Non-consumer (Pro/Live) approaches to DSP?

[adam_dan]

Now that I've graduated college and have a little more time and money, I figure it's time I start putting together my dream listening room. I've been poking around the forums a bit, and have been trying to figure out how to solve the DSP link in my signal chain. Unfortunately, the DAC8 pro has too few channels for my needs - here's the system I'm thinking of:

- LX521.4 with all the signal processing in the digital domain (the balanced analog signal processor is proprietary and I don't particularly want to go through multiple stages of circuit prototyping for an analog crossover) - 8 Channels (L/R Hi/HiMid/LoMid/Low)
- 4 Sealed subwoofers, still working out the details on these, but they will require 4 DSP channels.


This brings me up to 12 analog outputs, which it seems like I could do with something like the Motu 16A + a mac mini running CamillaDSP. However, here are my "wants"

- Sonically transparent DAC
- Analog inputs, switchable without logging into a computer.
- A physical volume control, ideally with a remote and a display to show volume level.
- Dirac Live? If I have a standalone DSP solution I'd be OK creating all these filters myself with other software (MSO + Audiolense)
- Balanced interconnects throughout. I'd like to have XLR cables running to 5-channel "monoblocks" next to each main speaker, and to each sub, with integrated plate amps, and have all the DSP hardware off to the side with my turntable (I like vinyl, I know it sounds worse, don't come at me haha)

Curious if the audio scientists have any ideas on hardware I could use in this case. I've been looking at the Symetrix Prism line, but would like high sampling rates, bit depth, and SNR.

 

[Sokel]

I've been looking at the Symetrix Prism line, but would like high sampling rates

Particularly on this,have a look at Merging,can do DXD,DSD,etc with the right add-on cards.

Merging Technologies | Hapi MkII Mic Pre & AD/DA Converter with RAVENNA & AES67

Hapi MkII is a small format RAVENNA and AES67 compatible Network Mic Preamp and ADDA converter with operating sample rates from 44.1 to 384kHz DXD/DSD256. Supporting the same I/O options as its big brother, Horus; Hapi is perfect for smaller budgets where compromise on sound is not an option.

www.merging.com

 

[adam_dan]

Particularly on this,have a look at Merging,can do DXD,DSD,etc with the right add-on cards.

Merging Technologies | Hapi MkII Mic Pre & AD/DA Converter with RAVENNA & AES67

Hapi MkII is a small format RAVENNA and AES67 compatible Network Mic Preamp and ADDA converter with operating sample rates from 44.1 to 384kHz DXD/DSD256. Supporting the same I/O options as its big brother, Horus; Hapi is perfect for smaller budgets where compromise on sound is not an option.

www.merging.com

This is quite interesting, I like the add-on card approach, could start with 8 channels, and the DACs seem fantastic if they perform as stated. This still necessitates a PC in the signal chain for DSP and crossover duty though.

 

[Sokel]

This is quite interesting, I like the add-on card approach, could start with 8 channels, and the DACs seem fantastic if they perform as stated. This still necessitates a PC in the signal chain for DSP and crossover duty though.

I think that the kind of control you want can't avoid a PC.
Note that these need RAVENNA (not USB) and can be completely controlled over LAN.

You can also include the analog inputs you like and are usually rare at such schemes.
Most important is that as a pro interface adds an extra layer of safety over your gear.

 

[mdsimon2]

That is a pretty tough list of requirements.

Most pro audio DSP stuff like QSC Q-sys doesn't have exceptional analog performance and pro audio interfaces rarely have remote control.

If you went with CamillaDSP on a raspberry pi it is pretty easy to add a display / remote control (see link in my signature) along with input switching. But this also has drawbacks. In my experience the MOTU AVB line (like the 16A) doesn't play well linux and you give up the ability to use Dirac. On the plus side you would have more than enough FIR power to do anything you wanted. Something like a MOTU Ultralite Mk5 has decent analog performance, 10 analog outputs + enough digital outputs where you could add other DACs to reach your 12 analog outputs. However, you are limited to 96 kHz if you want to use the digital outputs.

Michael

 

[gnarly]

Two basic paths it seems, studio or install.

