Tempo Detection question

Post general topics related to Cubase Pro 8, Cubase Artist 8 and Cubase Elements 8 here.
Post Reply
User avatar
raino
Grand Senior Member
Posts: 4679
Joined: Tue Dec 21, 2010 2:41 am
Location: San Francisco
Contact:

Tempo Detection question

Post by raino »

I have a couple of hand entered midi tracks where the notes are all snapped to the grid. Then I had a part played against those tracks and there is a slight variation in the played part vs. the original midi parts.

I want to have the midi parts match up to the more fluid played part. So I used Tempo Detection on the played part and that worked fine - the grid now aligns with all the slight variations in the playing. However I thought that this would also adjust the timing on the midi tracks, but it didn't. Now midi notes that were initially exactly on the grid are not exactly on the grid. Likewise the midi parts don't begin and end on bar lines where they initially did.

Looking closer I found that the midi tracks, which were initially set to musical mode, were now set to linear mode. It appears that the tempo detection process changed the mode :?: . So it turns out the midi notes are at the same time location as before and their positions relative to the played part hasn't changed. Not my intention at all.

Is there any way I can use Tempo Detection to set the grid to follow the played part AND have the midi data retain its position relative to the musical grid and not the linear timeline?
rodger

Cubase Pro 10.5, Win10 Home, Halion6, Groove Agent 5, WaveLab Pro 10, BFD3, GPO, Komplete Ultimate 9, Waves Gold, Misc. other

i7 4771 Haswell Processor 3.5GHz, ADK LGA1150, 16GB RAM, GeForce GTX970, RME Fireface 400, 2x UAD-2

Latest: Bawdy Politik https://rodgerraino.bandcamp.com/releases

User avatar
GPT
New Member
Posts: 31
Joined: Wed Dec 12, 2012 7:07 pm
Contact:

Re: Tempo Detection question

Post by GPT »

Hi Roger,

Not long ago I posted about the exact issue you're having. Here is my post:

http://www.steinberg.net/forums/viewtop ... 16#p495771

I tried to explain the symptom with my limited English. but it's exactly what you described. The main issue here is that when you start the tempo detection analyse, all the track that are in musical time base are switch to linear mode from the start of the analyses. One would expect for the tracks in musical time based to be left as is; just like it behave if you use the time warp alternate tool (Time Warp(musical events follow)) to adjust the tempo manually.

I also opened a ticket at SB support. The behaviour has been acknowledge and passed at higher up support team. Still waiting for some news.

I can confirm the behaviour on Windows 8.1 with Cubase8.0.30 and with MAC-OSX10.11 with Cubase8.5

I hope this get solved sometime...
Cubase Pro 8.5 - Mac Pro - OSX 10.11 ElCapitan (8Core 32G) - Focusrite 18i20
- PC Laptop - Windows 8.1 64bits (Asus S56CM i5@1.70 - 16G) - Focusrite 2i2

User avatar
raino
Grand Senior Member
Posts: 4679
Joined: Tue Dec 21, 2010 2:41 am
Location: San Francisco
Contact:

Re: Tempo Detection question

Post by raino »

GPT wrote:Hi Roger,

Not long ago I posted about the exact issue you're having. Here is my post:

http://www.steinberg.net/forums/viewtop ... 16#p495771

I tried to explain the symptom with my limited English. but it's exactly what you described. The main issue here is that when you start the tempo detection analyse, all the track that are in musical time base are switch to linear mode from the start of the analyses. One would expect for the tracks in musical time based to be left as is; just like it behave if you use the time warp alternate tool (Time Warp(musical events follow)) to adjust the tempo manually.

I also opened a ticket at SB support. The behaviour has been acknowledge and passed at higher up support team. Still waiting for some news.

