Multichannel audio on a Pi will get a whole lot easier and cheaper!

[Frunse]

Robert, I am not comparing digital vs. analog volume control, but digital in the player/DSP vs. digital in the DAC. Volume control in vast majority of DACs is digital too.

If your tweeters hiss at zero volume, simply your analog chain (including the DAC) is electrically too noisy. Of course analog volume behind the DAC will attenuate the DAC noise (and introduce its noise). It depends on your setup what is better.

I got me this day an nice Present

TinySA Ultra realy nice little Puppy, but damm, i need new Glasses

 

[Frunse]

The Day of the truth is there

I conected both Motu UL mk5 with USB to an PI 5 with CamillaDSP Software.
Build up the Alsa Device and made that /etc/asound.conf with 36 Channels at all.
Then i configured CamillaDSP.
And made the Measurements between the Channel Out One, on both Devices.

Timedifferences is 462us, even if you conect both devices with Toslink and sync the second Device to Optical IN, this Timedelay will be the same!
But:
Without the Optical Sync between the two Motus you get an Drift of that Delay Time, it is Moving over Time like an Slowmotion Movie on Osciloscope.

With Sync it is Stable at 462us

Even i switch of the Input and start it again the Delay is fixed at 462us.

On first Motu i measured the Time Betwen Channel Out 1 and 2. Timedelay on Scope was around 0-2us.
After i add an fixed Delay to Motu 1 Out 1 of 462us
I got also Timedelay, between Both Motus Out 1, of 0-2 us fixed!!!

So now i need to Check all chanels in between.

And of course making an Settup with Speaker and Check how it sounds

Checked it between Motu 1 Channel Out 1 and Motu 2 Channel Out 10 the Delay is the same. after adding 0.462ms to Motu 1 Channel Out 10,
Both Channel 10 on Motu 1 and 2 had 0-2us again

It looks like CamillaDSP Channel 1-18 is now Motu 2 ((hw:CARD=UltraLitemk5_1,DEV=0) and Channel 19-36 is on Motu 1 (hw:CARD=UltraLitemk5,DEV=0)
Not sure why.

After reboot the Motu 1 was again Channel 1-18 and Motu 2 was 19-36 !
Need to watch at this may be some Timing problems at boot for USB Detection not sure.

I changed also the Buffer on the Motus from 512 to 256 and the Delay after Boot was at 362us between both Motus on CamillaDSP.

I guess 400us isn't at all an Problem because some Systems could set A/V sync between 0 and 300ms.
I read some People using 60-80ms (Lipsync, A/V sync ???) on an Surroundreciver, so not an Brainer i guess.
But need to check it with aplied FIR Filter by sure.

New Problem found and can't realy see why.
If i use 1khz Tone to detect the Delay between the Channels i see that 364us Phaseshift Between Channel Out on Motu 1 and Motu 2.
1 add an Delay to Motu 1 with 364us on Camilla and the Phaseshift on 1 kHz is fine at 0-1us.
Now i change The Inputfrequenz to 5 kHz and get an new Phaseshift of 41.5us now i am a bit confused.
And at 100Hz 1.06ms thats realy Bad but why?

I used REW Generator to feed in the Tone from HDMI L+R on Windows to "HDMI to I2S" Board on RPI 5

And between OUt 1 and 2 on Motu 1 no Phaseshift, anyway what kind of Frequenz i Use how suggested.

Probably not the right way to make it.

I go now for the 16 Channel Setup with adding the Motu 2 thru LWL (4 Channels) and SPDIF (2Channels) instead of 2x USB

And do the Test again.

Robert

 

[Frunse]

I am not the Man with much Luck!

Found an Bug on Motu UL mk5 Device.

All Digital Signal going into Device (ADAT 1-4 and SPDIF, all at 96kHz) i can see it on CueMix 5.

But i cant get the SPDIF Signal routet to any Analogue Outs on CueMix but all ADAT Digital In i can Route to any Analogue OUT.

Thats makes no sense it must be a bug in Firmware.

CueMix and Firmware all is Updatet to last Version!

Robert

 

[mdsimon2]

I am not the Man with much Luck!

Found an Bug on Motu UL mk5 Device.

All Digital Signal going into Device (ADAT 1-4 and SPDIF, all at 96kHz) i can see it on CueMix 5.

