Forgive me if this is a stupid question, hardware connections and datasheets are still kind of a mystery to me.
Among other things, Zynthian V5 has two TRS inputs and allows you to choose between unbalanced mono, stereo TRS to mono and balanced mono for each jack. All of these result in a mono input signal per jack.
Given that it’s apparently possible to mix stereo TRS into a mono channel, I was wondering if it would theoretically be possible to interpret the TRS signals as stereo and use them as two separate input channels per plug. Or is this electrical nonsense, and the stereo-to-mono mix is happening in the ADC circuits without ZynthianOS/Linux/JACK ever being able to access the stereo signals individually?
Background is that I want to connect my sound module’s stereo outs to Zynthian, as well as a microphone. I have a USB interface, but it comes with the obvious latency drawbacks. It’s a bit unfortunate that the plugs handle TRS jacks, but are limited to mono nonetheless.
Sorry, it’s not possible. The R & T can be balanced signals (a differential pair), but they can’t be independent signals. The ADC chip is 2-channels only.
Hi @jpetso, have you considered simply adding an inexpensive but effective Behringer or Yamaha 4-6 channels sub-mixer to your setup, in order to get a mixed 2-channel stereo output of your synths, before entering in the two Zynthian TRS audio inputs and then ADCs?
I’ve considered it! If I really want to mix stereo to mono, it would probably be for Zynthian’s own sounds rather than the beautiful stereo piano from the sound module. I may also in the future consider buying Pianoteq and switching entirely to Zynthian for sound generation, but that’s a different topic and we’re still talking about a Raspberry Pi 5 doing multi-track synthesis and effects.
The other issue for me is simply space: I live in a small-ish apartment, and unless I displace my TV or dining table, I only really have the top of my Yamaha Clavinova digital upright piano to fill with audio devices. So anything that doesn’t take up extra space makes a huge difference.
Right now I’m still happy that I managed to make a two-tier stack of two devices with a laptop stand, Dexibell sound module on the bottom (front panel) and MIDI fader/knob controller on top. My new Zynthian can sit upright on the music stand, where one would usually find sheet music. The Zoom U-44 audio interface is compact enough to hide behind the music stand. A mixer would also have to go there. Anyway, you can see that this is a rather makeshift amateur setup, and any simplification of the cabling setup is valuable (also in case I ever decide to bring some of that setup into a performance space).
My current plan is to abuse the audio interface’s direct monitoring functionality in a somewhat unconventional way:
Sound module connects to Zynthian via USB, but only MIDI is used.
Microphone connects to Zynthian’s native audio in jack.
Not a condenser, so I don’t need phantom power.
U-44 audio interface connects to Zynthian via USB, but I guess I only need it for its direct monitoring functionality in this setup. Essentially turning it into a mixer.
Zynthian stereo out connects to U-44 audio in, ch. 1+2.
Sound module stereo out connects to U-44 audio in, ch. 3+4.
Turn the direct monitoring mix knob on the U-44 all the way up.
U-44 line out (with direct monitoring audio from Zynthian + sound module) goes into piano speakers or headphones.
This way I should be able to get a latency-free mix of Zynthian and sound module, with no loss of stereo channels. Connecting the mic to Zynthian directly means I can still get low-latency effects or looping for vocals. Whereas the sound module is operated purely via MIDI, so any kind of looping or sequencing happens via zynseq. With this setup, I should even be able to leave the USB hotplug option off for some extra latency savings, unless I want to record sound module audio, but recording MIDI should be good enough if I want to mix something later.
Thanks for the feedback, now I’ll have to see if this is actually going to work.
Okay, so I prototyped this with my two audio cables (need to get more) and headphone. Looks like abusing the audio interface as a poor man’s 4-channel mixer is going to work.
Funny thing though.
My Zoom U-44 only has two built-in XLR/TRS combo plugs, in order to make it an actual 4+4 audio interface one needs to use the clip-on EXH-6 expander for channels 3+4. Now, if I plug in the U-44’s USB cable to my sound module to power it, and then also plug the two audio cables from the sound module into the U-44 channel expander, then I get pretty bad high-pitched noise in the direct monitoring headphone out.
Also, if I power the U-44 from the Zynthian’s USB port, and then connect Zynthian’s audio outs with the expander plugs, I get the same high-pitched noise.
However, if I USB-power the U-44 from my sound module while getting the audio cable connection from Zynthian, then the signal for U-44 channels 3+4 is crystal clear. Also, only the expander module seems to be affected by this, the built-in channel 1+2 plugs are never causing any noise.