I don't know much about studio, because I wanted a path that could handle any conceivable I/O, any conceivable system setup, and use pro-audio networked amps.
Should say DACs/ high sample rates are not on my radar, other than good solid specs. Rather have good dsp specs, and huge processing capability.
So install for me. Q-sys, or .....Symmetrix, Xilica, Biamp, BSS-BLU, etc.
Different world than home audio... and offers a big new experience ime.

 

[adam_dan]

That is a pretty tough list of requirements.

Most pro audio DSP stuff like QSC Q-sys doesn't have exceptional analog performance and pro audio interfaces rarely have remote control.

If you went with CamillaDSP on a raspberry pi it is pretty easy to add a display / remote control (see link in my signature) along with input switching. But this also has drawbacks. In my experience the MOTU AVB line (like the 16A) doesn't play well linux and you give up the ability to use Dirac. On the plus side you would have more than enough FIR power to do anything you wanted. Something like a MOTU Ultralite Mk5 has decent analog performance, 10 analog outputs + enough digital outputs where you could add other DACs to reach your 12 analog outputs. However, you are limited to 96 kHz if you want to use the digital outputs.

Michael

Michael,

I've spent a lot of time reading your posts, both here and on OPLUG, thanks for chiming in. I do think that a PC-based solution is going to be the way to go with this one, and it comes down to finding interfaces with enough outputs that work on linux (I'd go for MacOS if I could tolerate a fan in my listening room, but will probably do a linux PC and have it double for video capabilities). Screw Dirac, I'll learn to do room correction myself.

I assume if one is software-inclined, adding a display+remote to any linux box should be relatively straightforward.

 

[stultifera navis]

Spatial Audio Real-Time Applications

A collection of open-source VST/LV2 audio plug-ins for producing and reproducing spatial sound scenes.

leomccormack.github.io

Cavern

Download a free room correction software and Dolby Atmos decoder for PC. Improve the sound quality of your home theater and enhance your listening experience.

cavern.sbence.hu

 

[mdsimon2]

Michael,

I've spent a lot of time reading your posts, both here and on OPLUG, thanks for chiming in. I do think that a PC-based solution is going to be the way to go with this one, and it comes down to finding interfaces with enough outputs that work on linux (I'd go for MacOS if I could tolerate a fan in my listening room, but will probably do a linux PC and have it double for video capabilities). Screw Dirac, I'll learn to do room correction myself.

I assume if one is software-inclined, adding a display+remote to any linux box should be relatively straightforward.

The two advantages a raspberry pi has over a conventional Mac / PC is small form factor and GPIO pins which making adding a simple OLED or LCD display very easy.

Mac is also a great a platform for CamillaDSP but I haven't personally looked in to integrating a remote / display with Mac. I am a big fan of conventional IR volume control with a simple volume display. I purchased a USB to GPIO board at the suggestion of @jdubs but I haven't gotten around to experimenting with it yet -> https://www.adafruit.com/product/2264. That might be a good way to use Mac and implement a simply display. I am sure there are other ways to implement an HDMI display but from what I've seen they are too large for my liking.

Michael

 

[stultifera navis]

https://eu.mouser.com/c/semiconductors/audio-ics/audio-dsps/

https://eu.mouser.com/c/semiconductors/audio-ics/audio-d-a-converter-ics/

 

[adam_dan]

The two advantages a raspberry pi has over a conventional Mac / PC is small form factor and GPIO pins which making adding a simple OLED or LCD display very easy.

Mac is also a great a platform for CamillaDSP but I haven't personally looked in to integrating a remote / display with Mac. I am a big fan of conventional IR volume control with a simple volume display. I purchased a USB to GPIO board at the suggestion of @jdubs but I haven't gotten around to experimenting with it yet -> https://www.adafruit.com/product/2264. That might be a good way to use Mac and implement a simply display. I am sure there are other ways to implement an HDMI display but from what I've seen they are too large for my liking.

Michael

USB to GPIO can't be that hard. Do you know where the code for the rpi GPIO lives/runs, or more generally, how can I hook external software into camilla? I'm imagining a sort of "control console" that lives in my rack, that talks to the DSP linux instance over USB, and in there just have an arduino to take care of IR receiver + volume knob functionality.

 

[mdsimon2]

@gnarly How do you do volume control and indication with Q-sys? Over some sort of network control? Or upstream analog preamp?

