Mike, would you be willing to zip a project file and folder and post here, or PM me with it? I would like to test.
Sure, give me a minute and I’ll do so.
Although it shouldn’t matter because simply creating a blank project in Cubase 8 and adding a midi track and trying to record will show the problem.
Sure, it’s simple to reproduce with these steps. I am definitely not a very precise piano player.
#1. Add a midi track to an empty project. (Connected to a vsti or not, doesn’t matter.)
#2. Turn on the metronome and record quarter notes along with it for say 16 bars.
#3. Open up the part in midi list editor to find out how bad the bug is affecting you.
In Cubase 7, the note on times and the note lengths will be completely random (like human playing is), with no discernable pattern.
In Cubase 8, if the bug is in full effect, the note on times will exhibit obvious artificial patterns. Such as you can only hit notes 3 ticks late or 7 ticks early, but nothing in between, every time.
If the Cubase 8 bug is in partial effect, then your note on times may be fine but the note lengths may exhibit the artificial patterns, such as only multiples of 10 ticks being possible.
Is my timing too far off? 
Cool, it looks like the bug may be affecting your note lengths or maybe the bug isn’t affecting you. Can you repeat the same test maybe twice as long and post the screenshot and I’ll point out what I mean?
Thanks for doing that. As you can probably tell, you’ve got the bug (my guess is everyone does and they’re unaware.)
Your note lengths are all multiples of 10, plus or minus one. Almost nothing between 42 and 48, 52 and 58, 62 and 68, etc. That’s definitely not the way you played it.
If you repeat this test in Cubase 7, you’ll find the note lengths don’t suffer from jitter like that, they’re totally random.
An even worse manifestation of the Cubase 8 bug is it will start doing to your note on times the exact same thing that it did to your note lengths above.
Cubase 8 Midi recording is currently unusable if you want it to sound like how you played it.
I see what you are saying about the values of the lengths being in steps of 10 ticks, but I have not seen the same thing for start times. Also I can confirm that in C7.5 it is different.
I guess this would show that when 120 ticks= a 16th note, Cubase 8 is only capable of recording lengths in steps of something between 128th and 256th notes.
Yeah it will randomly start affecting the start times too.
128th note steps is a very coarse resolution, particularly when it’s affecting note-on times. The difference in human feel between Cubase 7.5 and 8 is very audible.
I think the best way to report this is with screen grabs as I have been doing. If you and a couple others can post some images showing this affecting start times it might be helpful.
I sent project files a couple months ago to Steinberg with affected start times, so they’re certainly aware of it and don’t know how to fix it.
But I suppose your suggestion could be useful in the sense that it might make more users aware that they’re using buggy broken software, which might cause Steinberg to care more. So I’ll do it in a bit.
I sent project files a couple months ago to Steinberg with affected start times, so they’re certainly aware of it and don’t know how to fix it
That is an uninformed assumption.
I’m not sure what the cause is behind this, and really, using the Issue Reports forum and guidelines is the way to. That forum is viewed by SB staff and the reports are tested.
I posted it in Issue Reports too, with the same zero response as over phone and email.
When I posted it here, they combined my new thread with this old one and buried it, leading me to conclude their current strategy is to hide this bug unless and until they figure out how to fix it.
Link to your Issue Report please?
They moved/hid it into the “Miscellaneous” Issue Reports:
It was moved because it didn’t have a step by step repro. Posts that follow the guidelines do not get moved.
Anyway, You posted a hard-to-describe/understand problem, and I (could have been any interested user ) took interest and the time to figure out what you were saying and I tested the problem and I provided concrete repro info. Add to this screen grabs from a couple users (no project really needed) and a nice simple step by step recipe and we will have a wonderful report. Mainly, stay calm.
Well, thanks for your help. I’ll post some screen grabs too…
Wow this is amazing to me!
Have you tried raising your ppqn?
all daws have issues with latency and jitter by the way. Just a matter of how much.
Also I agree that a lot of people use at least some percentage of quantize, or on the other hand they are fine with lack of exact reproduction. I checked my system and cubase 7.5 is pretty accurate with midi recording but a few other daws are more accurate, I’m sure some settings tweaking would help but like I said it does not effect my ability to produce as it’s not a musically relevant amount still ‘‘feels’’ good…
Am reluctant to move to 8 because of this though. I’ll have to demo it sometime soon to see if this is an issue for me.
Ok, here are screen grabs demonstrating the bug.
First, here’s Cubase 7.5, which has no problems. You can tell because all the data is completely random, the way I played it:
And here’s Cubase 8, a recording which captures both versions of the bug - it switches at bar 9. Prior to bar 9, the bug is only affecting the note lengths - jitter rounds them to every 10 ticks or so.
This behavior - only messing up the note lengths - is the way Cubase 8 behaved me for the first 4 months or so.
Then one day - without changing any settings - the bug migrated to included note-on times. That is how it behaves almost all the time for me now, including in a fresh new project.
The below example was a project I started when Cubase 8 was only screwing up note lengths. I’m now able to record notes with normal start-on times on it, but only up to bar 9.
At bar 9, you can see the bug starts screwing up the note-on times - almost all the values are 103 or 113/112 or 2/3, with a 94 and a 12 and 22 thrown in.
There is a slight tempo change at bar 9 of that project, which could have something to do with it. However, there’s also a slight tempo change at bar 5, and that doesn’t affect anything. Also, fresh new projects without any tempo changes exhibit the full bug behavior for me including note starts. The randomness of whether this bug will affect note start times or just note lengths is further proof of how unstable this bug behavior is.
Lastly, here’s the exact same recording from image 2, with a PPQN resolution of 1000. Basically zooming in on the problem.
Now please your system info OS etc. Include midi and audio interface and anything else hooked up to you computer. Also your motherboard.