Perhaps their “questionable” techniques present me with “questionable” use of their product.
Wow. No kidding. Hope you don’t have to release a special “Cantabile- Roland Users” edition
The makers of OP-X had to make a special change for proper C3 use and came through great (they tried to help with a fix on their end). Could it be tied to the fact that Mainstage was the original intended host for the live softs by Roland? Just wondering how that live host and other DAWs can use it and not get the same troubles. Definitely a buger-bear (a little Canadian term I picked up while there)
When I get back from this business trip I will see how Roland Cloud users can report issues. Maybe if enough do it…
Firstly these problems aren’t unique to Cantabile - Roland has confirmed these issues occur in some hosts, but there’s two things about Cantabile that make the problem worse:
Many hosts employ a user-interface where the plugin editors appear as a child window of the main window. Cantabile’s plugin editors appear as popups meaning a different top level window so the techniques used by Roland don’t work in this situation.
Cantabile’s per-window resolution support allows plugin editors to appear at low resolution (correct size) while Cantabile’s main UI runs at high-resolution. I don’t know of any other hosts do this (in fact for user interfaces like described in 1, this is currently impossible because the Win32 API dictates that different resolutions are only supported on different top level windows). When the plugin arbitrarily decides to use another window (Cantabile’s main window), the resolution context gets all screwed up, the UI is no longer sandboxed to that environment and things go haywire resulting in crazy oversized, non-functioning context menus and other glitches.
That second point makes what might be benign bugs in other hosts detrimental to Cantabile.
This is what happens when a developer goes outside the spec and tests only in a few different hosts. Really they should be constraining themselves to the spec, or at least checking which host they’re running in (the VST spec provides an API to find out the host name) and limiting their funky behaviour to the hosts they know about.
What’s so frustrating about this is I can’t see what they’re trying to do that can’t be done using what’s already available in the VST API.
Ugh, end of rant for now, back to debugging… (while I’m at it I might put on The Killers “For Reasons Unknown” - seems appropriate)
Thank you for the extended explanation.
Who did they hire to code those things? They do sound really good but everything about them; the wonky UIs, the CPU usage… it’s just not what you expect from one of the world’s major corporations.
OK - I think I’ve got this sorted.
Build 3532 (available now) introduces a number of work arounds to fix the known issues. I’ve tested it with a handful of Roland Cloud plugins and it seems to be working but I haven’t had time to test every plugin.
Please let me know if you notice anything else amiss. These fixes won’t be retro-fitted to 32xx because the issues are more benign there.
Here’s build 3531 showing the problems:
Here’s build 3532 looking much better:
Is it advisable to move from 32xx to 35xx for performance yet? I’m having to create a couple of rigs for gigs later this year, and the option to reliably run Roland cloud instruments would be helpful.
So far I’ve heeded the warning on the download page.
I had played a couple gigs with 35xx - good to go!
Hi @The_Elf, it’s basically stable but not as well tested yet because most people are still on 32xx. Be sure to test it well in a practice session before going live.
Soon I’ll be marking it as stable to get more people moved over but wouldn’t be surprised if that shows up some (hopefully) minor issues.
I have 35XX on my backup laptop. I used it on 4 gigs so far with no problems. I am going to transition my main laptop as soon as Brad pulls the trigger on stability, and I get time to do it.
Who's working with 3500+?
Shame you have had to spend effort to work around another company’s problems. I’ve had to do that myself, if it is any consolation. My ex.factory librarian has to work around bugs in the Yamaha EX5 MIDI SYSEX implementation (which can corrupt your data). In 2003 I gave Yamaha UK a detailed bug report, they sent it to Yamaha Japan. Yamaha Japan eventually came back and denied that the EX5 even supported SYSEX data transfer, despite that being patently false!!!
35xx is looking pretty stable over here as well. Tomorrow I will be back home, and will download and test the latest C3 build with the Roland Cloud synths.
Good to hear positive reports of 35xx. I’ll soon begin building the new Catabile racks for the gigs later this year (I’ve learned a lot since my previous attempts, so it’s a start-over!), so I’ll use the newest version.