Help needed to find the optimal way to use DSP capabilities of my setup

[MCH]

Hi,

My stereo + 2 subs system as of today consists on the following:

Stereo input -> toslink -> DDRC2x4 (Dirac) -> 4 channels USB -> raspberry pi w/camilladsp currently doing only volume/tone control and signal routing -> USB -> Motu Ultralite mk5 -> stero amp + 2 subs

I know, it is a bit convoluted, but for different reasons, the hardware setup needs to be like this.

I remember when I set it up I tried doing the xovers in camilladsp and a couple other things but for whatever reason I don’t remember I settled on this. I am happy with the result so no concerns here.

There is nothing that I hate more than the measuring and running room correction, but because I will be changing some things soon, I will have to do it again, so I thought that I could be a good chance to try to get the most of the two DSP steps that I have in the chain. But I have no idea about what would be the best way nor the correct way to do the measurements.

I wanted to launch some thoughts so that you guys can tell me if they make sense or not or how to change them to work best:

1. Do the xovers and an initial “manual” correction in Camilladsp with REW and then run Dirac on top of that. This is the sequence of steps I was planning to follow:
   1. Disconnect the DDRC2x4 and do the xovers and initial REW correction The signal path would be:

Computer running REW and UMIK -> USB to spdif card -> spdif -> Motu Ultralite mk5 -> USB -> raspberry pi/camilladsp xovers and eq -> USB -> Motu ultralite mk5 -> amp+subs

1. Once I have the xovers and initial room coorection set up in camilladsp, connect again the DDRC2x4 to set up Dirac filters on top of it:

Computer running minidsp Dirac plugin and UMIK -> USB -> DDRC2x4 (dirac 2 channels full range) -> 2 channels analog out -> Motu Ultralite mk5 analog in -> USB -> camilladsp xovers and REW made EQ -> USB -> motu ultralite mk5 -> amp+subs

1. Once I have the Dirac filters, load them in the DDRC2x4 and reassemble the system as it initially was, but with the DSP tasks now shared with camilladsp:

Stereo input -> toslink -> DDRC2x4 (Dirac) -> 2 channels USB -> raspberry pi w/camilladsp doing xovers and REW made EQ -> USB -> Motu Ultralite mk5 -> stero amp + 2 subs

Does this have any chance to work?
If not, is there anything I could do differently so that it works?

Can it bring any benefit over the current setup?
If not, do you have other suggestions?

Thanks for your comments/questions/input in general.

 

[JeremyFife]

Bumped, this is beyond my ken

 

[Keith_W]

Hello MCH. I don't know if you have found a solution to your question, but I have to admit my head spun when I read the complexity of your setup. There are multiple redundancies there and I can see no good reason to have function replicated across multiple devices, e.g. DSP performed twice on two separate devices, with two different DSP packages (REW and Dirac). Why do you need a USB to SPDIF converter as input to your Motu, when the Motu can accept USB directly? I can only see disadvantages - VERY SERIOUS disadvantages! It muddies the waters considerably, makes measurement difficult, and makes filter generation difficult. Devices that perform DSP have to be in the signal chain during measurement, and use of a USB mic + multiple USB devices makes timing and phase measurements and correction all but impossible. It adds expense, increases points of failure, reduces usability, and increases potential for audible degradation.

May I suggest that you keep it simple?

For measurement:
PC (running REW) -- (USB) --> Motu Ultralite Mk.5 --> amps and speakers
Microphone (UMIK) --> PC (running REW)

For playback, I see that you need a toslink input. You could do this:
toslink --> Motu Ultralite Mk.5 (loopback) --> Raspberry Pi (with CamillaDSP) --> USB --> Motu Ultralite Mk.5 --> amps and speakers

Even better, I would suggest you sell the UMIK (I hate USB microphones!!) and sell the MiniDSP (it is less capable than your RPi running Camilla) and use the money to buy a calibrated 48V Phantom Power mic like the iSemCon EMX-7150. Or if you want to save money, a calibrated Behringer ECM8000. Your Motu Ultralite Mk.5 already has a mic preamp capable of powering these microphones. Decide if you want to use Dirac or REW and use only one software package (I suggest you use REW). Alternatively, Audiolense, Acourate, Focus Fidelity, etc. if you feel like spending money.

If you want to add complexity, you need to have very good reasons. Those reasons are what we should be discussing, because I can not make sense of what you are trying to do here. Sorry if I sound harsh, I do not mean to come across that way.

 

[MCH]

hey @Keith_W , no worries, i appreciate your answer, i will try to answer your questions.
I am running this setup for 1 or 2 years now and it works smoothly without any issues whatsoever, everything turns on and off by pressing one remote control button and it does exactly as i want and need. Everything is signal sense and goes to standby when not in use, and everything is super robust and i had zero issues so far. To change or eliminate any of the devices is out of question.

