Why is midi clock generated by Renoise so good?

Sorry if this is the wrong subforum, but this question has been puzzling me for ages.

Renoise has a very stable midi clock, I would say it is the best DAW for generating MIDI Beat Clock good enough to sync external instruments.

I don’t know why, but it functions almost like a rock solid hardware sequencer. Especially when using a setup with a midi to dinsync converter, Renoise will shine.

I’ve done some very non-scientific tests today which show this immediately.

My setup:

Mac Mini (end 2011) with 2.7 Ghz i7 with 4GB RAM running 10.10.2 Yosemite. Midi out from MOTU 828 mk3 firewire. Split up by Emagic AMT8 (functioning as a midi thru device), Syncing Yamaha RY30 pattern and a phil rees MIDI->Dinsync converter driving a Roland TR606 hihat pattern.

Daw’s tested:

  • Cubase Elements 7 (set to send midi clock on stop, like renoise does)

  • Renoise 3

  • Numerology

  • Logic Pro X

  • Digital Performer 8

  • Reaper V4.76/64

  • Live 9

Results:

https://soundcloud.com/rv0/sets/crappy-sync-test

Identical test, but using the ESI Midimate II instead of the Motu:

https://soundcloud.com/rv0/sets/more-crappy-sync-tests-with-ploytec

As you can see, every single DAW has terrible results compared to Renoise.

Why?

I’ve also noticed its much better than other DAWs. I usually use a syncgen and for whatever reason didn’t have it hooked up properly and had to sync a drum box, used Renoise midi clock out and it was just dead on in a way other daws just are not at all. Would be very curious as well. Perhaps just coded better all around, although I always thought it was a system thing and nothing DAWs could do about it? But your tests definitely prove that theory wrong. Just one of many reasons its my favorite DAW.

Thanks for recording those test btw!

Wow, those are impressive results. Before I got into Renoise I used to struggle all the time with getting a tight midi sync. In the end I sold all my hardware and went completely ITB so never really got to try it with Renoise.

As to why, I’m guessing it may have something to do with a tick based audio engine (if that’s what it still is under the hood?).

However the trade off may be why we have some limitations such as Signal followers only being able to modulate tracks to the right of them and limited audio routing options.

I dunno, I have a limited understanding of DSP so might be wrong…

Would be interesting to see if adding VST plugins or using Rewire affects the results.

I’ve been discussing this on IRC yesterday (#x0xb0x on freenode) and someone suggested (after checking sonogram) that it may not be jitter, just the dinsync converter not liking the clock from DAW (could be related to the order of the MIDI start signal / CLOCK pulses) and therefor syncing the 606 too late. At his request I did one more test where I pan the 606 and the RY30 to their own channel.

The first 3 files use a x0xb0x (running latest n0nx0x2) as a dinsync converter, all the others are the old Phil Rees MDS.

https://soundcloud.com/rv0/sets/more-crappy-sync-tests

Because the pattern doesn’t use any accents it does not have a “groove”, so the result is harder to hear, this is one for people who want to actually analyse the audio instead of listen. I haven’t done any analysis on this yet.

Would be interesting to see if adding VST plugins or using Rewire affects the results.

Friend of mine has been producing acid techno records with Renoise and dinsync converters for years, also using VST’s combined with TB303’s… It never affected his timing.

I think it feels like Renoise is more CPU efficient and rock steady at the core than other DAWs, maybe that has something to do with it? I can’t see how following a clock should be so hard as long as the resources can handle it? I don’t know crap though, just speculating.

You can connect a Signal Follower to a Hydra and make the Hydra control a track on the left of the current track

I don’t think i have tried this before, wouldn’t it result in awful feedback? I have to try this. :stuck_out_tongue:

Edit: Hmmm, that did not exactly like i suspected, but i got it looping out of control with the flanger at least. :slight_smile:

I can’t really understand how Renoise calculates this?

small bump…

so far no progress in this, I guess I’ll have no other option than to look at the actual midi data.

Hm cannot believe this. I am sure that logic’s and Cubase’s midi clocks are much more stable than this, they are of course proven to be stable! 10 million people do not have this problem. Is it possible that the other daws also send other midi data next to the clock, more than renoise? This would maybe flood the lame midi bandwidth or overcharge lame midi hardware.

