Hi Philip, I'm not sure. I'm pretty sure though that qtractor does use alsa midi (it's also mentioned on the front page- search for "alsa sequencer") and it appears as that in, say, QjackCtl
This is the reason why I started the discussion on the mailing list: unsure if this was Rosegarden or Pipewire. However I will say that in Qtractor renaming a 'MIDI bus' (Qtractor Alsa midi in/out) as well as loading a file with such 'buses' named the names are correctly picked up by the Pipewire MIDI bridge as well. I'm not sure what other application with similar features (arbitrary Midi in/out name and renaming) I could test with
Thanks for the merge, tested and it seems to work. I find it quite nicer for projects with around 10-12 but even 15 tracks
Allow to hide (remove) all controllers from instrument panel
Merge request #144 - obviously after 24.12 release
Track name label length feature #525
Tooltip for track width setting in Settings > Presentation
set the track label length (just the text, i.e. after the mute, record and solo buttons) based on the stting in Settings > Presentation
Settings (configuration): add track name length dropdown option in Presentation tab
Ok, how would one assess what are 'easier to find' plugins? I'm on Manjaro (Arch derivative) and would go for plugins which are only in the official repos. Is there any source of information to see availability across distributions? Agreed that maybe LV2 issues should now live in specific bug reports, this 'feature request' is now 9 pages long on sourceforge :-)
I quickly (am a bit full at the moment) tested sfizz and liquid loading a few sfz and playing and they seem to work. I only got one error when it says processor power is not enough but that happens also in some situations without LV2. With LV2 instruments there's still to look into one possible issue (next version I'd say), which is gain... It seems that the gain of instruments is always 'higher' than usual, especially if you chain other effects (e.g. reverb). From just the audio it seems as if there's...
Works for me as well. Thanks!
I'll just quickly report them here as I think the light theme pattern is similar. So, again, just in the light theme In both Matrix and Notation Editor all of the: - Controllers > Insert Expression Controller Sequence... - Controllers > Insert Pitch Bend Sequence... - Controllers > Insert Controller Sequence... (I think these are probably all the same dialog or the same template...) See an example screenshot. QT 5.15.15 (reported by Rosegarden) KDE Plasma: 6.1.5 QT Version (reported in system info):...
I've tested with the latest (just re-built) git and it seems to be fixed, in particular in the light theme (which is the one I use) for those specific dialog. I'm afraid that a similar issue still persists in other dialogs (which I probably hadn't reported here), but will maybe collect in a different bug report? Or is it better to add them here?
Italian translation for 24.12
some additional unfinished translations
Additional translations and some corrections
Italian translation as of today
Allow configurable track label width
Thanks Ted!
If it makes sense to add this see merge request https://sourceforge.net/p/rosegarden/git/merge-requests/141/ Link to screenshot
If it makes sense to add this see merge request https://sourceforge.net/p/rosegarden/git/merge-requests/141/
Added an extra large track size with 194 size factor. Also added a tooltip for the setting
Have an extra large setting for track size
I have tested with the latest git and this seems to be fixed now. Thanks! To 'stress test' a bit I inserted notes and weirdly placed clefs as well as with quarter, 1/8 and 1/16 notes, including overlapping.
