• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Help - Motu M2 Not Working with Dirac Live 3

Χ Ξ Σ

Senior Member
Forum Donor
Joined
Jul 16, 2020
Messages
457
Likes
1,969
Location
UTC-8
I received the Motu M2 yesterday. It functions normally as a DAC but not so well with Dirac Live Processor on my MacBook.

The digital signal path is Local Files-->Audirvana hosting Dirac Live Processor-->USB-->Motu M2. Sometimes it works, but it is a hit or miss, and I am getting more misses than hits.

I have installed (and later uninstalled) the Motu driver for Mac. With the Motu driver, the Motu M2 simply cannot produce a normal sound with Dirac. It either produces a fragmented, randomly paused sound, or a mono clipping signal. Uninstalling the Motu driver and using the driver came with macOS, the Motu can sometimes produce normal music with Dirac, but more often than not, it has a weird effect on the music as if it was playing mono or 8bit.

Other things I have tried: switching between using Audirvana's library and Music (iTunes) Integrated Mode, line-out vs headphone out, using Dirac with Audio Hijack and without Audirvana, turn on-off software volume control. As long as I remove Dirac from the signal path, the problem seems gone. However, Dirac is not the cause because it works with my other DAC.

So, How do I make them work together? Please help me troubleshoot this issue.
 
Last edited:

M00ndancer

Addicted to Fun and Learning
Forum Donor
Joined
Feb 4, 2019
Messages
719
Likes
728
Location
Sweden
So, How do I make them work together? Please help me troubleshoot this issue.
Check the CPU load on your MAC. It is possible that using Dirac live taxes the cpu. Shouldn't be that but it's worth having a look. Just to make sure. How much memory do you have free when the application is loaded. What goes on in the background?
 

Vasr

Major Contributor
Joined
Jun 27, 2020
Messages
1,409
Likes
1,923
Hard to make a diagnosis remotely but one thing to eliminate would be digital clipping. The way to debug these things is to keep eliminating variables until the problem disappears. Make sure you use the same piece of audio content over and over for each test, to eliminate issues with different sample rates, different loudness levels, etc.

Some DACs can behave badly with inter-sample peaking with content that has 0db peaks. Using room correction aggressively can create 0db peaks in digital.

1. Try using Dirac plug-in with NO filters installed (or selecting a preset with no filters installed). If this still has the problem then it isn't a Dirac induced clipping issue downstream.
2. To verify further, slide the gain in the Dirac Plug-in settings panel lower. This should reduce the volume and give more head room for filter correction. See if this makes a difference.
3. Reduce the volume control associated with the USB device. See if that corrects the problem.

If the above did not change the symptoms then it isn't likely a digital volume issue. Modify the device driver parameters if any in sample rates and depth to see if it makes a difference.

Other possibilities include some kind of buffering or sample rate conversion issue in the chain. On the PC, you can try installing the Dirac plug-in in a VST host as an alternative in the simplest chain possible to the USB driver preferably in exclusive mode.

Dirac support should be able to guide you through trouble-shooting but having the results of the above tests might help them diagnose.
 
OP
Χ

Χ Ξ Σ

Senior Member
Forum Donor
Joined
Jul 16, 2020
Messages
457
Likes
1,969
Location
UTC-8
Thanks for everyone's reply. I have contacted Dirac and asked this in the Motu community, haven't got a diagnosis yet.

I reinstalled the Motu driver again. With the Motu driver Dirac produce no sound. After I re-uninstalled the Motu driver for the second time and started using the M2 with the Mac's CoreAudio, The clippings and pauses are magically gone. Instead, it shows a different problem than what I described in the original post. It no longer sounds mono or 8-bit. (It's quite baffling how inconsistent these problems are.) Rather, it sounds as if Dirac's filter is not working. When the Dirac Live Processor is running, the existing filters (created when using Topping D90) have little effect on the frequency response (confirmed by REW measurements). I then tried to redo the measurement and create new filters in Dirac Live application, but the sweeps became really long, and the process always ended in error reports of low SNR or lost connection with DLP, and I could not proceed beyond the measurement step.

I can rule out Audirvana as the cause because using a different host (Audio Hijack) makes no difference.

Same settings on My MacBook and Audirvana for both DACs, yet the D90 never has a problem and the M2 always has. I wonder what's in the D90 that the M2 doesn't have.

Will keep trying.
 
Last edited:

Vasr

Major Contributor
Joined
Jun 27, 2020
Messages
1,409
Likes
1,923
I reinstalled the Motu driver again. With the Motu driver Dirac produce no sound. After I re-uninstalled the Motu driver for the second time and started using the M2 with the Mac's CoreAudio, The clippings and pauses are magically gone. Instead, it shows a different problem than what I described in the original post. It no longer sounds mono or 8-bit. (It's quite baffling how inconsistent these problems are.) Rather, it sounds as if Dirac's filter is not working.

The above is also consistent with Dirac not being in the new path at all rather than solving the original problem!

Are you aware of how to open the two UI panels associated with the Dirac VST plugin? One is a settings panel with some sliders to select settings. The other is a panel with filter selection settings and a "led meter" for signal going through. You can verify if Dirac is processing the sound that you are hearing with the latter and if the proper filters are selected. The latter will allow you to move the gain to see if it affects the volume so you can be sure if Dirac is processing the signal.

If you don't know how to open the above, you should figure it out for the set up you are using. It will help you try the troubleshooting I posted earlier to narrow down the problem. Otherwise, with this large scale changes you are not getting anywhere close to narrowing it down with multiple reasons of where the problem might be.
 