So it looks like the USB power and audio circuits for the channel expander module are not separated, and if I connect both USB and an audio cable to the same device then I get some kind of acoustic short? Really weird!
So it looks like I’ll have to USB-power the U-44 from a different USB host if I want to connect both Zynthian and my sound module to it at the same time. No simultaneous recording of sound module audio, I guess. Perhaps I’ll use the nearby wifi router’s USB ports to power the U-44. Yes, great idea to have a radio-frequency-emitting device right next to my audio equipment
Wow, I really appreciate the thoroughness of your efforts and investigation, to consolidate your proof-of-concept for a mini studio setup that fits in your very limited living space, possibly affording portability when required.
Still, are you sure that a small, inexpensive and lightweight mixer, like the ones pictured in the first five slots of this page, wouldn’t solve you a lot of problems and give you various handy options? (like on-board FX if useful):
They are reliable, and can be easily placed in any residual space, starting from the tiny 502S.
Alternatively, you could go for a slightly costlier and really quality portable mix station, of approximately the size of 1202SFX, from Soundcraft:
It looks like I’ll have to consider some kind of proper mixer solution - turns out the direct monitoring in my audio interface mixes all input channels into both L+R, which destroys the stereo split that I was going for in the first place.
So, those Behringers are dirt cheap. And also fairly small. Definitely an option. I think the upgrade pick for my available space would be a Zoom Livetrak L6 (or better, L6max with a few worthwhile fixes), however that’s about five times as expensive so maybe it doesn’t pay off.
Hence it would be really awesome if a future version of the Zynthian soundcard could use its two TRS plugs to the fullest, allowing a choice between balanced mono or unbalanced stereo. Perhaps in the same revision that could also provide independent main vs. headphone levels.
The soundcard in the V5 (and I think V4) has 2 TRS input sockets. Each may be configured individually. It may be balanced/differential, using top & ring for hot & cold. It may by unbalance/single-ended, using tip and sleeve. It may be unbalanced, using tip + ring and sleeve. This last option allows mixing two signals to a single mono on each input. There is no control over individual level of the tip and ring signals. I like to use this unbalanced configuration because it works fine with a standard mono jack but allows a stereo source to be plugged in, giving a mono sum of the T+R.
Of you are considering buying a mixing desk to provide more inputs, remember that you still need to mix them all down to stereo or dual mono to feed the zynthian.
An alternative is to purchase a multichannel sound card to give more inputs. There are 4 channel devices for about £100 and 16 channels for a bit more. I have a Beringer UMC1820 which gives lots of inputs and outputs but prefer my Tascam US-16x08 which has DSP giving compressor and 4-band EQ on each input without any RPi CPU load. Both devices are 1U rack mount, connect via USB and are fully supported by zynthian.
So, I was going to reply that I tried a bunch of configurations before I posted this, and all the various USB setup attempts that all added perceptible latency - not like “unplayable”, just like “this feels a little off”.
But now I’m thinking maybe I wasn’t exhaustive enough, or maybe I feel the latency less today, because one of the configurations I thought I had tried is now totally working for me.
This is what works or fails in terms of latency:
Zynthian → USB audio out → sound module → from there, analog to speaker: fail, too much latency on Zynthian instruments.
Sound module → analog out → Zynthian analog in → analog to speaker: works, but doesn’t leave room for an extra mic connection.
Sound module → USB audio out → Zynthian → analog to speaker: I thought I had tried it and that it felt slow to play, but now I can’t feel much of a difference?
Sound module → analog out → USB audio interface → USB audio out → Zynthian → analog to speaker: also seems to work for me now.
So, perhaps this whole thread was based on a mistake from my end. I’m not sure where I went wrong testing all the possible permutations. I really wanted to have the speaker cable with the combined sound go out from my sound module, so that I can continue to play it without turning on the Zynthian for its audio passthrough. I guess that isn’t possible, because the Zynthian → USB audio → sound module path adds too much latency on one or both sides. The other direction seems to work though, when I thought I had tested it and judged it a fail. I have to revise that finding - Zynthian’s USB audio in (even with hotplug) is indeed fast enough for playing.
So in the end, I don’t even need my USB audio interface. I do, however, need to turn the Zynthian on if I want to just play piano with my sound module. From that point of view, maybe a stand-alone mixer is still worth considering.
Thanks everyone for indulging! I’ll show myself out now.