Or is it a Yosemite problem? Did you deactive app nap in the app props, for example? Can you try OS X 10.6.8, which is rock stable?

Did you do a proper quantization on the track in the other daws? :slight_smile:

I think it feels like Renoise is more CPU efficient and rock steady at the core than other DAWs, maybe that has something to do with it? I can’t see how following a clock should be so hard as long as the resources can handle it? I don’t know crap though, just speculating.

I don’t think i have tried this before, wouldn’t it result in awful feedback? I have to try this. :stuck_out_tongue:

Edit: Hmmm, that did not exactly like i suspected, but i got it looping out of control with the flanger at least. :slight_smile:

I can’t really understand how Renoise calculates this?

I gotta say, I love Renoise, but since 3.0 I think they lost their efficiency throne to Reaper.

Hm cannot believe this. I am sure that logic’s and Cubase’s midi clocks are much more stable than this, they are of course proven to be stable! 10 million people do not have this problem. Is it possible that the other daws also send other midi data next to the clock, more than renoise? This would maybe flood the lame midi bandwidth or overcharge lame midi hardware.

Or is it a Yosemite problem? Did you deactive app nap in the app props, for example? Can you try OS X 10.6.8, which is rock stable?

Did you do a proper quantization on the track in the other daws? :slight_smile:

quantization? I’m just sending clock??

from looking at the files I posted in #4, I suspect the problem to be enlarged by the midi->dinsync conversion. For some reason my philrees converter likes the Renoise midi much more than other DAW’s midi clock. I debunked my own theory that this is related to the “constant midi clock” sent by Renoise, as my test with Cubase also had the setting to send midi clock on stop checked.

I’ve been having Sync issues for ages (it’s just the first time I actually document them) so this is not Yosemite specific. It’s a fairly clean install, no settings tweaking.

Doubt other DAW’s would send more midi than bandwidth allows, all my recordings were done on empty projects.

Sorry if this is the wrong subforum, but this question has been puzzling me for ages.

Renoise has a very stable midi clock, I would say it is the best DAW for generating MIDI Beat Clock good enough to sync external instruments.

I don’t know why, but it functions almost like a rock solid hardware sequencer. Especially when using a setup with a midi to dinsync converter, Renoise will shine.

I’ve done some very non-scientific tests today which show this immediately.

My setup:

Mac Mini (end 2011) with 2.7 Ghz i7 with 4GB RAM running 10.10.2 Yosemite. Midi out from MOTU 828 mk3 firewire. Split up by Emagic AMT8 (functioning as a midi thru device), Syncing Yamaha RY30 pattern and a phil rees MIDI->Dinsync converter driving a Roland TR606 hihat pattern.

Daw’s tested:

  • Cubase Elements 7 (set to send midi clock on stop, like renoise does)

  • Renoise 3

  • Numerology

  • Logic Pro X

  • Digital Performer 8

  • Reaper V4.76/64

  • Live 9

Results:

https://soundcloud.com/rv0/sets/crappy-sync-test

Identical test, but using the ESI Midimate II instead of the Motu:

https://soundcloud.com/rv0/sets/more-crappy-sync-tests-with-ploytec

As you can see, every single DAW has terrible results compared to Renoise.

Why?

RVooh van synthforum ?

haha you need t otest EnergyXT :smiley: it is the worst of all

simple midi out and one note and it cant sync… terrible

now im glad not im am the idiot! thank you

Hm cannot believe this. I am sure that logic’s and Cubase’s midi clocks are much more stable than this, they are of course proven to be stable! 10 million people do not have this problem.

Check over at Five12’s (Numerology Sequencer) Forum. They will laugh in your face about “how stable Logics Midi Clock” is… Lol… :smiley:

Its utter shite IMO. Can’t speak for Cubase though. Never had to sync up stuff to Cubase in the short time I had to use it.

I’ll probably be told off because of this, but Renoise’s actually got the best ITB filters out there as well. So I guess it’s all just about being really honest to whatever you offer customers. That’s what we/they expect: quality stuff.

All today’s DAWs have good enough midi clocks. It’s the hardware that sucks usually, doesn’t it? What do they use inside, a MC68000, or something (where did they get this)?

Actually I still have a MOTU midi router (timepiece I guess) that use as far as I remember an 68000, and this device works really good…

It’s the bios! Or the JIT-compiler!