I can confirm the behaviour on Windows 8.1 with Cubase8.0.30 and with MAC-OSX10.11 with Cubase8.5

I hope this get solved sometime...
Thanks (sort of). I was hoping I was just missing a setting somewhere. But it sounds like this is how it is designed to behave - weather by intention or carelessness.
rodger

Cubase Pro 10.5, Win10 Home, Halion6, Groove Agent 5, WaveLab Pro 10, BFD3, GPO, Komplete Ultimate 9, Waves Gold, Misc. other

i7 4771 Haswell Processor 3.5GHz, ADK LGA1150, 16GB RAM, GeForce GTX970, RME Fireface 400, 2x UAD-2

Latest: Bawdy Politik https://rodgerraino.bandcamp.com/releases

User avatar
GPT
New Member
Posts: 31
Joined: Wed Dec 12, 2012 7:07 pm
Contact:

Re: Tempo Detection question

Post by GPT »

Hi Roger,

I just had my reply from Steinberg... not the one I was hoping for as I was also hoping for some preference setting to resolve the issue.

As the support explained, Steinberg said that the result I am expecting is not how the tempo detection was designed to work. They recommend I go the way I was with the warping method for this example. The tempo detection was meant to be the first thing done in a new session then you build on from there.

I threw them back the question to know if that could be implemented in the future, but I think we're at a dead end on that one!

Regards,
Guy
Cubase Pro 8.5 - Mac Pro - OSX 10.11 ElCapitan (8Core 32G) - Focusrite 18i20
- PC Laptop - Windows 8.1 64bits (Asus S56CM i5@1.70 - 16G) - Focusrite 2i2

User avatar
raino
Grand Senior Member
Posts: 4679
Joined: Tue Dec 21, 2010 2:41 am
Location: San Francisco
Contact:

Re: Tempo Detection question

Post by raino »

After reading your posts I did figure out a workaround of sorts. I thought that if it is going to turn off musical mode before doing the detection, why not preserve the midi data from before detection.

1. Make sure the existing midi tracks are in musical mode.
2. Make sure there is only one midi part on each track, glue multiple parts together (this is really just for convenience).
3. Drag each midi part onto the desktop where it will create a .mid file.
4. Do the tempo detection.
5. Set the pre-existing midi tracks back to musical mode.
6. Remove the midi parts on those tracks (I just moved them to some empty tracks).
7. Drag the .mid files on your desktop back onto their respective tracks.
8. The midi data on those tracks is now aligned to the grid just like before detection.

While not elegant it is fairly easy to do. Also do it on a project copy just in case.
rodger

Cubase Pro 10.5, Win10 Home, Halion6, Groove Agent 5, WaveLab Pro 10, BFD3, GPO, Komplete Ultimate 9, Waves Gold, Misc. other

i7 4771 Haswell Processor 3.5GHz, ADK LGA1150, 16GB RAM, GeForce GTX970, RME Fireface 400, 2x UAD-2

Latest: Bawdy Politik https://rodgerraino.bandcamp.com/releases

-steve-
External Moderator
Posts: 9670
Joined: Wed Dec 15, 2010 7:42 pm
Contact:

Re: Tempo Detection question

Post by -steve- »

I was thinking the same thing. You might be able to reduce the number of steps by exporting the tracks and importing them back in after you do the tempo detection.
independent manufacturer rep (not a Steinberg employee)
[safe mode] [cubase manual] [score editor manual]

User avatar
GPT
New Member
Posts: 31
Joined: Wed Dec 12, 2012 7:07 pm
Contact:

Re: Tempo Detection question

Post by GPT »

Good thinking here, but I was already doing it. That work flow implies that you re-import after you are finished with TD. My wish is to do it while still in Tempo Detection mode.

If the time base would be respected all through TD, you could see were TD missed and make corrections along the way as the right most tempo re-adjust following a correction. In other words I'd like to hear the Musical mode tracks to bar&beats as an alternative to the stupid 1/4 signature metronome which throws me off every single time while correcting the TD misses.

One way to kind of work around it would be to use your method and then use the timewarp tools to complete manually from where Tempo Detection missed...