Agreed: new (untitled) document: last opened/saved directory document with a name and directory: start from that document's directory Also OK to have an option to maintain the existing behaviour
I know... and quoting myself from that thread: "Of course a humanizer/randomiser could be part of a wider 'groove quantize' feature for Rosegarden, but I imagine that would be rather complicated." "Randomization, 'humanisation', groove quantize etc. (the subject of the thread) would be cool and are IMHO part of those possible improvements ;)" ;-)
Simple 'humanise'
This is a bit old but I'm finding myself wanting to agree and second the feature. The interpretation, especially the auto placement of accents based on time signature (which in turn automatically modifies velocities) is really powerful IMHO and I'm finding myself using it quite a lot, so yes being able to apply it to multiple segments would be really useful
Save As directory should be the same one of the file being saved
Notation Editor, clef causes cursour position to be inaccurate
Canont recover autosaved 'unsaved' file
Hi @tedfelix, I can't seem to remove all MIDI track (and therefore devices). When I try to remove a device I get an error that it's being used by n tracks and therefore cannot be removed; but then there seems to have to be at least one track. I did try to save the autoload.rg file with all connections removed. But when I do new 'something' really wants to autoconnect to whatever is there (I've tried with e.g. Yoshimi, Fluidsynth). I'm attaching my current autoload.rg
Optionally have the timecode part of the transport window embedded in the main window
The preferences flag was it... I didn't know about it until now :-D
I don't seem to have a LV2_PATH but lv2ls works and I also tested a re-scan in Carla. Allowing debug I don't see anything about LV2 but just mention of DSSI and laspa..
Yes those are all in the output of cmake .. (from the ./build directory)
no error... but I only see dssi plugins?
Hi all, I'd like to maybe re-test this if you think it can be useful, but I can't seem to build with LV2 support any more? Is this still in the lman/rosegarden repository with the feature-462-lv2 branch?
It would still be nice to 'predict' what Rosegarden does in terms of auto-connection because often the connection comes with firing CC events etc. and could disrupt something like a soft synth's settings
I guess I can, thanks for looking into this. I'll investigate the Yoshimi side of it.
Hi Philip, thanks for testing this, and indeed you are right! I would have never found that setting in Yoshimi to be honest (and had never used it), but it seemed to be the culprit. I did set Yoshimi's setting to "External" and now when opening a rosegarden file which had a yoshimi connection it inded seems to work. Rosegarden (or Yoshimi??) still seem to 'auto-connect' the "General MIDI" output to when starting Rosegarden. I even tried to explicitly save the autoload.rg file (in $HOME/.local/share/rosegarden/autoload...
ALSA Midi autoconnect is totally unpredictable
I've tried the latest git but the issue still seems to be present, and yes just with the Native (light) theme - I overlooked that in my report as I've never used the other themes.
A maybe more 'user-oriented' POV yes, eventually that would really go hand in hand with LV2 This seems to have improved with your latest updates. It might also intersect with 8. and some instability specific with some plugins At the moment this seems specific also to looping - as said an old problem but rather annoying, but yes maybe unrelated to LV2 Actually, now that this is fixed I'd say we could keep the 'channel 1' approach. After all we are doing one plug-in per 'slot'. IMHO Midi submasters...
Some further testing. The attached starts to be a more complex (and maybe realistic) use case which uses multiple LV2 instrument, a sub and a few LV2 effects on single tracks. I have spotted a couple of plugins which really seem to disrupt Rosegarden, and also make it 'unrecoverable', meaning that even removing them will mangle the tempo (everything seems to slow down): - amsynth seems to grind evrything to a halt and slow down, xrun, etc. and once gave me an 'out of audio processing power' warning...
Yes there seem to be no hanging notes. There's still some weirdness with dexter... something like 'unstopped' notes. Unfortunatelly dexter is quite hard for me to test as LV2 as it's really small and cannot be scaled (nothing to do with rosegarden)... Maybe someone could do a bit of testing, or.. I could set-up a tune which uses the 'standalone' version (which can scale) and see how it sounds in rosegarden.
Well I can confirm this with 24.06 with DSSI synths see audio (and .rg) example attached of playing a looped segment via hexter
First... This seems to be coming along pretty nicely. I was now able to create a couple of tracks with different synths and send them to a sub1 with effects which were therefore common (but see about crashing if changing subs while playing). Adding a few random things I noticed in further testing today beyond what I reported as responses to the previous issues: - There is still a click on loop start when looping - Maybe related and seems only with certain plugins like Yoshimi or Dexed (polyphonic...
