123 postsPage 6 of 7
1 ... 3, 4, 5, 6, 7
Yeah - for me "it is broken" is the REASON it bothers me. I'd prefer not to have to think about it being a contributing factor. Of course, as I stated above, the amount of brokenness is (for me, to a large degree) somewhat irrelevant.

But I still don't want to have to think about it or test it and instead just go back to knowing that it really is my playing that is at fault and not some technical problem.

Drummer and Singer mostly with some limited Guitar/Bass/Keys and a love for tech.
My personal website, with some useful videos and other music related things: https://blades.technology

Using:
Self-built i7 8700k/16gb/all SSD, S1 4 Pro, +Sonar Platinum/Cakewalk by Bandlab, PreSonus Studio 18|24 USB, Faderport 8, Tranzport, Edirol PCR-800, Pearl Mimic Pro and Roland V-Drums kit & Console 1 with SSL4000e + British Class A..
User avatar
by niles on Thu Jul 16, 2020 12:27 am
admiralbumblebee wrote
robertgray3 wroteYour video said "broken". Is ~1.6ms of jitter on Windows "broken", or is it safe to say that we're basically only talking about macOS at this point?


I would absolutely consider that broken assuming that the current testing is correct.
From @robertgray3's video I've measured the distance between the audio (constant) and monitored signal and what we see in the scope corresponds to the measurement.

So I'm inclined to say we have narrowed it down to macOS. To make sure the Notes are at their correct place you could simultaneously record the notes and audio and measure the relation between the note and the constant to determine if the notes land (approximately) where they should.

To rule out macOS as the initiator it would be good to do the exact same tests (with audio this time) in e.g. Ableton Live and Reaper.

If the discussion becomes 'we should have sample accurate monitoring', I'm out. I simply don't expect my program to beat physical laws of time and as a MIDI user, in the worst case, I tolerate 1ms jitter alone for pumping MIDI around.

admiralbumblebee wroteI did encounter something strange though: With Maximum DP, LL on and buffer sizes below 2048 (1024 and lower), the scope did not work for track A. Recording the output gave some strange results where the first and second note were close together and the rest were evenly spaced.
Meters, scopes, analyzers etc. don't work within a DP enabled path simply because the monitored instrument you are hearing then is a duplicate residing in it's own time critical path.

I'm not a fan of Studio One's DP and LLM system and never use it, so I have no further ideas about that.

OS: Windows 10 Pro | HW: P9X79 • i7 3930K • 16GB • 3x EVO 860 • NVIDIA GT1030 (@WQHD) • RME AIO

Vote ❤ for Retrospective Record / Capture Recording in Studio One
User avatar
by robertgray3 on Thu Jul 16, 2020 12:54 am
niles wroteTo rule out macOS as the initiator it would be good to do the exact same test (with audio this time) in e.g. Ableton Live.


So basically do the original test you did where the stereoscope is used? Ok.

Just an aside, this wouldn’t be the first part of the program to work differently on Mac than a Windows- there are some differences that aren’t bugs, such as some of the graphical differences between Mac and PC.

niles wroteIf the discussion becomes 'we should have sample accurate monitoring', I'm out. I simply don't expect my program to beat physical laws of time and as a MIDI user, in the worst case, I tolerate 1ms jitter alone for pumping MIDI around.


Ditto :thumbup:
Last edited by robertgray3 on Thu Jul 16, 2020 1:02 am, edited 2 times in total.

Mac OS X Mojave 10.14.6
Mac Pro (Late 2013)
3 GHz 8-Core Intel Xeon E5
32 GB 1066 MHz DDR3
Dual AMD FirePro D500 3072 MB
Quantum 2
User avatar
by niles on Thu Jul 16, 2020 12:56 am
robertgray3 wrote
niles wroteTo rule out macOS as the initiator it would be good to do the exact same test (with audio this time) in e.g. Ableton Live.


So basically do the original test you did where the stereoscope is used? Ok.
Yes, I don't know if Live has such a stereoscope, most important is sound. If we have sound we can measure.