Not holding my breath, but the representative on my support ticket sent it as a feature request... whatever that mean!
Hey, still gotta love the tool!
Cubase Pro 8.5 - Mac Pro - OSX 10.11 ElCapitan (8Core 32G) - Focusrite 18i20
- PC Laptop - Windows 8.1 64bits (Asus S56CM i5@1.70 - 16G) - Focusrite 2i2

-steve-
External Moderator
Posts: 9670
Joined: Wed Dec 15, 2010 7:42 pm
Contact:

Re: Tempo Detection question

Post by -steve- »

Feature requests do get taken seriously, and the devs do work on them. Personally I think this would be very useful. It takes a while for feature requests to get through the system, though, so not holding your breath would be a wise course of non-action. :D
independent manufacturer rep (not a Steinberg employee)
[safe mode] [cubase manual] [score editor manual]

User avatar
GPT
New Member
Posts: 31
Joined: Wed Dec 12, 2012 7:07 pm
Contact:

Re: Tempo Detection question

Post by GPT »

I didn't mean that features request were not taken seriously. I think that dev team might be pretty busy for the next few months with the continuous work on current 8.0 trend version and the upcoming bugs and issues that 8.5 will bring to the table(as expected with every new software versions...)

When I say not holding my breath, I just mean that I won't be waiting for that feature to keep using Cubase as I don't think it could get implemented in any time (very)soon. But yes I'll be the first one to be happy when they decide to add this feature! I don't want in anyway to discredit dev team, to me, Cubase is a gem compare to some other DAWs on the market (the ones I tried, anyway).

cheer!
Cubase Pro 8.5 - Mac Pro - OSX 10.11 ElCapitan (8Core 32G) - Focusrite 18i20
- PC Laptop - Windows 8.1 64bits (Asus S56CM i5@1.70 - 16G) - Focusrite 2i2

alexis
Grand Senior Member
Posts: 4417
Joined: Tue Dec 21, 2010 7:55 pm
Contact:

Re: Tempo Detection question

Post by alexis »

raino wrote:I have a couple of hand entered midi tracks where the notes are all snapped to the grid. Then I had a part played against those tracks and there is a slight variation in the played part vs. the original midi parts.

I want to have the midi parts match up to the more fluid played part. So I used Tempo Detection on the played part and that worked fine - the grid now aligns with all the slight variations in the playing. However I thought that this would also adjust the timing on the midi tracks, but it didn't. Now midi notes that were initially exactly on the grid are not exactly on the grid. Likewise the midi parts don't begin and end on bar lines where they initially did.

Looking closer I found that the midi tracks, which were initially set to musical mode, were now set to linear mode. It appears that the tempo detection process changed the mode :?: . So it turns out the midi notes are at the same time location as before and their positions relative to the played part hasn't changed. Not my intention at all.

Is there any way I can use Tempo Detection to set the grid to follow the played part AND have the midi data retain its position relative to the musical grid and not the linear timeline?
Dumb question almost certainly: after the part in bold above, would the desired results be obtained if after putting the tracks back to Musical Time Base (not musical "mode", right?) all the "in the wrong spot" MIDI parts were selected then commanded to return to their point of origin (Shift-O, iirc) ... or does that only work for audio?
Alexis

-Cubase "Safe Start Mode" (CTRL-ALT-SHIFT)
-Get variable-tempo audio to follow a grid here,
-Replacing freely-timed section into a variable tempo project

Cubase 9.0.20; i5-4570 3.2GHz, 16GB RAM; W10 Pro 64-bit on Samsung SSD 840 Pro 256GB; Seagate 1TB SATA 600 Audio; UR28M; Motif8; UAD-2 Solo; Jamstix 3.6; RevoicePro3.3; EZDrummer 2

User avatar
GPT
New Member
Posts: 31
Joined: Wed Dec 12, 2012 7:07 pm
Contact:

Re: Tempo Detection question

Post by GPT »

alexis wrote:
raino wrote:I have a couple of hand entered midi tracks where the notes are all snapped to the grid. Then I had a part played against those tracks and there is a slight variation in the played part vs. the original midi parts.

I want to have the midi parts match up to the more fluid played part. So I used Tempo Detection on the played part and that worked fine - the grid now aligns with all the slight variations in the playing. However I thought that this would also adjust the timing on the midi tracks, but it didn't. Now midi notes that were initially exactly on the grid are not exactly on the grid. Likewise the midi parts don't begin and end on bar lines where they initially did.

