versions of REW though, the older version that I was using, and the new Beta 5.40 version.

Actually, there is another diagnostic step I could try. I will connect my DAP to the optical input of my RME and see if ASIO Line in works. I thought of that earlier today, then went hunting for my optical cable. I can't find the bloody thing, and the shops are closed. Something to try tomorrow.

[mattkhan]

I am also sceptical it will be different but it's useful in so far as it would add support for the idea that the MC is the software that has the problem

[Amfibius]

I am also sceptical it will be different but it's useful in so far as it would add support for the idea that the MC is the software that has the problem

I can confirm that Acourate can not pass signal to JRiver through loopback. The same as REW.

BTW to be fair to the MC team, one could equally say that RME software is the problem because MC is known to work with other brands of interface. To me, it's either MC or RME, and as a consumer I have no way of knowing. I was hoping that I made a stupid mistake - no matter how careful I am, I should never underestimate my own stupidity. I think I have covered all bases, and if there is nothing else, then maybe the MC team should have a chat with RME and sort this problem out between them.

[Manfred]

I had in some ways a similar Use Case a few years ago discussed in this forum.
I have an rme HDSPe AIO PRo PCie Soundcard.
Goal was to have a unique MC DSP Processing for Computer, TV and Other Input Devices befrore routing the audio Signal to the DAC.

1. route Audio Signal from the Computer using the rme Madiface AISO Driver through AES to my Devialet and having MC EQ or other MC DSP applied (This works with out any issues)
2. route the Audio Signal from TV or other input devices through the HDSPE AIO Pro to MC, apply MC EQ and other DSP settings  and then reroute the signal through AES to my Devialet. That does not work!

Routing the audio signal back through USB to the Devialet works (not stable) and creates so large latencies that it was useless.

Only chance to getting something like 2. working is that the MC Team speaks with the rme Team how to make this happen at very low latency. It needs a special driver or plug-in?

[mattkhan]

Works fine with other interface, no special drivers required. I would think that an MC developer with an rme device locally for testing would get it working soon enough