OS: Windows 10 Pro | HW: P9X79 • i7 3930K • 16GB • 3x EVO 860 • NVIDIA GT1030 (@WQHD) • RME AIO

Vote ❤ for Retrospective Record / Capture Recording in Studio One
User avatar
by robertgray3 on Thu Jul 16, 2020 1:03 am
niles wrote
robertgray3 wrote
niles wroteTo rule out macOS as the initiator it would be good to do the exact same test (with audio this time) in e.g. Ableton Live.


So basically do the original test you did where the stereoscope is used? Ok.
Yes, I don't know if Live has such a stereoscope, most important is sound. If we have sound we can measure.


Wouldn’t we also need a comparison of Studio One and Live on Windows to compare this too, though?

Mac OS X Mojave 10.14.6
Mac Pro (Late 2013)
3 GHz 8-Core Intel Xeon E5
32 GB 1066 MHz DDR3
Dual AMD FirePro D500 3072 MB
Quantum 2
User avatar
by niles on Thu Jul 16, 2020 1:37 am
robertgray3 wroteWouldn’t we also need a comparison of Studio One and Live on Windows to compare this too, though?
Sure, I can do the test too, but I only have the Lite version of Live so I don't have any way to visualize it like we did in Studio One with scope. Of course I can measure the sound. I can set up an Ableton test song later today if you like.

OS: Windows 10 Pro | HW: P9X79 • i7 3930K • 16GB • 3x EVO 860 • NVIDIA GT1030 (@WQHD) • RME AIO

Vote ❤ for Retrospective Record / Capture Recording in Studio One
User avatar
by robertgray3 on Thu Jul 16, 2020 2:12 am
niles wrote
robertgray3 wroteWouldn’t we also need a comparison of Studio One and Live on Windows to compare this too, though?
Sure, I can do the test too, but I only have the Lite version of Live so I don't have any way to visualize it like we did in Studio One with scope. Of course I can measure the sound. I can set up an Ableton test song later today if you like.


No rush, you’ve done plenty!

Mac OS X Mojave 10.14.6
Mac Pro (Late 2013)
3 GHz 8-Core Intel Xeon E5
32 GB 1066 MHz DDR3
Dual AMD FirePro D500 3072 MB
Quantum 2
User avatar
by niles on Thu Jul 16, 2020 5:18 am
I've ran the same test with Live. Because I have to measure each trigger by hand I only lift out 128 (because that's what I use) and 2048 (because people seem to use that :?).

On my Windows system the average deviation between the constant and the monitored signal for 8 triggers at 128 samples / 44.1kHz is ~1.3ms. The overall latency lays between 7 and 10 ms, so the maximum deviation is ~3ms (just like Studio One).

The average deviation between the constant and the monitored signal for 8 triggers at 2046 samples / 44.1kHz is negligible (0.4ms). The overall latency is 95ms and an ocasional 96ms, so the maximum deviation is ~1ms (1ms less than what I measured with Studio One).

Here is the video, you can do additional measurements with the audio of it (for instance run it through the Scope).
phpBB [video]


So to summarize, I don't see a big difference here between live and Studio One. The differences are really marginal (within 1 ms) on Windows and certainly not something to worry about when you understand what and what not to expect from the technology driving it.

To show Windows users what this actually means in context I've triggered both sequencer with the MIDI from the other. The left audio signal is Studio One and the right audio signal is Live.
So what you are hearing here is the difference between Studio One and Live with the same MIDI triggers.

When I measure 128 and 2048 again:
At 128/44.1kHz, Live's average latency is ~2ms higher than Studio One.
At 2048/44.1kHz Studio One's average latency is ~1ms higher than Live.

phpBB [video]


I've attached a .zip with the Studio One file and Live file.

Attachments
Easy_Buffer_Jitter_Test_3_SO_and_Live.zip
(71.28 KiB) Downloaded 8 times

OS: Windows 10 Pro | HW: P9X79 • i7 3930K • 16GB • 3x EVO 860 • NVIDIA GT1030 (@WQHD) • RME AIO

Vote ❤ for Retrospective Record / Capture Recording in Studio One
User avatar
by admiralbumblebee on Thu Jul 16, 2020 9:09 am
Live on macOS:

The provided Live test project gives an error for me "File .wav not found" when I try to arm anything.