Looking closer I found that the midi tracks, which were initially set to musical mode, were now set to linear mode. It appears that the tempo detection process changed the mode :?: . So it turns out the midi notes are at the same time location as before and their positions relative to the played part hasn't changed. Not my intention at all.

Is there any way I can use Tempo Detection to set the grid to follow the played part AND have the midi data retain its position relative to the musical grid and not the linear timeline?
Dumb question almost certainly: after the part in bold above, would the desired results be obtained if after putting the tracks back to Musical Time Base (not musical "mode", right?) all the "in the wrong spot" MIDI parts were selected then commanded to return to their point of origin (Shift-O, iirc) ... or does that only work for audio?
This is not a dumb question to me (because I thought the same thing at first!) but this is not the case as the tempo shifts happen after the Musical mode are switch to linear mode. So even after switching the tracks back to musical after being finished with Tempo Detection, it remains misplaced and now respond to bar&beats, but with the wrong relative positions. I tried both types, MIDI and audio, and I get the same results. I don't know about the "Shift-O" but in my situation, it doesn't re-adjust anything... :(
Cubase Pro 8.5 - Mac Pro - OSX 10.11 ElCapitan (8Core 32G) - Focusrite 18i20
- PC Laptop - Windows 8.1 64bits (Asus S56CM i5@1.70 - 16G) - Focusrite 2i2

User avatar
raino
Grand Senior Member
Posts: 4679
Joined: Tue Dec 21, 2010 2:41 am
Location: San Francisco
Contact:

Re: Tempo Detection question

Post by raino »

SteveInChicago wrote:I was thinking the same thing. You might be able to reduce the number of steps by exporting the tracks and importing them back in after you do the tempo detection.
Duhh, that worked just fine & is much easier :D
rodger

Cubase Pro 10.5, Win10 Home, Halion6, Groove Agent 5, WaveLab Pro 10, BFD3, GPO, Komplete Ultimate 9, Waves Gold, Misc. other

i7 4771 Haswell Processor 3.5GHz, ADK LGA1150, 16GB RAM, GeForce GTX970, RME Fireface 400, 2x UAD-2

Latest: Bawdy Politik https://rodgerraino.bandcamp.com/releases

User avatar
raino
Grand Senior Member
Posts: 4679
Joined: Tue Dec 21, 2010 2:41 am
Location: San Francisco
Contact:

Re: Tempo Detection question

Post by raino »

alexis wrote: Dumb question almost certainly: after the part in bold above, would the desired results be obtained if after putting the tracks back to Musical Time Base (not musical "mode", right?) all the "in the wrong spot" MIDI parts were selected then commanded to return to their point of origin (Shift-O, iirc) ... or does that only work for audio?
Shift+O isn't the command, but I found it in the Key Commands & it has no key attached - it also is audio only.

The behavior that is occurring is actually just what you'd want & expect if the midi tracks were set linear. But if the tracks are set to musical you'd want them to remain at the same grid location. The problem is that Tempo Detection decides all on its own to change the tracks from musical to linear.

So even if the command worked on midi, I'm guessing it would see all the notes as already being at their origin since they are at the same place on the timeline as they originally were.

Interestingly any chords on the Chord Track and any Markers do not maintain their place on the grid, instead staying at their "linear" location (although those tracks can't be set either way). Quantize will move the chords to the right place. But for the Markers you have to hand drag them to snap to the grid.
rodger

Cubase Pro 10.5, Win10 Home, Halion6, Groove Agent 5, WaveLab Pro 10, BFD3, GPO, Komplete Ultimate 9, Waves Gold, Misc. other

i7 4771 Haswell Processor 3.5GHz, ADK LGA1150, 16GB RAM, GeForce GTX970, RME Fireface 400, 2x UAD-2

Latest: Bawdy Politik https://rodgerraino.bandcamp.com/releases

Post Reply

Return to “General”

Who is online

Users browsing this forum: No registered users and 3 guests