Motu 828es Phase Issues

[GianlucaDelfino]

Dear fellow MOTU 828es users,

I am experiencing phase issues on my MOTU 828es: when I feed the same mono signal to inputs 1&2 (front) I should get two identical mono signals in both channels.
However, when I analyse them with MOTU Audio Tools Phase Analysis in rectangular view, it shows a line that tilts right very much like the example MOTU gives for the one-sample delay case in their MOTU Audio Tools user guide.
It looks like channel 2 has a delay of one sample.
On further inspection, I could verify that at 48Khz sample rate, there is one-sample delay between the channels. It does vary at different sample rates. In Ableton Live, as a workaround, I managed to compensate using either the AlignDelay plugin or a custom built max4live patch: by delaying channel number one of exactly one sample @ 48khz sr I can perfectly phase-align the two channels (straight line in Phase Analysis).
This happens with all input pairs 3-4, 5-6, 7-8, but curiously the delay for all rear inputs is reversed (line tilting to the left, showing a one-sample delay in channel 3, 5, 7). On the other hand, it does not happen if I pair the inputs this way: 1-3, 2-4, 5-7, 6-8.

My workaround allows me to fix my recordings, but I can’t do anything for live monitoring nor for standalone mixer use.

I contacted MOTU support in September 2023 and I sent them screenshots, audio clips and detailed description. They recognised the issue and told me they were going to check with their engineers. I have regularly asked for updates every 15 days or so, but my issue has been constantly postponed until yesterday, when I received a message stating “I do not want to get your hopes up for an update in the near future”. After waiting for more than 4 months, my hopes are very low indeed.I guess this is how it goes with out-of-warranty products (my 828es was 3 years old when I first contacted them).

Hence, I am asking the community: could you please check if you can reproduce the issue on your units?

Please let me know, that would be much appreciated.

Best,

Gianluca

[mikehalloran]

Yikes!

How exactly are you feeding the same input to two channels? Please describe the signal chain.

[James Steele]

Yikes!

How exactly are you feeding the same input to two channels? Please describe the signal chain.

I was kind of wondering the same thing. Makes me wonder if the analog signal is arrive at the inputs out of phase? Like maybe one channel is slightly delayed on the way in? Seems like this would be a glaring fault and not sure how the engineers that designed it would let that get through?

[HCMarkus]

Good work and interesting results, Gianluca. Thank you for sharing.

In all honesty, unless we are feeding a mono signal to two channels simultaneously (and why would we need to do that?), I don't see this as a major issue. The phase differences you will see with even an XY coincident mic pair will be orders of magnitude more significant.

If this is just too much for anyone, it looks like running stereo inputs on even-even or odd-odd pairs of inputs will alleviate the issue. Easy to set up in Bundles if running DP. Matter of fact, I think I'll just do a little cable swapping and bundle adjust now. Should take me all of a couple of minutes.

[James Steele]

...looks like running stereo inputs on even-even or odd-odd pairs of inputs will alleviate the issue.

Hmmmm... if others confirm this issue, I suppose that would be a good workaround. Are you saying, though, Markus, as I think you are, that if recording stereo program material into an input pair... for example room mics for a drum kit... that for all practical purposes this isn't much of an issue?

[HCMarkus]

Are you saying, though, Markus, as I think you are, that if recording stereo program material into an input pair... for example room mics for a drum kit... that for all practical purposes this isn't much of an issue?

Indeed I am. Likely isn't an issue at all, especially for a spaced pair Even a coincident pair is going to have out of phase stuff, especially up high, where you start to see the phase anomaly become more graphically apparent due to the shorter wavelengths involved. At 10k, the wavelength is a tad over an inch.

Just because I got a wild hair, I put the two pairs of mic Pres I tend to use as stereo pairs on even inputs. Took but a couple of minutes to switch a couple of plugs and fix my template Bundles, and the label on my 828ES. But I ain't gonna' worry about it more.

[mikehalloran]

That’s why I asked about the signal chain. I learned a thing or two studying for a Class 1 FCC License a half century ago.

With a stereo source, a one sample delay just isn’t an issue and with a pair of identical mono inputs, panning each hard R/L makes it go away. If there really is a need, the odd or even bundles workaround makes this a non-issue.

MOTU’s response tells us that this is in one of the processors, possibly the DAC IC which handles many functions nowadays. They would have to find a replacement that does everything they need without this flaw that has no practical consequences. Good luck with that.

[GianlucaDelfino]

Hi,

Thanks everyone for your replies.

I agree this would not be an issue if you were recording a stereo mic pair in any configuration.

However, if you’re trying to record a stereo signal from a mixer through inputs 1&2, this becomes an issue as all mono components are misrepresented in terms of phase.

When I record a stereo mix down of my tracks, whenever a mono element plays on its own (let’s say a perfectly mono kick drum), it shows phase alignment problems.

I have carried out an experiment by recording the same source through inputs 1&2 and through an external Adat box (behringer ada 8200) and the phase misalignment shows up only through the 828. I can even venture to say that I can hear a difference in lower end tightness and high frequency spectrum. This is confirmed by checking with a spectrum analyzer that shows noticeable differences in high frequency slope.

I hope this clarifies my point

[stubbsonic]

I would have thought that 1 sample delay would only affect frequencies at around 20K.

[GianlucaDelfino]

You a are right, probably the low frequency tightness problem is just in my head…

On the other hand, the high frequency difference is there, and can be measured. Whether it is relevant or not, that depends on usage, but my point here is to investigate whether the technical issue of phase misalignment is a fault in my unit or a design flaw.

[James Steele]

Well I guess if this is a real issue and I’m going to be wiring a patch bay soon I may decide to wire mic pres that might record stereo sources into even+even and odd+odd inputs like HCMarkus mentioned.

[stubbsonic]

I'd be inclined to just nudge a track by 1 sample only as needed.

[Phil O]

I would be inclined to ignore it except in a situation where that type of precision was absolutely needed (essentially never in my work). One sample at 48K works out to be about .28 inches (using 1120 ft/sec as speed of sound). Try moving one mic a quarter of an inch in a coincidence pair and see if you can actually hear the difference. If you can, then look for a remedy. If you can't, don't fuss over it.
If I needed a remedy for this, I would document the time relationship of every channel, not just adjacent pairs. Then when recording multiple tracks, nudge (as Mr. Stubbs suggests) the recorded material as needed. Now you are sample accurate. Done.
For live situations, when would you ever need sample accurate signals between channels?

Your friendly neighborhood Phil

[GianlucaDelfino]

Dear Phil, thank you for your reply.

As stated earlier, I am not concerned with microphone placement.

Whether the problem is relevant or not, is a matter of what you use the interface for.

I am just trying to understand whether this is a fault in my unit or a design flaw.

[stubbsonic]

... I would document the time relationship of every channel, not just adjacent pairs.

This. If you could sent a short tick of audio to all the analog inputs simultaneously then analyze.

It would be even better (easier) if MOTU would just run these tests and document what is happening.

EXAMPLE TEXT: "We've tested both our A-to-D and D-to-A and have determined ... " followed by...
"A firmware fix is forthcoming."

Then, it might be interesting to see if this issue is common with a particular chipset or across different manufacturers' interfaces.