Synth drift

A space for learning and studying the Scope environment and music-making in general.
Sormast
Posts: 13
Joined: Sat Dec 14, 2013 2:59 am

Re: Synth drift

Post by Sormast » Tue Dec 17, 2013 5:52 am

Mr Arkadin wrote:Personally I never use Cubase's mixer. I route via ASIO from Cubase into a Scope mixer. You can plug all your Scope synths and effects into the Scope mixer. You can then feed the result back into Cubase for recording. There is no latency this way.

Have you adjusted the ULLI in Scope:Set>Settings> for a low setting?
ULLI - ~22. I have Powercore, Uad. For me only External instruments?
"The sample you uploaded is around 145.7 bpm and has some randomness in the attack of the sound" - and how to be with it? really people sample analog?

Sormast
Posts: 13
Joined: Sat Dec 14, 2013 2:59 am

Re: Synth drift

Post by Sormast » Sat Dec 21, 2013 9:03 am

Sormast wrote:
Mr Arkadin wrote:Personally I never use Cubase's mixer. I route via ASIO from Cubase into a Scope mixer. You can plug all your Scope synths and effects into the Scope mixer. You can then feed the result back into Cubase for recording. There is no latency this way.

Have you adjusted the ULLI in Scope:Set>Settings> for a low setting?
ULLI - ~22. I have Powercore, Uad. For me only External instruments?
"The sample you uploaded is around 145.7 bpm and has some randomness in the attack of the sound" - and how to be with it? really people sample analog?
all thanks. I sold the card

Yogimeister
Posts: 307
Joined: Sat Jan 12, 2008 10:12 am

Re: Synth drift

Post by Yogimeister » Thu Jan 01, 2015 10:53 am

I actually have both these problems ...

(I guess they are more noticeable when working on "high percision/surgical" 145psy ...)

I have been getting pops and volume irregularities in the attack end of the synths since day one (only on some patches/synths ) - i thinkit is more about those synths having no short attack values - so its either click or fade in (im exaggerating)

Secondly, not all synths have oscillator sync and I think this causes ringing/phasing between oscillators on some settings ...




My major concern now (Im doing a psy track and trying to be very surgical about it - which means making sure the samples are phase-aligned ) And I have been noticing that the synths (or midi) DO DRIFT ....

The midi notes are fully quantized in cubase and everything is well - but the re-recorded audio "drifts" - some notes hit before, some after and some dead on (relative to the first note - not the actual midi ) ....
I triple checked the midi parts ....



Midi>system sync is on (always)


I tried ASIO2 and ASIO1 (32) IOs (I didnt connect a clock to the ASIO2 - and I didnt see any clock settings in cubase ...
In the cubase (5.5) device manager - hw pull up/down is off, clock source is set to sync to SC and "externally clocked" is unchecked.

Vst system link is not active. (Will it help to turn it on - not just for sync but also for transfer bits, etc... ?)

User avatar
faxinadu
Posts: 1201
Joined: Wed Nov 01, 2006 3:12 am
Location: israel
Contact:

Re: Synth drift

Post by faxinadu » Sun Feb 21, 2016 2:16 am

- always use scope synths/modular internal bpm and use the textbox/fader to set to your track bpm. never sync them to daw bpm (external) as this is buggy and has been an issue since first scope until today.

- make sure your system likes the ulli you have setup.

- not all synths having osc sync(retrig) is a good thing not a bad thing ;) many sounds you can't achieve with retrigger oscs (for example a supersaw and other detune based stuff).

- regularities of the attack can be an inherent problem of some scope synths if the designer wasn't careful with the min/max values. If taking into account all samplerates, then the minimum attack time in scope is 3.7ms. Any synth that lets you go to values below the ones in the table "lie" and can cause clicks.

Sample Rate Minimum Value Maximum Value
32.0kHz 3.75ms 131.0s
44.1kHz 2.70ms 95.1s
48.0kHz 2.50ms 87.4s
96.0kHz 1.25ms 43.7s


extra thought: so you see guys, beyond subjective "how good does it sound" differences between samplerates, there are also objective real world behavior differences :)

fra77x2
Posts: 155
Joined: Sun May 03, 2015 3:23 pm

Re: Synth drift

Post by fra77x2 » Sun Feb 21, 2016 2:45 am

What are "regularities of the attack" ?

If you put values below it would just stay the same.
Any synth that lets you go to values below the ones in the table "lie" and can cause clicks.
Clicks are regularly created by envelopes. This is not a problem, it is a standard behavior of signals.
Raise the attack a bit to avoid these.
then the minimum attack time in scope is 3.7ms.
This is for the standard envelope atoms, for custom made a 0 sample attack time is possible.