I recreated the project, and have supplied a video of the project below:

  • Quarter note MIDI notes
  • Loopback to sampler with settings modified to not mess with the sample
  • Record output of sampler track

At all buffer settings in Live, every hit except the first is sample perfect. Studio One also does this "first note is wrong" behaviour in a few different setups, notably with loop on. (likely to avoid the 'first note doesn't play' behaviour that plagues DAWs in a number of circumstances)

I'll check windows with the same project (attached).

phpBB [video]

Attachments
EATER Project.zip
(10.61 MiB) Downloaded 7 times

User avatar
by niles on Thu Jul 16, 2020 9:17 am
admiralbumblebee wroteI'll check windows with the same project (attached).

phpBB [video]

Is there sound?

admiralbumblebee wroteLive on macOS:

The provided Live test project gives an error for me "File .wav not found" when I try to arm anything.
It's in the \Media folder in the .zip and below.

Attachments
metronome.wav
(172.31 KiB) Downloaded 8 times

OS: Windows 10 Pro | HW: P9X79 • i7 3930K • 16GB • 3x EVO 860 • NVIDIA GT1030 (@WQHD) • RME AIO

Vote ❤ for Retrospective Record / Capture Recording in Studio One
User avatar
by admiralbumblebee on Thu Jul 16, 2020 10:01 am
niles wrote
admiralbumblebee wroteI'll check windows with the same project (attached).

phpBB [video]

Is there sound?


No, sound isn't necessary to demonstrate. Utilizing the capture drivers I have available complicates the issue as described previous. This is using the same audio device I've used in other tests, with the same settings.

I've re-run these with 2 other devices (System audio and Komplete Audio 6) each time to double check the results on macOS. They're not shown to avoid complicating the discussion as the results are identical.

I can do another video with a correlation meter, but it shows the exact same results. Correlation is identical on all but the first note.

niles wrote
admiralbumblebee wroteLive on macOS:

The provided Live test project gives an error for me "File .wav not found" when I try to arm anything.
It's in the \Media folder in the .zip and below.


Yes, it sees the metronome file and will play it back as audio and the sampler.

Live gives a generic "File ".wav" could not be opened" for arming any track when using that project. Appears to be a bug with Live. I've attached a video of the issue, but I already fiddled with it enough.

Attachments

[ Play Quicktime file ] NotFound.mp4 [ 370.72 KiB | Viewed 738 times ]


User avatar
by patricemazmanian on Thu Jul 16, 2020 10:10 am
niles wroteI've ran the same test with Live. Because I have to measure each trigger by hand I only lift out 128 (because that's what I use) and 2048 (because people seem to use that :?).

On my Windows system the average deviation between the constant and the monitored signal for 8 triggers at 128 samples / 44.1kHz is ~1.3ms. The overall latency lays between 7 and 10 ms, so the maximum deviation is ~3ms (just like Studio One).

The average deviation between the constant and the monitored signal for 8 triggers at 2046 samples / 44.1kHz is negligible (0.4ms). The overall latency is 95ms and an ocasional 96ms, so the maximum deviation is ~1ms (1ms less than what I measured with Studio One).

I don't understand all this enthusiasm on this subject!
To play on my midi keyboards most of the time, my buffer size is on 64 samples. If my song is heavy, it is 128 samples (with the "Hollywood Orchestra" library)
I would be unable to play a correct note with a buffer at 512, 1024 or other ...

Summary:
If I understood correctly, with a small buffer size, the jitter in studio one is imperceptible.
If the buffer size is large (which I never use outside of mixing), the jitter is more important.
I do not see where the problem is...
cordially
Patrice

S1 pro 5 - Melodyne - Notion 6 - Win10 - jBridge 1.65 - Motherboard ASUS Z87-K - i7 4770K 3.5Ghz, 32 Go RAM - GeForce GT 610 1024MB - Audio interface RME Hammerfall multiface II - Faderport - ADAT Interface Focusrite Scarlett OctoPre - MIDI Interface Motu MIDI Express 128

http://mazmusic.free.fr/
User avatar
by robertgray3 on Thu Jul 16, 2020 12:35 pm
niles wroteIs there sound?.


@niles Thanks for setting that up. Here's my video results. With audio. I wasn't sure what scope plugin would be the most useful. The first two I tried didn't play nice with Ableton (so much for greener grass and all) so I just used SPAN.