OP
Χ

Χ Ξ Σ

Senior Member
Forum Donor
Joined
Jul 16, 2020
Messages
457
Likes
1,969
Location
UTC-8
Dirac has always been in the path when any of the aforementioned anomalies happened. My Dirac AU plugin only shows one panel, which has filter selections and a signal meter slider. See attached Pic 1. The signal meter demonstrates no problem regardless of whether the sound is broken or coherent, whether I was using M2 or D90, whether CoreAudio or Motu driver is in use, or whether the filter has an actual effect on the speakers' frequency response.

I have measured the speakers' FRs using different DACs with CoreAudio. Attached Pics 2-5 are the results. When M2 is working with Dirac, the filter has minimal effect on the FR other than lowering the level and adding some choppiness. When the filter is turn off and Dirac is left on, the M2 is unable to produce frequencies high than 4kHz. When switching back to D90, everything goes back to normal.
 

Attachments

  • Pic 1.png
    Pic 1.png
    192.3 KB · Views: 122
  • Pic 2.jpg
    Pic 2.jpg
    104.7 KB · Views: 119
  • Pic 3.jpg
    Pic 3.jpg
    107.5 KB · Views: 106
  • Pic 4.jpg
    Pic 4.jpg
    104.4 KB · Views: 106
  • Pic 5.jpg
    Pic 5.jpg
    102.6 KB · Views: 111

Vasr

Major Contributor
Joined
Jun 27, 2020
Messages
1,409
Likes
1,923
This is a real puzzle.

Assuming the M2 isn't faulty...

1. Is the only difference in using the M2 or D90 in switching just one cable from one DAC to the other with no other changes in the end-end chain? If not, there may be an explanation somewhere in there. If it is the same, then there might be an issue with your M2.

2. What is the REW measurement via the D90 with Dirac filters off? Is it identical to the M2 graph?

3. When content is playing through the M2, reduce the gain in the Dirac panel down to about -10. Does the volume change correspondingly? Do the REW graph with that low gain set both with Dirac filters on and without for the M2. Do they remain the same as before (but at lower dB?). Do they still exhibit the same problem at lower gain?
 
OP
Χ

Χ Ξ Σ

Senior Member
Forum Donor
Joined
Jul 16, 2020
Messages
457
Likes
1,969
Location
UTC-8
Mystery solved! The Software Volume Control on Audirvana has to be turned on!

I had tired turn on and off the Software Volume Control, but the software was always able to control both D90's and the M2's volume (with or without Dirac running), so I thought this setting made no difference. However, with it was turned off, Dirac was unable to filter the frequency response when connected to the M2. Again, not an issue with the D90.

This also explains why Dirac Live Application could not perform measurement: the Dirac Live Processor must be online, which was hosted by Audirvana. Fixing the setting in Audiravana also fixed the issue with Drac Live.

Now that the M2 functions normally with Dirac, it does not work as seamlessly as the D90 does. When the sampling rate changes from tracks to tracks in the shuffle mode, say 44.1kHz to 96kHz, the M2 takes a few seconds to change the sampling rate and inevitably skip the first few seconds of the track. Changing the sampling rate for the D90 is considerably faster, and it won't skip the first few seconds.

What could that be? USB powered vs plugged into an outlet? 192kHz vs 768kHz? 24bit vs 32bit? Clocking? I checked Audirvana's debug info, while the current buffer frame size is the same for both DAC at 512, it says the audio buffer frame size is 50 to 4096 frames for the M2, and 15 to 4096 frames for the D90. Is that why there is a big difference in sampling change latency?
 
Last edited:

Vasr

Major Contributor
Joined
Jun 27, 2020
Messages
1,409
Likes
1,923
Glad to hear the mystery is solved.

Different implementations have different latencies for detection of incoming sampling rate. No way around it. No way to predict it. Depends on the receiver card implementation within the unit I guess. Don't really understand why some can do it so much faster than others. Size of the input buffer? the speed of the processor? Perhaps someone who has the technical knowledge can explain.
 

AnalogSteph

Major Contributor
Joined
Nov 6, 2018
Messages
3,335
Likes
3,278
Location
.de
Now that the M2 functions normally with Dirac, it does not work as seamlessly as the D90 does. When the sampling rate changes from tracks to tracks in the shuffle mode, say 44.1kHz to 96kHz, the M2 takes a few seconds to change the sampling rate and inevitably skip the first few seconds of the track. Changing the sampling rate for the D90 is considerably faster, and it won't skip the first few seconds.
There is no specific technical reason why changing sample rates would have to take several seconds, let alone skip some of the audio (some rather overeager momentary muting being triggered?). I would try to reproduce the issue in a stripped-down setup with just audio player > M2, and if it persists, contact MOTU support to hear what they've got to say.

As a workaround, you could try enabling upsampling to e.g. 192 kHz in Audirvana, avoiding the need to switch sample rates altogether.
 
OP
Χ

Χ Ξ Σ

Senior Member
Forum Donor
Joined
Jul 16, 2020
Messages
457
Likes
1,969
Location
UTC-8
I guess the D90 has a shorter/broader digital path than the M2, or perhaps their USB interfaces implemented PLL differently. Motu does offer a Mac driver that supposedly reduce latency, but as I mentioned before, I had no success with it.

Under my original settings, the skipping happens whenever the sample rate changes. I hear the skipping with both Audirvana and BitPerfect, two music players that handle sample rate automatically. Dirac is turned off.

The skipping in Audirvana can be eliminated by adding latency to the track. I have added a whole 3s of delay before the skipping is gone. However, changing the sample rate already takes 5s. Now it takes 8s before the next track is playing. Yikes.

I also tried using forced upsampling. It gets rid of skipping and inherent latency altogether. However, a short blip before the next track is inevitable. I tried turning on Mute When Changing Sample Rate and adding latency, but the blip is still there.
 
Top Bottom