Midi sync is unrelated to midi jitter.
never sync them to daw bpm (external) as this is buggy


It is not buggy. It is susceptible to midi jitter. You can smooth midi clock if you prefer. It works.

User avatar
faxinadu
Posts: 1201
Joined: Wed Nov 01, 2006 3:12 am
Location: israel
Contact:

Re: Synth drift

Post by faxinadu » Sun Feb 21, 2016 2:48 am

how do you smooth midi clock?

User avatar
faxinadu
Posts: 1201
Joined: Wed Nov 01, 2006 3:12 am
Location: israel
Contact:

Re: Synth drift

Post by faxinadu » Sun Feb 21, 2016 2:50 am

mate but you forget when someone like yogi is having problems, he isn't an sdk user, he is making a track with scope devices... info like costum built envelopes with 0 attack and smoothing midi jitter is not really relevant to him :)

he needs to know why he is having problems, and what is the quick solution :)
Last edited by faxinadu on Sun Feb 21, 2016 2:51 am, edited 1 time in total.

fra77x2
Posts: 155
Joined: Sun May 03, 2015 3:23 pm

Re: Synth drift

Post by fra77x2 » Sun Feb 21, 2016 2:51 am

Interpolate the f signal

User avatar
faxinadu
Posts: 1201
Joined: Wed Nov 01, 2006 3:12 am
Location: israel
Contact:

Re: Synth drift

Post by faxinadu » Sun Feb 21, 2016 2:52 am

fra77x2 wrote:Interpolate the f signal

cool i will try, but wondering if it is that simple how come it has never been fixed in modular or any cw/sc synths hmm

fra77x2
Posts: 155
Joined: Sun May 03, 2015 3:23 pm

Re: Synth drift

Post by fra77x2 » Sun Feb 21, 2016 2:52 am

Midi jitter is a general problem with external hardware (like scope). There are several remedies.
If a man knows to identify midi jitter he is a step before finding a solution.

fra77x2
Posts: 155
Joined: Sun May 03, 2015 3:23 pm

Re: Synth drift

Post by fra77x2 » Sun Feb 21, 2016 2:58 am

if it is that simple how come it has never been fixed in modular or any cw/sc synths hmm
If you send a reliable midi clock it works as opposed. This is a problem of midi interfaces, computers, software midi etc

There are a millions ways to synchronize your sounds, noone really cared about the old and problematic midi clock synchron.

Personally i never use midi clock because it "pollutes" the midi channel with data.

User avatar
JoPo
Posts: 1899
Joined: Wed Jun 05, 2002 4:00 pm
Location: VRRAAaaooOôOooommmh
Contact:

Re: Synth drift

Post by JoPo » Sun Feb 21, 2016 3:23 am

Yep ! It sounds like midi jitter...

To be sure it's midi jitter, you can play solo midi track and see if there is the same sync issue.

I use Cubase and when there is too much midi data to read on some parts, I get midi jitter. You could record the most full of midi data tracks in audio and when you'll mute them, normaly, midi sync issues are gone... I must say that midi jitter are always on hardware midi gear : hardware synths or Scope. On VST's, I've never noticed midi sync issue as long as Cubase CPU monitor doesn't go in red.
> > > > > > > > > > > > --- Musica --> here ! ---< < < < < < < < < < < <

fra77x2
Posts: 155
Joined: Sun May 03, 2015 3:23 pm

Re: Synth drift

Post by fra77x2 » Sun Feb 21, 2016 3:48 am

Vst 's are always sample accurate.

By my experience the best midi timing (for external hardware) is offered by Reaper.

For critical sounds, percussive etc midi jitter is unacceptable. For the rest 85% of sounds it is workable.

There are several ways to work against it.

To test for midi jitter simply send a midi track with 64th notes and hear the output.
Be sure to listen to some oscillator that it's phase is synced to gate.

Do the same for a vst or with a hardware sequencer (scope modular seqs are fine)
to get a reference of high quality midi timing.

User avatar
JoPo
Posts: 1899
Joined: Wed Jun 05, 2002 4:00 pm
Location: VRRAAaaooOôOooommmh
Contact:

Re: Synth drift

Post by JoPo » Sun Feb 21, 2016 4:06 am

Reaper ? Ah.. Here, when I tested Reaper for one full month, I had same or even more midi timing issue. Perhaps, it depends of which computer is used.

Anyway, the workaround is easy : record the most complicated midi tracks in audio and say bye-bye to midi jitter...

Also, for controling with more accuracy Scope devices, I use Xite midi input : I assign midi track to my midiman midi output and go in Xite with a hardware midi cable, the timing is (just) a bit better. Or with Copperlan.

