what are the specifics of the windows timing errors?
Archive · 200710 old posts
From the original SynthMaker forum (2004–2014), with its members’ names and dates.
gregh ·
From digging around i find that timing will be much better in version 2 but at the moment is subject to windows timing errors. Is there more detail on the nature or distribution of these errors?
For example, is the occurence of any particular event plus or minus about 15ms? Does this mean that four sequential events planned at 100ms intervals might occur at intervals of 85:115:85:115 or some such. Or could they even be 85:85:85:85
Is that how the error plays out?
I'd really appreciate some more detail on this
thanks, gregh
doug4350 ·
Hi Gregh, ( & welcome to the forum )
Basically, it boils down to the timer components rely on the Windows timer, so SM is at the mercy of Windows in that regard. The Windows timer is on a low priority process/thread, so it's not accurate. It's impossible to say how inaccurate, as it depends on how busy the system is. The more busy, the more inaccurate.
For example, is the occurence of any particular event plus or minus about 15ms? Does this mean that four sequential events planned at 100ms intervals might occur at intervals of 85:115:85:115 or some such. Or could they even be 85:85:85:85
It's unlikely that an event will be triggered less than the timer duration, sort of. From your example, you might get a pulse at 115 ms, and the next at 85 to keep linear time in track, but you wouldn't get an 85 ms pulse and then a 115 ms pulse -if- the 85 ms pulse is on the beat. I hope that makes sense . In some (not so) extreme cases, pulses could even be missed, as in if the system is busy for > 200 ms.
Aloha,
Doug
gregh ·
Thanks doug4350 - that's really disappointing as i was hoping to use synthmaker mainly for making algorithmic control devices rather than as a sound generator.
I can see how to use the current system if the timing was accurate. However, is there another method that will allow for accurate step sequencing or more complex timing of events?
I don't want to wait for V2
And thankyou for the welcome to the forum,
Gregh
GaryG ·
I think Dougs covered it.
From my experience, I built a little 16step sequencer and it generally runs ok but it will pause/stutter occasionally (when other tasks are grabbing more of windows time). ie. sounds like someone with bad timing struggling with a piece (a bit like me )
.g
gregh ·
thanks GaryG - unfortunately I'm very particular when it comes to randomness
sambean ·
It is possible to lock to the host timing using stream connectors, and generate events using stream triggers. The disadvantage of this is that more cpu will be consumed, and it can be a bit of a challenge to get your trigger events to activate sound, but it's not impossible.
Here's a sound mangler that I wrote that locks to the hosts timing:
Itchy?
SynthmakerX, which has sample accurate timing is under development and is available for registered users to test upon request (I believe!) It's limited at the moment but it's currently possible to get some basic event triggered stuff going (There's an example sequencer with synth), I'm sure Jorge would be happy to have more folk trying it out and assisting with it development.
Sam
gregh ·
thanks sambean - that's very positive. One more question (for which I'll start another thread)
Dewdman42 ·
I'm extremely interested in this thread, as this is one of the main reasons I am interested in synthmaker also. I am surprised at two things that i am reading from this thread, but maybe I am misunderstanding the capabilities of SM, so please correct me if I'm wrong.
1- Synthmaker uses the windows timer for producing midi events. The windows timer is well known to be not so great. I am wondering if Outsim has looked into or considered using the timer built into DirectMusic which is possibly more accurate?
2 - Am I to understand that there is not an automatic means to sync an SM plugin to the host? The method described earlier in this thread sounds complicated and CPU hungry.
trip_out ·
Sorry to dig this one up (thought it better than starting a new one on the same thing), but what is the timescale for getting this sample accurate timing?
Whilst I'm here I don't know if this is possible, but could we have a list of any upcoming plans, priorities, timescales, etc? Given that v1 has only (fairly) recently been released what are the plans for this version (not v2)? This would be good to have with all the upcoming interest from CM we will have.
trip_out ·
Jorge, Malc, anyone?
Wrote something here? To have a post removed or your name hidden, write to [email protected].