First... This seems to be coming along pretty nicely. I was now able to create a couple of tracks with different synths and send them to a sub1 with effects which were therefore common (but see about crashing if changing subs while playing). Adding a few random things I noticed in further testing today beyond what I reported as responses to the previous issues: - There is still a click on loop start when looping - Maybe related and seems only with certain plugins like Yoshimi or Dexed (polyphonic...
First... This seems to be coming along pretty nicely. I was now able to create a couple of tracks with different synths and send them to a sub1 with effects which were therefore common (but see about crashing if changing subs while playing). Adding a few random things I noticed in further testing today beyond what I reported as responses to the previous issues: - There is still a click on loop start when looping - Maybe related and seems only with certain plugins like Yoshimi or Dexed (polyphonic...
yes this seems to work now!
Yes, this seems to work now. It does however seem to crash still if you change / add submasters while playing (with LV2 plugins).... What happens is you hear sound for a bit (with parameters / patches which seem to get reset), and the Rosegarden crashes
Some dialongs have weird text placement by default
Mixers don't visually scale
I tried testing with 3 synth (instrument) plugins and it seems the instrument selection dropdown (in the Track parameters panel) gets populated for all of the channels with the last selected plugin and not with the various plugins on each channel
With the latest pulled from the branch yoshimi seems to work as expected (on its single channel).
Tried this and situation seems improved. Unfortunately I'm still getting the clicks (which sound more ass 'chopped' notes if I loop a part on the very first note if it is at the beginning of a loop
OK I see your point and can agree about instruments. However for actual effects (like delay, reverb) I feel that some some sort of 'bus' / routing would be very useful. Think of a a set of instruments (synth) where you want to have the same reverb for all to make them sound as if they were in the same 'space' you'd send all their outputs to the same bus. The same would go for a limiter you want to put on the whole song etc. But again this would be another feature request
Hi Philip, I haven't had a chance to test yet... But your discoveries about MIDI channel make me think that we'd probably need something like a "MIDI Bus" too. Yoshimi is a good example. It allows up to 16 channels, so it would make sense to place it on a MIDI bus and then have tracks with different tracks send to that. Same goes for plugins like the Calf Fluidsynth where you can have up to 16 instruments on 16 MIDI channels etc. Not sure how complex this would be and would really only make sense...
Hi Philip, nice progress! Reporting again on a few tests based on your latest messages: Unfortunately not as extensive as last time due to time constraints on my side but hopefully useful. Let me know if there's something specific I should test more. Black Pearl now works even without notes (from the editor / GUI) There are several plugins which now work both which in my previous testing crashed or were unstable. Didn't test them all but definitely (and including editors) - J303 - Dexed - OB-Xd Yoshimi:...
Hi Philip. I'll try and build and test as soon as possible and report back. In the meantime, I'm not sure if this info is useful / relevant, but in the MIDI spec there's both a note off 'status' as well as a note with velocity 0 (zero) meaning note off and I think especially with plugins (but also some hardware) this might have created issues at times. Not sure what rosegarden is doing. I find the first answer in this forum sums it up well :-) https://www.kvraudio.com/forum/viewtopic.php?t=29189...
Hi Philip, thanks for looking into the instruments. Will answer here to the various questions and based on pulling and re-compiling your latest changes in the branch. - I do have the gtk2 package installed on manjaro (arch based). That should include development files IIRC - it seems indeed that now the calf plugin guis (for the instruments which I have tested), open and work - kars plugin: the editor button is grayed out. Maybe my version of the plugin? I did notice trying it in Carla that no 'native'...
A couple of general notes to add: for instrument plugins there should be a way to route a mono plugin to play to the the two main stereo (Rosegarden) outs, essentially merging the mono out into a stereo. Otherwise the result is hearing the plugin only on the left channel there seems to be no way to get presets from plugins, even without the external editor?