But i cant get the SPDIF Signal routet to any Analogue Outs on CueMix but all ADAT Digital In i can Route to any Analogue OUT.

Thats makes no sense it must be a bug in Firmware.

CueMix and Firmware all is Updatet to last Version!

Robert

I never routed SPDIF to analog using Cuemix but you can definitely do it in CamillaDSP without issue.

Michael

 

[Frunse]

I never routed SPDIF to analog using Cuemix but you can definitely do it in CamillaDSP without issue.

Michael

Thats not the Problem!!! I have Camilla routed to SPDIF OUT of Motu 1
Then conected to SPDIF IN Motu 2 like an external DAC but i cant get anything to any OUTS there, so SPDIF IN is Usless on Motu Device at the Moment.
But all ADAT IN's i can Route to any OUT.
Thats stupid why it has SPDIF IN but cant get Use of it.

SPDIF OUT is on Channel 10-11 CamillaDSP.
0-1 and 4-5 are Tweeter and Subs so 1 kHz wont work there

I read the Manual but there isn't writen anything that this wont' work
SPDIF is deaktivated at 192kHz but i am at 96kHz and I see the input Levelmeter at full level for ADAT IN and SPDIF IN but can only route ADAT not SPDIF for what reason???

So why it has SPDIF IN may be only for USB Conection but then, why i can move an Slider for SPDIF IN on Mix BUS for MAIN 1-2 thats makes no sense???
Thats would be DUMP Realy DUMP and i will send it Back by sure.

I try to make an ticket on Motu Support but this wont work also can't get it done after hitting submit nothing happen, what a crapy day

Got at the End an Support Ticket Submitted now lets wait what they say

Robert

 

[Frunse]

Did the second Test with 2 Motu UL mk5 the second is connected as an Dac to the first one with ADAT and SPDIF synced to SPDIF.

Timedelay (Phaseshift) Measured between Channel 0 Motu Device one and Channel 0 Motu device 2:

1kHZ Sinewave aplied thrue HDMI to CamillaDSP = 70us

No Phaseshift between Channel 0 and 1 on second Motu Device

After adding 70us Delay on CamillaDSP to Channel 0 and 1 on Motu Device one:

1kHZ Sinewave aplied thrue HDMI to CamillaDSP = around 1us
100Hz Sinewave aplied thrue HDMI to CamillaDSP =around 30 us
16kHz Sinewave aplied thrue HDMI to CamillaDSP =around 950ns

So Much better than using 2 Motu Devices with both connected with USB to CamillaDSP
But still little Timedelays between different Sine Frequenzies.

That realy need an Expert tell to me how this come's

By sure i understand the delay thrue ADDAT because it needs abit longer till the Data arrived on second Motu but why Different timeshifts on different Frequenzies???

The data transferred has every time the same ammount off Bits no Matter what kind of Data recordet there only some Jitter would addad but the devices are synced together so i am abit confused.

I am not an Expert on Digital Math, could be my Measurments are not done right but on one Device there are no different Phaseshifts between 2 Channels on different frequenzies only on That ADDAT Stream to Second Motu UL

Conclusion:

Use One Device with 16 Channels would be the best, even ADDAT has some little Timedelay problems not sure how they do it in Studios.

But no Device there for now with 18 or 24 Channels and SabreES9028Pro or ES9038Pro, so for me an good Compromise for now if this SPDIF bug would gone on Motu UL let see.

Robert

 

[HenrikEnquist]

1kHZ Sinewave aplied thrue HDMI to CamillaDSP = around 1us
100Hz Sinewave aplied thrue HDMI to CamillaDSP =around 30 us
16kHz Sinewave aplied thrue HDMI to CamillaDSP =around 950ns

Sine waves are not great for measuring the delay accurately. At 100 Hz the period is 10ms, and 30 us is only 1/300 of a period. Can't do much better that that!
Try with a slow square wave instead, like 10 Hz. The sharp flanks are much easier to compare.

 

[Frunse]

Sine waves are not great for measuring the delay accurately. At 100 Hz the period is 10ms, and 30 us is only 1/300 of a period. Can't do much better that that!
Try with a slow square wave instead, like 10 Hz. The sharp flanks are much easier to compare.

i tried it yesterday but i have realy much Ringing on Rectangel out Of Motu i will make the Measurments again with Rectangel, let see.