Why do i have the minidsp? two reasons:
1. I have multiple spdif sources (TV, a CD player, streamer and a hdmi extractor), each one outputs a different sample rate, that everyone using camilladsp knows is a pain in the rear, and the ARSC of the minidsp takes all that and does rate conversion smoothly, no clicks, pops, nothing to worry about. On top of that, i am using a diy spdif switch with signal sense, so i don't even need to bother selecting source, y press "play" in the CD player, and there you go.
2. Dirac live: when i started using REW i was very happy. My room has a strong mode at 50Hz and being able to solve that changed my (audio) life forever. But besides that, i never managed to get 100% satisfying results. I am very noob and there are many things about room correction that i don't understand, and to be honest, i don't have the interest to learn about it. As i said in post #1, i hate doing measurements. At the time, saturday mornings was measurement day. I hated it, my family hated it. I decided to try Dirac, and bip bip bip, in one hour i had something much better than what i had achieved with a lot of effort before that. No way i want to renounce to this now.

Why do i use an UMIK?
when i started with REW i only had a topping DAC, no way to use a non USB mic. I bought an UMIK and i never felt the need to change it, but i hear what you are saying.

Why do i have a Motu UL?
When i bough two subwoofers, before having the minidsp, i bought a MOTU M4, that served me well. But then i had the chance to get an UL with a very good discount (a VERY good one). And as i always had in mind to set a multichannel system when we move to a bigger place (next year), well, i said, now or never.

Why do i use a raspberry pi and camilladsp?
I started using it before having the minidsp. but still today it does all the signal selection and routing, streamer, system on/off (the raspberry pi) volume control, remote control, display, tone control (camilladsp). It is the heart of the system, it works perfect and i don't see a reason to change it. Remember the Motu does not have remote control and the minidsp does not have display, and for me the lack of either is very inconvenient.

Why do you need a USB to SPDIF converter as input to your Motu, when the Motu can accept USB directly?

May I suggest that you keep it simple?

For measurement:
PC (running REW) -- (USB) --> Motu Ultralite Mk.5 --> amps and speakers
Microphone (UMIK) --> PC (running REW)

I thought on using the USB to SPDIF only to do the measurements because i don't know any other way to do the room correction on top of the camilladsp crossovers. remember i have two subs. Maybe i can set up the crossovers in whatever software in the computer and do the measurements and room correction as you say computer > motu > amps > speakers+subs? I have no idea about audio in PC in general, and i touch REW with a stick only any time i must do it. But i hear any advice

For playback, I see that you need a toslink input. You could do this:
toslink --> Motu Ultralite Mk.5 (loopback) --> Raspberry Pi (with CamillaDSP) --> USB --> Motu Ultralite Mk.5 --> amps and speakers

wont work with multiple sample rates + for the reasons above, i want to keep dirac

In general: why do i want to run Dirac and camilladsp simultaneously? It is not that i want, i was just wondering if there would be any advantage and given that i will be redoing the measurements and i already have the necessary hardware, i thought i would ask here if there is a good idea and how to do it. But if it is not, well, much easier then

 

[mdsimon2]

What you describe will work, I used a similar workflow a few years ago -> https://www.audiosciencereview.com/forum/index.php?posts/851557/.

Thinking about it a bit more, the only potential issue I see is the magnitude / phase response of the DDRC-24 DAC / UL Mk5 ADC would be considered in the Dirac correction. However, I believe both the DDRC-24 DAC and UL Mk5 ADC filters are linear phase so I don’t think there should be much impact.

In any case it is easy to view the electrical response of the DDRC-24 correction using the bidirectional USB audio and REW. I’d probably run a quick frequency response sweep in REW using the DDRC-24 as a DAC and the UL Mk5 as an ADC to see what it looks like. If you really wanted to you could also develop / apply a FIR filter in CamillaDSP to completely flatten the magnitude / phase response from the DDRC-24 DAC / UL Mk5 ADC in the audible band using REW impulse inversion and only use it when developing the DDRC-24 correction. I show the results of such correction using a UL Mk5 analog loopback here -> https://www.audiosciencereview.com/...sparent-adc-actually-exist.52223/post-1884969.

Michael

 

[MCH]

What you describe will work, I used a similar workflow a few years ago -> https://www.audiosciencereview.com/forum/index.php?posts/851557/.

Thinking about it a bit more, the only potential issue I see is the magnitude / phase response of the DDRC-24 DAC / UL Mk5 ADC would be considered in the Dirac correction. However, I believe both the DDRC-24 DAC and UL Mk5 ADC filters are linear phase so I don’t think there should be much impact.

In any case it is easy to view the electrical response of the DDRC-24 correction using the bidirectional USB audio and REW. I’d probably run a quick frequency response sweep in REW using the DDRC-24 as a DAC and the UL Mk5 as an ADC to see what it looks like. If you really wanted to you could also develop / apply a FIR filter in CamillaDSP to completely flatten the magnitude / phase response from the DDRC-24 DAC / UL Mk5 ADC in the audible band using REW impulse inversion and only use it when developing the DDRC-24 correction. I show the results of such correction using a UL Mk5 analog loopback here -> https://www.audiosciencereview.com/...sparent-adc-actually-exist.52223/post-1884969.

Michael

Thanks Michael, much appreciated. I will study your answer carefully over the weekend and come back with any questions.