This took a while to set-up and test So, first good news, I was able to build this fork and run, it listed LV2 plugins and some play :-) yay! What I tried doing is quite systematically try out all 'instrument' plugins I have installed (not some of these I hardly ever use, sometimes install with 'group' in manjaro (arch based). Why instruments? I think these are the most relevant and should definitely work in Rosegarden. Of course effects are important too, but my main motive to push LV2 is to: a)...
Hey Philip this sounds amazing honestly and I'm eager to test (time permitting). Could you recap how to build a test version for this?
Update to Italian translation 2023-12
Thanks! Yes I removed files in ~/.local/share/rosegarden/library and re-installed latest git and it works. I think with the updated script I should be able to keep more up-to-date with Yoshimi and push the updated rgd file more regularly!
Yoshimi rgd file is out of date
Tested with latest git. Seems to work as expected now. I tested by - Creating a new file - Changing the Program to 122 (in Rosegarden, that is MIDI program 121 in 0-based) - Exporting to MIDI I checked the output midi file with: - Rosegarden midi import of the just exported file: the track is correctly set to program 122 - Checked the file with the midicomp (https://github.com/markc/midicomp) - Imported in another sequencer (Muse): the program change is there
Fix #1663 - check that the midi event is a controller
Fix #1663 - Added check to make sure this is a control change with data1 == 121 othewise any event with that data1 would be dropped
Program change 122 (121) erroneously skipped when exporting to MIDI file
IMHO from a user perspective it's acceptable that if I try to send a stereo signal (instrument or audio) to a mono plug-in this wouldn't be allowed. In this case maybe it would be good to have a warning for the user. Even better... if the instrument / audio is stereo don't even show mono plug-ins as an option. The scenario where one might want to send a stereo instrument to a mono plug-in is because this plug-in only exists in its mono version (e.g. guitar emulation amp or similar), but in that case...
Indeed with latest git (just pulled and compiled) it works as expected, i.e. re-labelling has the current label in the text field ;-) Thanks!
Indeed with latest git (just pulled and compiled) it works as expected, i.e. re-labelling has the current label in the text field ;-)
When selecting Relabel for a segment the existing label should appear in the text entry
Now that #1656 is possibly solved IMHO this is definitely something to have in the next version. Re-testing it makes it so much more usable to enter e.g. fast arpeggio (16th) notes, or other parts where both horizontal but also 'vertical' (i.e. timing) is relevant. As per comment above maybe a lighter shade of green for the preview, but that's maybe personal taste - of... even better configurable somewhere?
Now that #1656 is possibly solved IMHO this is definitely something to have in the next version. Re-testing it makes it so much more usable to enter e.g. fast arpeggio (16th) notes, or other parts where both horizontal but also 'vertical' (i.e. timing) is relevant. As per comment above maybe a lighter shade of green for the preview, but that's maybe personal taste - of... even better configurable somewhere?
Now that https://sourceforge.net/p/rosegarden/bugs/1656/ is possibly solved IMHO this is definitely something to have in the next version. Re-testing it makes it so much more usable to enter e.g. fast arpeggio (16th) notes, or other parts where both horizontal but also 'vertical' (i.e. timing) is relevant. As per comment above maybe a lighter shade of green for the preview, but that's maybe personal taste - of... even better configurable somewhere?
Works for me too. Compiled the latest git and created a segment tried doing what caused this before (quickly scrolling up and down) and nor the piano nor the note 'preview' seem to have that artefact.
Matrix vertical 'piano view' refresh problem in matrix editor when mouse scrolling
Hi Ted, I'm a little confused... How can you see those in 22.12.1? Was the green indicator already there? To clarify I only see them now with the latest git
So I pulled the latest git and tested. The vertical lines are great, make it so much easier to place e.g. 16th notes! The 'preview' is nice, maybe the green should be a bit lighter, something no more than #c8ffc8 ? Also I have an issue where if I scroll vertically with the mouse wheel 'traces' (the green rectangles) remain on notes where the mouse pointer isn't any more, and also the piano display on the left is mangled). Changing zoom level removes all of these graphical quirks, see screenshot of...