Always interested in how people make it work with the pro audio stuff.

Michael

 

[adam_dan]

https://eu.mouser.com/c/semiconductors/audio-ics/audio-dsps/

https://eu.mouser.com/c/semiconductors/audio-ics/audio-d-a-converter-ics/

If this isn't a joke, I'm not particularly trying to write TOO much code, but would consider it if USB weren't so annoying to implement. Some of the SHARC chips seem to have the DSP horsepower to meet my needs, and C++ libraries so I'm not writing assembly. However, there MUST be a way to do this with off the shelf hardware (The merging stuff seems promising, but I have no clue about RAVENNA and am hesitant to introduce proprietary formats, even digital, into my signal chain.)



Right now, I think my best bet is the RME Fireface UFX II. I can take the subwoofer outputs from the headphone jacks on the front, and use balanced driver chips to drive long XLR runs to sub amps. I've also heard it's compatible on linux.

 

[mdsimon2]

USB to GPIO can't be that hard. Do you know where the code for the rpi GPIO lives/runs, or more generally, how can I hook external software into camilla? I'm imagining a sort of "control console" that lives in my rack, that talks to the DSP linux instance over USB, and in there just have an arduino to take care of IR receiver + volume knob functionality.

I use pycamilladsp -> https://github.com/HEnquist/pycamilladsp to control CamillaDSP via python and also use it to retrieve data from CamillaDSP. I then use lgpio python -> http://abyz.me.uk/lg/py_lgpio.html to control an OLED display.

I agree that getting USB to GPIO working shouldn't be that difficult, I just haven't been very motivated to do it as all my setups use RPis. That Adafruit board I linked also uses python so porting the display code should be pretty easy. Prior to the RPi5 I used other python GPIO libraries and porting the code between them was simple. If you look at my oled.py code, all that needs to be ported is 1) claiming / assigning GPIO output pins and 2) writing to those GPIO output pins in the data / command functions. Everything else could be used exactly as-is (or modified for your requirements).

Michael

 

[Keith_W]

Now that I've graduated college and have a little more time and money, I figure it's time I start putting together my dream listening room. I've been poking around the forums a bit, and have been trying to figure out how to solve the DSP link in my signal chain. Unfortunately, the DAC8 pro has too few channels for my needs - here's the system I'm thinking of:

- LX521.4 with all the signal processing in the digital domain (the balanced analog signal processor is proprietary and I don't particularly want to go through multiple stages of circuit prototyping for an analog crossover) - 8 Channels (L/R Hi/HiMid/LoMid/Low)
- 4 Sealed subwoofers, still working out the details on these, but they will require 4 DSP channels.


This brings me up to 12 analog outputs, which it seems like I could do with something like the Motu 16A + a mac mini running CamillaDSP. However, here are my "wants"

- Sonically transparent DAC
- Analog inputs, switchable without logging into a computer.
- A physical volume control, ideally with a remote and a display to show volume level.
- Dirac Live? If I have a standalone DSP solution I'd be OK creating all these filters myself with other software (MSO + Audiolense)
- Balanced interconnects throughout. I'd like to have XLR cables running to 5-channel "monoblocks" next to each main speaker, and to each sub, with integrated plate amps, and have all the DSP hardware off to the side with my turntable (I like vinyl, I know it sounds worse, don't come at me haha)

Curious if the audio scientists have any ideas on hardware I could use in this case. I've been looking at the Symetrix Prism line, but would like high sampling rates, bit depth, and SNR.

I can't think of any product which fulfils all your requirements, sorry. As mentioned, if you want 12 channels you have to go pro audio. And pro audio tends to lack consumer features like physical volume controls and remotes.

If you want a remote, implement it in your PC. There are many ways to do this, and it all depends on what software you are using for playback. Some (like Roon and JRiver) have an app. Others (Camilla, Acourate Convolver) have a web interface. You may be able to build in an IR remote into your PC using FLIRC. FLIRC translates IR remote commands into keyboard strokes, so it then depends on whether your software supports keyboard shortcuts for remote control. I personally use BubbleUPNP, because it is the "least worst" solution.

