Using a Model 12 with Linux

[cakes]

Hello, I'm new here. I've been reading the forum for a bit, looking for information on this topic. I was unable to find the answers I'm looking for and the search here is... incomplete, to put it nicely. So, I apologize if I've missed a previous thread about this exact topic.

I am putting together a home studio. The setup I am building is: amps into IR boxes into a board, vdrums strait into board via stereo out (or MIDI into board input to the DAW), vocals through mic pre to board. Board will be used with Bitwig. I am currently using a Behringer UMC404, but it's time to upgrade to something with more inputs and versatility. The Model 12 looks like the perfect board for this.

The reason I like this board from what I've read, is the multi-use functionality. I intend to use it for band practice (as the hub for guitar/bass/vdrum/vocals input, output to studio monitors or individual in-ear monitors), demoing and recording, mixing, and if possible mastering (though I'm no dark wizard in that arena, so it's not a big deal to send the final mixes to a pro). The no-latency is key for practice, but I also like the idea of laying down live tracks on the SD to later move to the DAW for mixing. I get about 5ms of latency on my computer through the DAW. Not bad, but I would rather use the DAW for overdubs and mixing.

Using the board as a DAW controller is a nice to have, but not required. (I would assume I'd have more than 8 tracks anyway, so I am not looking for a DAW controller.)

I'm using Linux (pop_OS, but you could assume Ubuntu) and my box has USB 2 and 3 inputs. I have read that the Model 12 has a class compliance mode, which I assume is for USB-C, which should work with the USB 3 input. The fact that Tascam doesn't have Linux drivers is not an issue if class compliance works, as not having the DAW mode available is not a big deal. I've read posts on using the Model 16 and 24 with Linux, but those models are slightly different, using (I believe) USB 2 output, and I know that already works well with Linux class compliance. I mean, if it's class compliant, it should work, just looking to see what others may have come across using the Model 12 with Linux.

What I'm really concerned with is the ability to get this to work with JACK, and be able to export recorded audio from the SD card. As long as the card is Fat32, I think things are kosher, though I don't have a card reader, so curious as to how I would import that in. Would the Model 12 show up as a drive in Storage mode? I assume if it's FAT32 with class compliant USB C to 3, this theoretically should work.

If push comes to shove, I can move the DAW to Windows 10 on a dual boot, but would rather try to use Linux.

Thanks in advance!

 

[cakes]

I spent some time setting up the Model 12 on Linux and Windows and will document here for others looking for information like this. I am not intending to use the Model 12 as a DAW controller, which is not possible as far as I know on Linux. However, I don't own any of the DAWs that are supported by the Model 12 found in the menu controls.

I tried setting up with Reaper and Bitwig as DAWs, as those are the ones I tend to use.

I found using Windows with the Model 12 frustrating. Mostly because Windows is an overall unsatisfying and annoying operating system to do any production work on. There seemed to be setup issues with the Model 12. I was able to find the Tascam Model 12 drivers and install them. Right away, it opened up as expected and had the default buffer at 256. Next, I opened up Bitwig and found the only inputs were 1 and 2, defaulted to stereo. There was no option to get all 10 channels (8 mono, 2 stereo). I then opened up Reaper and found the same result. Next, I installed JACK, which also allowed only inputs 1 and 2, strangely. Digging into it further, I found the ASIO drivers. I installed them, and was able to map inputs however I wanted. The ASIO experience felt really uninspiring and even wrong. In the end, I wasn't able to get the input busses working with ASIO. The Windows environment was having a hard time with the latency on the default settings. It didn't perform well until I bumped things a little higher, but the latency wasn't good at all. It was very disappointing, and I don't know why I even wasted my time. I'm sure there's room for optimization and testing, but I was already exhausted by the experience so I abandoned it and immediately killed the Windows partition.