Possible bug with deleting notes when the controller ruler is open in the Matrix
Rosegarden screenshots and debug links
Rosegarden screenshots and debug links
Rosegarden screenshots and debug links
Rosegarden screenshots and debug links
The feature itself is so cool IMHO that I can personally live with a 1px discrepancy at high zoom levels. Just my two cents :-) On 18/05/2023 22:15, Ted Felix wrote: Pushed as [216583] https://sourceforge.net/p/rosegarden/git/ci/216583757bc4a6fba7708ea57ee5842612a3f22c/. Please test latest git. Ok. Doesn't really bother me. I'm pretty good with rounding errors. I might be able to find it. But if there is wildly different math, that does make it difficult to track down. We should probably address...
There is still a small glitch, especially visible at high zooming rates where if you drag the view around by clickin in the bottom preview the controller grid is slightly off compared to the main grid and measures. This disappears / is not effected if you drag around using the actual horizontal scroll bar (or similarly holding SHIFT while using the mouse wheel). It also seems to be 'reset' to correct view if you scroll vertically where the notes are or move the zoom with the wheel in the bottom right...
There is still a small glitch, especially visible at high zooming rates where if you drag the view around by clickin in the bottom preview the controller grid is slightly off compared to the main grid and measures. This disappears / is not effected if you drag around using the actual horizontal scroll bar (or similarly holding SHIFT while using the mouse wheel). It also seems to be 'reset' to correct view if you scroll vertically where the notes are or move the zoom with the wheel in the bottom right...
There is still a small glitch, especially visible at high zooming rates where if you drag the view around by clickin in the bottom preview the controller grid is slightly off compared to the main grid and measures. This disappears / is not effected if you drag around using the actual horizontal scroll bar (or similarly holding SHIFT while using the mouse wheel). It also seems to be 'reset' to correct view if you scroll vertically where the notes are or move the zoom with the wheel in the bottom right...
There is still a small glitch, especially visible at high zooming rates where if you drag the view around by clickin in the bottom preview the controller grid is slightly off compared to the main grid and measures. This disappears / is not effected if you drag around using the actual horizontal scroll bar (or similarly holding SHIFT while using the mouse wheel). It also seems to be 'reset' to correct view if you scroll vertically where the notes are or move the zoom with the wheel in the bottom right...
Quickly pulled and built from git (12367840fe (22.12-181-gfbfc27980) and both the 'notes deleted by changing velocity in the ruler' and the 'misaligned controllers' when dragging seem to have been solved
OK. I managed to download just the snapshot for feature-509-controller-grid as a zip, build and test, both the velocity -> same time deletion and move cotnroller / pitch seem fixed. Maybe I can test when it's all merged in the next useful git ;-) Thanks Philip!
Hi Philip, I think sourceforge has problems with comments, I always get an error from their CDN when posting (but then the post is there). I'll try with your git repository, it is really slow to clone now (in Italy at least)
I'm getting something similar where the notes starting at the same time all get deleted if you select all notes and change their velocities via the ruler (not if you change them using the keyboard shortcut (e.g. SHIFT UP/DOWN arrow). See attached gif
This is really really nice! I think I might have found a small bug both with controller and pitch bend where the grid is not kept when moving an already entered controller Activate grid (e.g. 1/16 or 1/8) Place a controller: it is correctly placed on a grid line Try 'grabbing' the controller and moving it... The controller is moved horizontally by something which seems a fixed space but which is not on the grid See also attached screengrab
This is really really nice! I think I might have found a small bug both with controller and pitch bend where the grid is not kept when moving an already entered controller Activate grid (e.g. 1/16 or 1/8) Place a controller: it is correctly placed on a grid line Try 'grabbing' the controller and moving it... The controller is moved horizontally by something which seems a fixed space but which is not on the grid See also attached screengrab