I wonder what other ASR members think about the importance of SNR/SINAD in multichannel amps which are going to be used for DSP. I personally don't care about SINAD at all, because if it is low enough (< -100dB) the chances of audible noise/harmonic distortion is extremely slim. But in theory: digital volume control removes the least significant bit, effectively reducing the bit depth and bringing the signal closer to the noise floor. (1) is this a problem if you DSP if you implement volume cuts AND use digital volume control, (2) can this be mitigated by increasing bit depth?

I see that Merging has been mentioned. Here are some pros/cons of Merging:

Pros:
- has Ravenna. This IMHO is the killer feature. This means your system is almost infinitely expandable. You can gang together dozens of DAC's and Ravenna products and send audio to all of them from one PC. Probably overkill for home users.
- DSD support. This depends on whether you think DSD is a pro or not.
- Excellent performance.

Cons:
- has Ravenna. Even the pro guys struggle to get it working. No USB input.
- horrendously expensive. Try pricing up a Hapi with enough option cards for 12 digital outputs.
- output is via D-Sub which means you need a breakout cable.

Two Merging Anubis interfaces ganged together with Ravenna will give you 12 digital outs. It is also cheaper than a single Hapi once the Hapi is optioned up with the cards that you need.

 

[adam_dan]

Volume control in the PC seems like the way to go. Maybe

I can't think of any product which fulfils all your requirements, sorry. As mentioned, if you want 12 channels you have to go pro audio. And pro audio tends to lack consumer features like physical volume controls and remotes.

If you want a remote, implement it in your PC. There are many ways to do this, and it all depends on what software you are using for playback. Some (like Roon and JRiver) have an app. Others (Camilla, Acourate Convolver) have a web interface. You may be able to build in an IR remote into your PC using FLIRC. FLIRC translates IR remote commands into keyboard strokes, so it then depends on whether your software supports keyboard shortcuts for remote control. I personally use BubbleUPNP, because it is the "least worst" solution.

I wonder what other ASR members think about the importance of SNR/SINAD in multichannel amps which are going to be used for DSP. I personally don't care about SINAD at all, because if it is low enough (< -100dB) the chances of audible noise/harmonic distortion is extremely slim. But in theory: digital volume control removes the least significant bit, effectively reducing the bit depth and bringing the signal closer to the noise floor. (1) is this a problem if you DSP if you implement volume cuts AND use digital volume control, (2) can this be mitigated by increasing bit depth?

I see that Merging has been mentioned. Here are some pros/cons of Merging:

Pros:
- has Ravenna. This IMHO is the killer feature. This means your system is almost infinitely expandable. You can gang together dozens of DAC's and Ravenna products and send audio to all of them from one PC. Probably overkill for home users.
- DSD support. This depends on whether you think DSD is a pro or not.
- Excellent performance.

Cons:
- has Ravenna. Even the pro guys struggle to get it working. No USB input.
- horrendously expensive. Try pricing up a Hapi with enough option cards for 12 digital outputs.
- output is via D-Sub which means you need a breakout cable.

Two Merging Anubis interfaces ganged together with Ravenna will give you 12 digital outs. It is also cheaper than a single Hapi once the Hapi is optioned up with the cards that you need.

Volume control in the box seems like the way to go. Maybe I can even do a motorized pot into ADC and use a remote to trigger the motor, make it feel like an analog control haha.

The merging stuff seems interesting... Does it just hijack the ethernet port on a computer? Seems targeted towards large live sound installs, which is probably the closest thing to what I'm going for. Keeping bit depth and SNR well beyond audible limits should allow me to not stress about fidelity lost with digital volume control. However, it is HORRENDOUSLY expensive... I could do a fireface UFX III for that price, which seems to meet my needs spec-wise, and should work in CC mode on linux.

Fireface UFX III | High-end USB Audio Interface - rme-usa.com

RME Audio Professional Live, Studio, Recording and Broadcast Solutions. Unrivalled Quality, Performance & Stability MADI Interfaces, Converters & Preamps.

www.rme-usa.com

 

[mdsimon2]

I wonder what other ASR members think about the importance of SNR/SINAD in multichannel amps which are going to be used for DSP. I personally don't care about SINAD at all, because if it is low enough (< -100dB) the chances of audible noise/harmonic distortion is extremely slim. But in theory: digital volume control removes the least significant bit, effectively reducing the bit depth and bringing the signal closer to the noise floor. (1) is this a problem if you DSP if you implement volume cuts AND use digital volume control, (2) can this be mitigated by increasing bit depth?