@admiral on Mac I just use Rogue Amoeba Loopback + Quicktime to record. Works fine, has sound, and doesn't glitch for me at 44.1 with my external wordclock.
.



patricemazmanian wroteIf I understood correctly, with a small buffer size, the jitter in studio one is imperceptible.
If the buffer size is large (which I never use outside of mixing), the jitter is more important.
I do not see where the problem is...


@patrice Agreed, but I said that on the last page and was called "dismissive" "insulting" and "insufferable." I keep testing to be a good sport but we should be realistic about this.

Mac OS X Mojave 10.14.6
Mac Pro (Late 2013)
3 GHz 8-Core Intel Xeon E5
32 GB 1066 MHz DDR3
Dual AMD FirePro D500 3072 MB
Quantum 2
User avatar
by niles on Thu Jul 16, 2020 2:21 pm
admiralbumblebee wroteNo, sound isn't necessary to demonstrate.
It is if you want to take your findings to the developers. Don't expect them to start investigating without a solid straightforward report. It's a digital audio software developer, the least we should do is give them digital audio examples of the issue.

robertgray3 wrote@niles Thanks for setting that up. Here's my video results. With audio. I wasn't sure what scope plugin would be the most useful. The first two I tried didn't play nice with Ableton (so much for greener grass and all) so I just used SPAN.
Perfect. When I measure the audio from this video the latency in Live at 128/44.1kHz is 11ms, nearly dead on.
When I measure the audio from the video of Studio One on macOS the latency for 128/44.1kHz is ~14ms and there's a deviations of max 2ms (latency varies from 13ms to 15ms).

At 2048/44.1kHz in macOS the latency in Live is 98ms dead on, zero deviation. In Studio One the average latency at 2048/44.1kHz is 119ms. There's a deviation of max 39ms (latency varies from 100ms to 139ms).

I think these figures speak for themselves and they are definitely different from Windows.
If we can rule out the MIDI we have narrowed it down. I think this can simply be done by recording the MIDI and audio all together, save it and pack it in a zip for further analyzes.

I would use a buffer size of 512 and sample rate of 44.1kHz, it's big enough to show the issue an not as excessive as 2048. Don't use input quantize.

Like this:

phpBB [video]


With this song:
Easy_Buffer_Jitter_Test_5.zip
(39.29 KiB) Downloaded 8 times

Edit: Added new attachment (5) due to an error I made in test song.
Last edited by niles on Fri Jul 17, 2020 7:04 am, edited 4 times in total.

OS: Windows 10 Pro | HW: P9X79 • i7 3930K • 16GB • 3x EVO 860 • NVIDIA GT1030 (@WQHD) • RME AIO

Vote ❤ for Retrospective Record / Capture Recording in Studio One
User avatar
by robertgray3 on Thu Jul 16, 2020 3:58 pm
OK packed it in a ZIP and attached it.

niles wrote
admiralbumblebee wroteNo, sound isn't necessary to demonstrate.
It is if you want to take your findings to the developers


Yup! Like I said in the other thread on this...

robertgray3 wroteif you’re not coming up with a short easily repeatable test ... then someone else, usually another user, has to do it for you.


On that note, here's my report for niles' short test :XD:

robertgray3 wroteIf we can rule out the MIDI we have narrowed it down. I think this can simply be done by recording the MIDI and audio all together, save it and pack it in a zip for further analyzes.

I would use a buffer size of 512 and sample rate of 44.1kHz, it's big enough to show the issue an not as excessive as 2048. Don't use input quantize.


@niles Here's one last video of it and a zip file with everything. DM me if you need anything else.

Also, kinda interesting that the latency (+jitter) seems to "settle in" after a while of monitoring, no?




Mac OS X Mojave 10.14.6
Mac Pro (Late 2013)
3 GHz 8-Core Intel Xeon E5
32 GB 1066 MHz DDR3
Dual AMD FirePro D500 3072 MB
Quantum 2
User avatar
by niles on Fri Jul 17, 2020 1:12 am
robertgray3 wroteAlso, kinda interesting that the latency (+jitter) seems to "settle in" after a while of monitoring, no?
Good find. Unfortunately it was an error from my side. I trigger the scope with a MIDI trigger (hidden Scope Trigger track in song), and totally forgot to copy that through. Sorry!