Using Pop_OS, this was an entirely different story. I plugged in the Model 12 and it showed up in my sound preferences without the needing to install the Tascam drivers (I assume those are more for the DAW controls anyway). I installed JACK, started it up and found all input busses connected. I fired up Bitwig (and Reaper later) and mapped out my controls: 8 mono, 2 stereo. I plugged in my guitar, and sound was going through the DAWs. I was able to change the sample rate and the buffer size in JACK to get down to 2.6ms without any performance issues.

I should note that my computer has a Ryzen 7 w/ 8 cores @3.35ghz, with 32gb DDR4 RAM. Latency shouldn't be a huge issue here, but Windows just couldn't handle it out-of-the-box.

Why did I try using Windows? There are more options for music software and midi controllers in Windows. However, I've decided to rethink some of my workflow and get creative on Linux. The Model 12 is already a flexible enough workstation, I feel like I can get a lot done with it before I even get to the computer.

 

[urgentmeow]

Thanks for this post! I wanted to chime in and say that I had exactly the same results running Linux Mint. Plug and play! JACK saw all inputs and outputs, and I was able to send audio back and forth from Reaper to the Model 12 without issue. I didn't try DAW mode but, like you, I find it unnecessary.

 

[AltTuning]

Thanks for your post -- I didn't think about it until just now but I wonder if the audio glitches I'm getting when using the M12 as an audio interface would be any different in Linux and probably worth a test to see if it's a hardware issue or Windows issue.

Nice about JACK working -- I looked into it, but never figured it out...

BTW, the M12 formatted my SDXC card as exFAT, which I believe is Linux-compatible.

 

[toupeira]

I can also confirm that JACK and USB storage works just fine!

I also wanted to use a mic on one of the channels for desktop usage, which turned out to be tricky with PulseAudio because it only configures a stereo input profile by default, meaning you'll only get the sound of channels 1 + 2 (which are also the default stereo outputs, so you can't really use them as inputs anyway).

I ended up switching to Pipewire where I could select the "Pro Audio" profile, and set up virtual devices for the mic (channel 3) and also the main bus (channels 11+12), in case I want to route something else to the desktop. These input devices can then be selected in apps like Zoom. My configuration for this can be found at https://github.com/toupeira/dotfile...bb30113c/etc/pipewire/pipewire.conf#L160-L195.

This could be extended to also move the default stereo output to other channels, see the example at https://gitlab.freedesktop.org/pipe...ger-umc404hd-speakersheadphones-virtual-sinks.

Pipewire also replaces JACK and works great through the pw-jack wrapper. I only did some quick testing with Audacity but was able to record all 12 channels without glitches, with default Pipewire settings and RT enabled.

 

[toupeira]

After some more Pipewire usage I ran into crackling issues and didn't find a solution yet, this seems to be specific to the Model 12 and I submitted a bug report at https://gitlab.freedesktop.org/pipewire/pipewire/-/issues/2009.

I'd be curious to hear if others are running into the same problems with Pipewire.

For now I switched back to PulseAudio and Jack (which works perfectly at low latencies), and had to create a custom card profile so I could access all channels in PulseAudio (since it doesn't have the "Pro Audio" profile which Pipewire offers). This can be found at https://github.com/toupeira/dotfiles/blob/main/etc/pulse/tascam-model12.conf but needs some additional setup, as described on e.g. https://wiki.archlinux.org/title/Pu...t_analog_and_digital_outputs_on_the_same_card.

 

[James Paskaruk]

Hey guys, I haven't picked up one of these mixers yet but this thread is making it highly likely that I will - I need something new and Linux is not up for discussion at this point, so hearing that things work out of the box is very encouraging. I'm thinking I'll be more along the lines of 16, or possibly one of the rack devices, I have a bunch of tabs open right now and I decided to google about Linux/Tascam first, which brought me here.

Anyways I thought I'd pipe in about Pipewire - it is now early 2023 and it seems to me that Pipewire is still not quite working at the level of a real Jack system. I have spent the last month or so in heavy experimentation, though I don't consider my results conclusive in any way and I plan to revisit it very soon.