But it's impossible to avoid completly midi jitter with Cubase because midi jitter is created inside Cubase. So even with the most faboulos midi interface, you'll still have issue on big cubase project.

Another solution is USAMO box : http://www.expert-sleepers.co.uk/usamo.html I never tried it ... If anybody did, it would be nice to know if it's worth it...

And I know that wellis is working on a ScopeSync VSTi that could help us a bit like USAMO box.
> > > > > > > > > > > > --- Musica --> here ! ---< < < < < < < < < < < <

fra77x2
Posts: 155
Joined: Sun May 03, 2015 3:23 pm

Re: Synth drift

Post by fra77x2 » Sun Feb 21, 2016 4:15 am

Anyway, the workaround is easy : record the most complicated midi tracks in audio and say bye-bye to midi jitter..
Just recording a midi track that suffers from midi jitter does not cure it...
You have to quantize the audio in a way or find a different solution.

By my testing most midi interfaces does not offer acceptable midi timing for the critical elements.
That also applies to cooperlan.
Here, when I tested Reaper for one full month, I had same or even more midi timing issue
This is strange.

User avatar
JoPo
Posts: 1899
Joined: Wed Jun 05, 2002 4:00 pm
Location: VRRAAaaooOôOooommmh
Contact:

Re: Synth drift

Post by JoPo » Sun Feb 21, 2016 4:47 am

Well.. Recording most complicated midi track is enough to avoid midi jitter that I hear..! So it's enough to make me happy.. It's better to not hear midi timing issue than to hear it ! :P

A video from expert sleepers shows usb midi jitter :
https://vimeo.com/30913010
But it's not so bad : a division = 1 ms.. But with their USAMO box or ES-4 & else product, nomore sync issue ! When Cubase has to send a lot of midi data, the jitter is much stronger than in the video ! 1ms is unhearable : human ears won't detect delay under 15-20ms. So if I can hear it, it means they are more than 15-20ms long !

USAMO box = 140€ without power supply.
> > > > > > > > > > > > --- Musica --> here ! ---< < < < < < < < < < < <

User avatar
faxinadu
Posts: 1201
Joined: Wed Nov 01, 2006 3:12 am
Location: israel
Contact:

Re: Synth drift

Post by faxinadu » Sun Feb 21, 2016 6:01 am

why all these workarounds instead of just using internal scope timing? almost every scope device that has a clock can be set to internal mode. problem solved.

fra77x2
Posts: 155
Joined: Sun May 03, 2015 3:23 pm

Re: Synth drift

Post by fra77x2 » Sun Feb 21, 2016 6:18 am

midi jitter and midi clock are (totally) unrelated.

Midi clock is used to give information about the tempo to lfo's, arpeggiators etc

Midi jitter is an error associated with the transmission of every midi message.

If you watch the video Jopo has posted
A video from expert sleepers shows usb midi jitter :
https://vimeo.com/30913010
It displays clearly what we are talking about. The small square displays the drift of the midi message.
When the "expert sleepers" is used check that it stays the same (sample accuracy).

Jopo knows to identify midi jitter and solves this by not overloading the midi channel and tries to minimize midi jitter to acceptable values

User avatar
faxinadu
Posts: 1201
Joined: Wed Nov 01, 2006 3:12 am
Location: israel
Contact:

Re: Synth drift

Post by faxinadu » Sun Feb 21, 2016 6:42 am

ahhhhhhhhhhh ok i gotcha now, i thought you were talking about clock drifts in lfos and such yes.

yah, that's probably what is going on with yogi as well. i had that in the past too thinking about it, my workflow that last years has basically been produce and record channels to audio one by one and not much midi information is going on so yeah that solves it.

User avatar
Bud Weiser
Posts: 1299
Joined: Tue Sep 14, 2010 5:29 am

Re: Synth drift

Post by Bud Weiser » Sun Feb 21, 2016 6:49 am

JoPo wrote: 1ms is unhearable : human ears won't detect delay under 15-20ms.
There´s lot of theory in the ballpark, but experience is all you need.

A musician will adapt to a degree of constant latency as long it is under ~10ms.

Jitter is what you feel more than you´ll hear it and it´s comparable to random non-periodic changing delaytime.
These random changes can be small, large, both and anything in between at any point on time axis.
Unforeseeable without testing.
There is no 15-20 ms, it can be any values depending on the system´s issues.
It´s pretty much recognizable far under 15ms.

MIDI Test 4.6 (32Bit only),- still downloadable http://web.archive.org/web/201002170606 ... l#download

You also have to differentiate between an average human´s ear and trained ears of musicians, producers and engineers.

Bud

Post Reply