 

[HenrikEnquist]

i have realy much Ringing on Rectangel

That is as expected! It's just the sinc reconstruction filter. Look around the zero crossing, there you should have nice steep flanks.

 

[Frunse]

That is as expected! It's just the sinc reconstruction filter. Look around the zero crossing, there you should have nice steep flanks.

Hello me again and did the Measurements, Rect is much more Precise and so you could see Delay at 10Hz and at 1kHz is neare the same around 73us and could good compensate with an Delay Filter on Camilla

to the higher frequenzies like 10kHz it changed a bit but the Oszilloscope cant cound it right too only 9600Hz he Count for that 30.3us Timedelay so i guess it could an Measurment error bebecaus of the Filtering at the ege of the DAC BW or so.

I need to do the Measurement again with Both aplied as two USB Devices to Camilla i guess.

And thank you Henrik you did an great Job with Camilla i love it

Gives me much Funtime to play with.

Robert

 

[Frunse]

Here the new Measurements with 2 Motu UL mk5 , both conected as USB Devices with 36 Channels at once.

Both Motus synced with SPDIF Cable Motu one, Spdif out to Motu 2 Spdif In.
Motu 1 Synced to Internaly and Motu 2 synced to SPDIF. Thats Importand or you get Drifting in Timedelay.

Looks very Promising.

I used only 10 Hz and 1kHz Rect at higher frequenzies the Rect are a bit worse, dont trust the Scope Measurement

115us Delay need to be Compensate to the first Motu Device on Camilla.

Now need to see how stable this will work on an Speaker Settup and how it sounds

Little Problem i run into after reboot the Motu 1 was before Channel 18 to 35 and Motu 2 was Channel 0 to 17.
Now it is again right Motu 1 is Channel 0-17 and Motu 2 is 18 to 35 and the Delay was now 311us i need to compensate \o/

Need to check if i could do an arangement which Device will detect before the other on boot for USB not sure about.

Or i need to made an Chain while Power on, first the RPI 5 with Time delay then Motu 1 and then motu 2 need to Look into it.
Alo need an start delay for CamillaDSP Software on Linux i guess.

Bad News even after reboot the Motu 1 and 2 dont swap, i get every time an different Timedelay between the 2 Devices and also sometimes the Phase need to be corrected in the other direction so need to switch the Channels to aplie the right delay, thats will not work that way

Not sure its possible to fix this/ compensate this on Software.

Robert

 

[somebodyelse]

You can add a udev rule to make sure the interfaces are ordered consistently.
https://alsa.opensrc.org/Udev

 

[Frunse]

You can add a udev rule to make sure the interfaces are ordered consistently.
https://alsa.opensrc.org/Udev

Oh Man thats looks realy complex to me but i will try.
Still the question is, will the Delay change every time after Boot, like now, or not and why is it?

Robert

 

[somebodyelse]

Still the question is, will the Delay change every time after Boot, like now, or not and why is it?

Can't help you there - never looked that closely at the delay.

 

[Frunse]

The Motu Support does the Job and helped me!
On that new Motu UL mk5 i had to make an Factory Reset, after that i setuped all again like before but it works now also with SPDIF.

16 Analogue Outs working now! (with ADAT and SPDIF)

Only 74us of Delay between the two Devices (ADAT) but every Time the same after Boot so it will work this way.

SPDIF i will Measure next day if its differ to ADAT or not but on Surround Settup you need to Measure each Speaker and add an Delay that all plays at the same Time at hearing Position, so no Brainer

So i did all Measurments and got this results (1kHz Rect):
Motu UL Device one CHannel 2 to 6 Delay : -500ns
Motu UL Device one CHannel 2 to Motu Device two Channel 2: 73.5us
Motu UL Device one Channel 6 to Motu Device two Channel 6: -32.5us
Motu UL Device two Channel 2 to Motu Device two Channel 6: -41.5us
Motu UL Device two All 4 Channels on ADAT have -500ns in Between
Motu UL Device two both Channels on SPDIF have -500ns in Between

Motu Device two Channel 1 to 4 using ADAT Connection to Motu Device one
Motu Device two Channel 5 and 6 using SPDIF Connection to Motu Device one
at 96kHz

Delay Compensation on Camilla DSP with Motu UL mk5:
Add to Channel 0-9 Delay 0.00735ms (Subsample aktiv)
Add to Channel 12-13 Delay 0.00415ms (Subsamples aktiv, SPDIF OUT Channels)

And all Channels have max 500ns Delay in Between

Configuration Motu UL mk5 Device 2 CueMix5:
ADAT on Aanalogue OUT 1-4
SPDIF on Analogue OUT 5-6

Happy Eastern

Robert

 

[Frunse]

Hello,
Next Step:
How to know what HDMI Channel is what kind of Multichannel Speaker and to what Channel on CamillaDSP Inputs it is conected.
HDMI to I2S Board, mentioned on this Thread (i use the Same).