My current audio is a Scarlett 6i6, and I just joined a new band which has incited a big interest in getting my audio systems back in shape, I've been working in animation the last few years and my creative energy has mostly been focused on Python and 3D stuff so I've just been floating along, musically speaking, and I even decided to go back to Windoze for a while because I thought it would make things simpler... it did not. It just made them more expensive and not any better.

Anyways we used Ubuntu at my last job and so I've just been using Ubuntu desktops and whatever they used for audio, and rarely touched anything musical, but once I decided to get back into it the first thing I did was upgrade to 22.10, which enables Pipewire by default, and tried it out a bit - results were promising. So, I immediately downloaded Ubuntu Studio, my old go-to which I've used onstage in the past as a B3, and was shocked to find that they had not implemented Pipewire for 22.10 like mainline Ubuntu did. "But what about the future!" said I, and decided to do something else.

That something else ended up being, I tried out Arch again for the first time in probably ten years, and I have been extremely pleased with it - I'm running KDE and things are just gorgeous. I do have a decent video card. I've actually gone back and forth a few times between straight Jack with pulse plugin and Pipewire, running mainly Reaper and Bitwig with some of the more mainstream plugins like ZynAddSubFX. By the way, I also plan to pick up a Zynthian kit very soon, yall should check that out if you're into Linux audio, it's an amazing project.

Anyways, what I've found, in January 2023, is that Pipewire has very comparable DSP performance to Jack - on that front they've more or less nailed it, as far as my ears can tell. I also found that wireplumber is pretty nimble at making connections with no user input at all. That's the good.

The bad - and this might just be a matter of configuration that I'm too dense thus far to figure out - is that the session manager is way too busy, and it resulted in a very annoying moment to moment experience. Specifically, when I played music on Spotify, it would play silence for the first second of every song, because for whatever reason the session manager was disconnecting and reconnecting between every song.

It did do sane connections automatically when I would load up a DAW which I had configured with a specific amount of inputs and outputs and midi devices - I never needed to tell it what to connect, so they have basically achieved parity with ASIO in terms of ease of use, so far as I can see as well. But overall, they need to figure out a better flow for session management with more persistence built in, or else make configuration a little less opaque, either better documentation or better examples or something, cause I'm not incompetent at this but I have no idea what's going on in those files.

I do, however, understand how to use the patchbay to create whatever perpetual connections I need (granted, that is a bit opaque too, but learnable if anyone needs help), so the active session management is a thing I can live without, particularly if it comes at the cost of one second out of every song I hear. And the Jack-Pulse bridging is very stable these days, I don't even think about it anymore, once I got the necessary Arch packages installed, and that's all in the Wiki.

So for now, I'm still on Alsa/Jack/Pulse, and in retrospect, the crew over at Ubuntu Studio obviously went through the same experience as I did when they made the decision not to move forward along with the mainstream distro. I could just as easily have stuck with them in the first place, and I remain a strong advocate for Ubuntu Studio more generally, as an onramp to Linux for creatives. They do a great job at what they do, which is to wrap everything up into a big one stop package for anyone who does creative work, musical or otherwise.

If you're more sharpish and can follow directions well, I can report a very happy experience with Arch as well, and I plan to stay with it as my daily desktop for the foreseeable future.

 

[Dale Nicholson]

Hello. I had some struggles when figuring out how to use the Model 12 that I thought I would share some of what I learned in case others run into similar problems or are trying to decide if a Model 12 will work for them on Linux.

I wanted to do multitrack recording using the Model 12 as my main audio interface for analog instruments, while at the same time sending out audio from my computer to the model 12. I also wanted to be able to push midi between all the devices as necessary and do various combinations of sequencing from in the box and from external devices.

