Windows: Tightness Of Midi-output

Sure.

Getting tight MIDI timing on Windows is really not so easy. Windows was never made for these user-level multimedia high resolution timings. You dont want to know the details - really. Anyway, I think we should focus this problem for an upcoming release and try to improve the timing. Using Directmusic is one thing that could be tested.


For now, to get good timing on Windows (On OSX the OS takes care about this and does an
exellent job) try the following:

  • When using ASIO, use as small buffersizes as possible. The smaller the buffers are, the less jittering will happen

  • Try to switch between DirectSound and ASIO. Directsound will in some cases (when the ASIO buffers are too big for example) result in a better MIDI timing (less jitter). This depends on the drivers and the system. Hard to explain - simply try which one does a better job with your setup. The buffersizes with Directsound will have no effect on the jittering though.

Thank you for replying Taktik! Sounds promising with a future focus on the problem. I also wasn’t aware of the trouble getting jitterfree-less timing on PC. I guess one gets used to how good it actually works in Cubase, of which I’m sure they’ve spent quite some development-time on. Anyway, it won’t be long before I register Renoise, so I can try out the tips you gave. Exciting. B)

oh wow. i’m suddenly really happy i got a powerbook. :P

Hi,

I’m using a Delta 44 and Ees Pc midi 2/4 interface (Parallel port) and i’ve got a better MIDI timing with Renoise than Cubase.

I’ve slaved my MIDI timing to my Delta 44’s audio clock to get this result and had to use Win2k instead of WinXP, because my MIDI interface works with interrups (using interrups under XP at LPT1 makes MIDI timing, in my case, very unstable).

Same problem here :( in the meantime i can’t use renoise to control my hardware with exact midi timing… :(

Hoping and prayin’ that devs focuse this bug and fix it in upcoming 2.0 release!

For the ordinary midi bus control, there might be room for improvement.
For USB i don’t think any host can perform an exact timing synchronisation, at least not without hickups and quirks.
It might also depends on how may USB devices you have connected to the same bus though.
But using a MIDI device through USB in general doesn’t seem to be an advisable setup from my point of view, certainly when having specific chipsets that are known to be troublesome, but i believe Bantai made a post about this somewhere.

Yeah! Please, improove and fix just ordinary MIDI bus control. It’s all we need for normal midi gear functioning with Renoise.

If that would be so easy on Windows, this already had been implemented, belive me. Getting tight timing on Windows is a pain. The OS was never made to support high precision multimedia timing.
We are using the oldschool MIDI interface in Windows (MME), like most others do, which gives no opportunities for better timings. Using DirectMusic (Cubase uses this) could fix/improve that, but will bring even more problems, as not many MIDI drivers support DirectMusic. To get a really tight timing, you will need to support the MIDI via a non standard Interface. Thats what Cubase does with its own distributed Harware, but thats nothing we can do.

For now you can improve the timing in Renoise by using very low Audio latencies. The lower the Audio latency, the less is the jitter you get. We will try our best to fix this in future. I think Vista has some opportunities to get that really fixed.

OSX doesnt suffer from all these problems.

Thanks Taktik! I understand the situation. Hoping for better !

Answer added to the timing issue faq.

Worth adding perhaps, though i agree on what’s been said about the slightly erratic midi behaviour of renoise, is that at least the earlier Yamaha DX7 synthesizers have pretty poor timing. I don’t know if this carried over to the DX7II but if this is the case, it might make matters even worse for the original poster.

polardark: I hear what you’re saying - it’s an error-source that’s needed to be thought of. I’ve had my share of synths being sloppy over MIDI, with Roland JX-10 being one of the worst examples. Here however I made sure that the synth itself wasn’t being the main problem, and the Cubase example made it clear that it really could be very responsive and that the poor timing was dependent on something other than that.

Seems the original files aren’t online anymore (I didn’t have much webspace then) but I could upload the files again if there’s a demand for it. I think I still have them somewhere.

I’m sure there will be a DX7 vsti that could compensate this matter.
But yeah, it will be at the expense of CPU consumption.

I just read the first 2 post… and i agree

I have the same feeling when i use renoise for sequencing a yamaha a5000 sampler. ANd its not due to my hardware, i use the midi out of a creamware pulsar 2… with very low latency.

Hope this will be sorted one day.