HDMI To I2S	CamillaDSP In 5.1	CamillaDSP In 7.1
1 Front Left	0 Front Left	0 Front Left
2 Front Right	1 Front Right	1 Front Right
3 Center	2 Subwoofer	2 Subwoofer
4 Subwoofer	3 Center	3 Center
5 Surround Left	4 Back Surround Left	4 Surround Left
6 Surround Right	5 Back Surround Right	5 Surround Right
7 Back Surround Left	------------	6 Back Surround Left

8 Back Surround Right

------------

7 Back Surround Right

How to Measure HDMI Multisurround on Windows with REW:
https://www.minidsp.com/applications/acoustic-measurements/206-umik-1-asio4all

I used the HDMI Output on my NVIDA GTX 1080TI on Windows10 Pro
You need to Konfigure it for 5.1 or 7.1 Speakersetup, by default it is Stereo only.

Last Picture Attached is my full Channel Setup.
All 2-way Speaker are the same and 2 Subwoofer, or without Subwoofer and instead 2 Frontspeaker 2x3 Way for 7.0

Robert

 

[Frunse]

Got this smaler HDMI Board today and will test it soon

https://de.aliexpress.com/item/1005...il_item.2.6b776368QIlqVU&gatewayAdapt=glo2deu

Tested it right now with REW and the Generator on all 8 Input Channels it works like the biger one, Channel 3 and 4 ar also rotated so Channel 3 HDMI Out on Computer is Channel 4 IN on the HDMI Board! and 4 Out is 3 IN.

Now i will play a bit of musik

Stereo Musik worked fine so far without any quirks or faults.

Aproved

Robert

 

[Frunse]

Hallo, me again.
Did the first Shematic this day.

Feature:

4-8 DLatch for Multichannel I2S Inputs
Samplerate Detection (LRCK) with Arduino and switching CamillaDSP Configuration's for right Incomming Samplerate
Communication between Raspberry and Arduino with I2C (Bidirectional I2C Level Conversion 3,3 / 5V)
I2C LCD 4x20 Character Display

This way it is Possible to make the Main-Software on Arduino or on Rasppbery (or Both) and its more Flexible i guess.

Its first Approach so may be it changed a bit on Building Prototype

Waiting for Parts again

Robert

 

[renne]

Hallo, me again.
Did the first Shematic this day.

Feature:

4-8 DLatch for Multichannel I2S Inputs
Samplerate Detection (LRCK) with Arduino and switching CamillaDSP Configuration's for right Incomming Samplerate
Communication between Raspberry and Arduino with I2C (Bidirectional I2C Level Conversion 3,3 / 5V)
I2C LCD 4x20 Character Display

This way it is Possible to make the Main-Software on Arduino or on Rasppbery (or Both) and its more Flexible i guess.

Its first Approach so may be it changed a bit on Building Prototype

Waiting for Parts again

Robert

Great work!
Do you still have plans for hardware ASRCs to allow different word-sizes and sample-rates on the I²S input and output of the RPi5?

 

[MCH]

Hallo, me again.
Did the first Shematic this day.

Feature:

4-8 DLatch for Multichannel I2S Inputs
Samplerate Detection (LRCK) with Arduino and switching CamillaDSP Configuration's for right Incomming Samplerate
Communication between Raspberry and Arduino with I2C (Bidirectional I2C Level Conversion 3,3 / 5V)
I2C LCD 4x20 Character Display

This way it is Possible to make the Main-Software on Arduino or on Rasppbery (or Both) and its more Flexible i guess.

Its first Approach so may be it changed a bit on Building Prototype

Waiting for Parts again

Robert

Glad to see good progress. I have a question, that we probably discussed like 3 times before but I don't remember now. Why don't you do the LRclk sample rate measurement directly in the RPi? Would save you a few parts and all the i2c communication with the Arduino.