I agree that THD+N doesn't really matter, because THD doesn't really matter, but noise performance is rather important. Some of these pro audio interfaces have a lot of noise that will be audible with decently sensitive unpadded tweeters. In some ways pro audio interfaces are worse because they often have high output voltage, so even if the dynamic range looks good (100+ dB), if you calculate residual noise it may be high in absolute sense. Newer interfaces are much better in this regard, but something like a Focusrite 18i20 2nd gen has a spec'd DR of 108 dB(A) which translates in to 24 uV. When you multiply that by amplifier gain you end up with a lot of noise at the speaker terminals.

Of course there are other ways of solving these issues. Sensitive drivers can use lower gain amplifiers and you can also attenuate using a voltage divider at the amplifier input. Some interfaces have the option to reduce output voltage which may result in lower dynamic range but improved residual noise.

I don't really think DSP has an impact unless you are doing something weird like cutting a lot in DSP and making up for it with a high gain amplifier.

Michael

 

[Keith_W]

Volume control in the PC seems like the way to go. Maybe

It is the ONLY way to go unfortunately! Unless you want to keep getting up to twiddle the volume knob on your interface. Think about where the volume control needs to be - either before the convolver (player software), or in the convolver. If it's in the multichannel DAC, you need one that controls all the outputs simultaneously and has a knob and remote control. I believe that Merging has a remote control (web interface for their mixer software? Can someone confirm?). I have an RME Fireface UC, and I don't think it has a web app. Maybe @MC_RME can chime in.

The merging stuff seems interesting... Does it just hijack the ethernet port on a computer?

It does. So you will need a switch. Not only that, you need a managed switch.

Seems targeted towards large live sound installs, which is probably the closest thing to what I'm going for. Keeping bit depth and SNR well beyond audible limits should allow me to not stress about fidelity lost with digital volume control. However, it is HORRENDOUSLY expensive... I could do a fireface UFX III for that price, which seems to meet my needs spec-wise, and should work in CC mode on linux.

Fireface UFX III | High-end USB Audio Interface - rme-usa.com

RME Audio Professional Live, Studio, Recording and Broadcast Solutions. Unrivalled Quality, Performance & Stability MADI Interfaces, Converters & Preamps.

www.rme-usa.com

Seems like a good product that will fit your needs. RME makes great stuff. Perhaps MC_RME can tell us whether you can expand the Fireface UFX3 by connecting it to another RME product to get more DAC channels? Can you use the MADI output as an input to a second multichannel DAC?

 

[EERecordist]

Are you aware of https://www.hairballaudio.com/catalog/linkwitz-asp4? It's drop in analog line level crossover for those speakers.

 

[Sokel]

Volume control in the PC seems like the way to go. Maybe

Volume control in the box seems like the way to go. Maybe I can even do a motorized pot into ADC and use a remote to trigger the motor, make it feel like an analog control haha.

The merging stuff seems interesting... Does it just hijack the ethernet port on a computer? Seems targeted towards large live sound installs, which is probably the closest thing to what I'm going for. Keeping bit depth and SNR well beyond audible limits should allow me to not stress about fidelity lost with digital volume control. However, it is HORRENDOUSLY expensive... I could do a fireface UFX III for that price, which seems to meet my needs spec-wise, and should work in CC mode on linux.

Fireface UFX III | High-end USB Audio Interface - rme-usa.com

RME Audio Professional Live, Studio, Recording and Broadcast Solutions. Unrivalled Quality, Performance & Stability MADI Interfaces, Converters & Preamps.

www.rme-usa.com

@Kal Rubinson can gives us an insight to Merging as he's using one.

(think he also reviewed it)*

*Edit: Yes,here:

Merging Hapi MkII multichannel digital processor

I've been running a 5.1 system for years. Recently, I expanded it to 5.3 with the addition of two more subwoofers. This system is manageable with one of my eight-channel DACs. Even more recently, I dipped my toe into Dolby Atmos, which made it necessary to lash up at least four more channels...

www.stereophile.com