FWIW the audio still had deviations although the scope looked steady.

So due to my error the last: S1 - macOS - 512 Jitter Issue.mov is not representative.

I've updated the test file to version 5 viewtopic.php?p=238931#p238931
If you are able to make the same movie again I can do the same for Windows and we can clearly show the difference side-by-side. If possible you can show the deviations extremes in the scope by clicking once on the lowest latency and move to the highest (approximately). It may give the developers a better idea of what they are looking at.

Image

Attachments
scope_meassure_approximate_latency.gif

OS: Windows 10 Pro | HW: P9X79 • i7 3930K • 16GB • 3x EVO 860 • NVIDIA GT1030 (@WQHD) • RME AIO

Vote ❤ for Retrospective Record / Capture Recording in Studio One
User avatar
by roland1 on Fri Jul 17, 2020 11:38 am
Blades wroteYeah - for me "it is broken" is the REASON it bothers me. I'd prefer not to have to think about it being a contributing factor. Of course, as I stated above, the amount of brokenness is (for me, to a large degree) somewhat irrelevant.

But I still don't want to have to think about it or test it and instead just go back to knowing that it really is my playing that is at fault and not some technical problem.


This is my issue as well. I would much rather be working to improve my timing/meter than worrying about whether improving my own performance will ultimately be meaningless in the end because of some unacknowledged software glitch or recording procedure handicap in S1. :(

I used to have confidence in my tape gear.

I use: Studio One Pro v5 on i7 7700 win10 PC w16GB RAM and 4.61 on a Mac Pro Tower (Sierra) w/RME & Focusrite interfaces.

Author: Reality Checks for Everyday Life — Keeping it real in a world of illusions. I can also be found loitering at: http://RolandK.ca

Attention all shoppers: The people in commercials are paid actors.
User avatar
by roland1 on Fri Jul 17, 2020 12:05 pm
Blades wroteOk - so I'll chime in my input here. I've read through (most of) the thread and I understand where both "camps" are coming from:

....


Thanks for your input, Blades. I appreciate it because I began this thread as a result of my curiosity/confusion about the timing of MIDI data being recorded from my V-Drums into S1.

Ultimately, I find that I have to do quite a bit of tweaking anyway, not only to adjust the velocities and timing, but also to get the overall effect that I want from the drum part.

I think it would just be easier to start the track with a pristine reproduction rather than a fairly close interpretation of what was played.

I use: Studio One Pro v5 on i7 7700 win10 PC w16GB RAM and 4.61 on a Mac Pro Tower (Sierra) w/RME & Focusrite interfaces.

Author: Reality Checks for Everyday Life — Keeping it real in a world of illusions. I can also be found loitering at: http://RolandK.ca

Attention all shoppers: The people in commercials are paid actors.
User avatar
by robertgray3 on Fri Jul 17, 2020 5:34 pm
roland1 wroteI think it would just be easier to start the track with a pristine reproduction rather than a fairly close interpretation of what was played.


Seeing as how it is much simpler and easier to measure, can you perform Niles’ test the same way the rest of us did? It only takes a couple minutes.

It would be good to see a current, real measurement from you as well as the other users who have the most trouble with this.

Mac OS X Mojave 10.14.6
Mac Pro (Late 2013)
3 GHz 8-Core Intel Xeon E5
32 GB 1066 MHz DDR3
Dual AMD FirePro D500 3072 MB
Quantum 2
User avatar
by kassra on Fri Jul 17, 2020 6:29 pm
so just wondering, if working on a mac, does the jitter show up when transforming (or freezing) midi tracks/virtual instruments to audio? does it matter what your buffer setting is when you freeze to audio? if i freeze at 2048 (as opposed to 32 or 64), will my resulting audio track suffer from this timing problem?

OS X 10.13.6, Mac Pro (Late 2013), 3.5 HHz 6-Core Intel Xeon E5, 64 GB RAM, Apollo Quad

123 postsPage 6 of 7
1 ... 3, 4, 5, 6, 7

Who is online

Users browsing this forum: No registered users and 16 guests