I'm using Garuda Linux with KDE Plasma, Ardour 8.1.0, Pipewire, Qpwgraph 0.6.1 for a pipewire patchbay, VCV Rack 2.4.1, Cardinal, Dexxed, OB_Xd, Hydrogen, and many more. I'm not trying to use anything that requires Carla or Wine, etc., only things that are meant to run on Linux. Most plugins just work now, whether created for Linux or not.


Multitrack Recording: Pipewire is a drop in replacement for pulse audio and Jack, and ALSA is still the underlying system. It's very confusing to many since Pirewire just implemented all the things that were in those systems so things like my DAW can still think it's using JACK. Initially I could only see two channels coming in from the Model 12. I had already changed the settings on the model 12 (menu->system->USB Audio->PC <-Multi). To get ALSA to present all the incoming channels you have to change the profile to "pro audio", default settings for my version of KDE were to "play hifi quality music" which masks the extra available channels and presents the main out instead. To change this in KDE, you just open your audio settings while you have the Model 12 connected and change the profile selection. I'm sure there's a settings file somewhere that you could change too, and other DE's no doubt have similar settings.

Firmware: Updating firmware works just like the directions say, once you switch to storage mode in the Model 12 settings the device will be mountable just like an external thumb drive in linux and you can upload the new firmware file with no issues.

PC Audio out: When I first bought my model 12, the only way to send sound from the PC was through channels 1 and 2. This was with firmware version 1.30. Starting with firmware 1.4.0 they added the ability to direct it to 9/10 or to main. In pro audio model the model 12 does not actually present the main outs however, so you really can only use channel 9/10 instead of 1 and 2 with ALSA despite main being an option on the Model 12.

Phantom Power: I have a Bass and Guitar that I wanted to record while also recording other items that do not need phantom power. The Model 12 has a button on each channel for turning it on, but there's also a global one. I did not need to use the global, I don't know why I would need that, and I'm also not sure why it's even there as you very likely never want phantom power on all channels.

Modular: I have some modular/semi modular gear too, I had no issues at all on the Model 12 with sending in the audio from modular gear. I've heard a lot of people talk about how hot a modular signal is, but I had no trouble, just adjust the incoming gain.

Record out: Don't forget about the settings for where in the signal chain you want to push into the PC for your multi tracks. You can set each channel differently or do them all globally. If you just want a clean signal to process in your DAW you can have the signal come right after the gain but before the Compressor.

DAW Control: I found this feature to not be useful to me. You have to switch into DAW control mode or keep things in normal mode, you can't mix both together. Beyond that, the Model 12 doesn't currently have a setting for Ardour.

Bluetooth: I have used the bluetooth function to play music from my phone, it works just fine. You can set it to go to channels 9/10, off, or direct to main out, no need to waste 9/10 if all you are doing is playing some music, would only need that if you wanted to use the channel strip or add effects. I can imagine some other ways to use this such as when I want to eventually integrate Koala, Animoog Z, Synth One, and Launchpad into the overall system, but that experimentation will have to wait until another day. It's not clear if the model 12 is the answer for that or just using Loopy Pro or something else instead to get the signal and midi into the PC.


Conclusions: I experimented with a lot of different configurations with a mix of midi and audio over usb, din midi, etc. In some cases I used the built in effects (bass and guitar) for others I didn't. I also sent guitar into VCV rack to process and send back the modified signal. I sequenced from VCV rack to Minibrute and Drumbrute, from MPK225 to VCV Rack, etc. Played directly into VCV rack using the keyboard to play one section in a patch while using bank A and B pads to play two other parts of the same patch in VCV. Also sequenced from various of my external gear in a lot of combinations to see what worked well and what didn't, and get a feel for how flexible the set up could be. My Roland devices send both midi and audio through the USB midi connection, while the Arturia synths only send midi, those had to have their audio out go though the model 12 and then to the PC.

You can most definitely do full scale audio production on Linux today. The main thing you have to do is be sure only to get gear that is USB compliant. I don't feel any sort of limitation, but that might be because I have only ever used Linux for audio, if I had gotten used to something else I would probably feel differently.