[CLOSED] Cakewalk 2019.09 Early Access

[Noel Borthwick]

On 9/27/2019 at 10:03 AM, huviolan said:

Hi Everybody, .

I don´t know if the bug i´m gonna describe was present before this Early Access version, (I have it installed so I don´t know if this bug was present in the last stable version of Cake).
The fact is:  if you want to go to full screen on the video view window AND you are using Direct Show like the engine for video media (because that´s the only way to be able to move around in the timeline for a video clip. As far as I know it is not posible to do that with Windows Media Foundation engine), you only get a flickering  on the screen and the video window reverts inmediatly to the default stretch mode (which is "preserve aspect ratio"). Is this the way it shoud be?? Because with the Media Foundation engine no problem going to full screen, but no way to get it if you are using Direct Show. 
I hope I've been clear enough...despite of my english
Thanks a lot in advance for any clarification about this!
 

Hi huviolan, this wouldn't be related at all to this months release. There have been zero changes to video for a very long time.
Media foundation doesn't support hardware full screen mode as far as I know. It has never worked. 

[OeAi]

1 hour ago, msmcleod said:

This is usually down to a MIDI prepare buffer that is too small. The default is 250, but I tend to use 750:

i had 512, set it to 1024 - yes there is difference, but not really, anyway there are same faults
it's like it's not reading pedal automation and some keys plays with different velocity - it's a different sound.
the footprint is more like equal, but you can listen - it's not what it must be.

syntr-test-audible-bounce.mp3 syntr-test-fast-bounce.mp3

Edited September 28, 2019 by oeai

[msmcleod]

2 minutes ago, oeai said:

i had 512, set it to 1024 - yes there is difference, but not really, anyway there are same faults
it's like it's not reading pedal automation and some keys plays with different velocity - it's a different sound.
you can listen - it's not what it must be.

syntr-test-audible-bounce.mp3 1.3 MB · 0 downloads syntr-test-fast-bounce.mp3 1.3 MB · 0 downloads

It's worth noting that not all synths support fast-bounce. 

What is the synth you're having issues with?

[Noel Borthwick]

22 hours ago, MarianoGF said:

I know that the following is previous to the 2019.9 Early Acces, but it could be great you consider to adressing it for the upcoming release:

When I edit real time stretched clips, everything sounds fine. But when I bounce them or the whole project, it appears clicks on the former edit points or crossfades between those time stretched clips.

Here I attach a .cwp example file with audio.

This issue occurs even if I choose Elastique Pro as the stretch method for both online and offline renders. And it happens with any kind of bouncing: bounce to clips, to tracks, freeze and even export (fast bounce or not).

The workaround I can do in the meantime is to bounce those clips INMEDIATELLY AFTER stretch, and BEFORE any subsequent editing.

If I'm not the only one experiencing this, hope you consider it for fixing.
Thanks!

Thanks for sending the project file. Are you referring to the small spikes at the point where you do the crossfades? The stretches are performed per clip when bouncing but the crossfades are rendered post stretch (after the stretch is applied) so it might be that the sample accuracy of the crossfade positions is not accurate. 
Jut a guess though. I havent investigated it. In your project I found that the clicks go away in the bounce if you have wider crossfades. 
Will investigate further, but this wont be in the 09 release since we are almost done here.

[OeAi]

4 minutes ago, msmcleod said:

It's worth noting that not all synths support fast-bounce. 

What is the synth you're having issues with?

syntronik, maybe sampletank too as it is one company and maybe rapture session.
i remember that i had such problem before, but not sure with instrument, maybe that was syntronik too.

Edited September 28, 2019 by oeai

[Noel Borthwick]

 

Just now, oeai said:

syntronik, maybe sampletank too as it is same company and maybe rapture session.
i remember that i had such problem before, but not sure with instrument, maybe syntronik too.

Are you sure that this is actually a new issue to this months release? As Mark said some synths don't render data properly during fast bounce. Have you tried unchecking fast bounce to see if the render is accurate yet?

[OeAi]

On 9/28/2019 at 10:56 PM, Noel Borthwick said:

Are you sure that this is actually a new issue to this months release? As Mark said some synths don't render data properly during fast bounce. Have you tried unchecking fast bounce to see if the render is accurate yet?

well, i'm not sure about month release, i don't see such fix for bouncing tracks, but you got something related, so it's easier to catch when you remember the code, because i think that it's something about reading automatization, maybe sync-related, but mostly bouncing.
Syntronik and sampletank uses mostly the same engine, and most of the time i had sampletank with 1 instrument bounced norm.
i've added fast-bounce and no-fast-bounce in attchment. What i can say that there is 2 instruments used in one for this track, there can be 4 for syntr and 16 for sampletank and maybe the problem is that some fast bounce engine code can't recognize that there is something else or work somehow different, so it's not sending some data or not taking the response (feedback problem ). or multithreading and multicore distribution support for different synths used in one vst.
Yes. i can believe that they've made some errors in their code, but maybe there should be not-so-fast-bounce, but not real-time too that will render good, because i think that real-time rendering can cause some problems with high-cpu load, so it should be somewhat different.

Edited September 29, 2019 by oeai

[Noel Borthwick]

1 hour ago, Jeremy Jensen said:

When I installed this EA release, it made it so my Arturia Piano V stopped making sound. How do I revert? I'm hoping I don't have to uninstall and reinstall.

@Jeremy JensenThanks for your report. I was able to reproduce this. This is caused by bugs in Arturia program management that I have reported to them.
I have worked around these issues for the other arturia plugins but apparently the workaround doesn't work for this specific synth. I'll look into it.

 

[Noel Borthwick]

16 minutes ago, oeai said:

well, i'm not sure about month release, i don't see such fix for bouncing tracks, but you got something related, so it's easier to catch when you remember the code, because i think that it's something about reading automatization, maybe sync-related, but mostly bouncing.
Syntronik and sampletank uses mostly the same engine, and most of the time i had sampletank with 1 instrument bounced norm.
i've added fast-bounce and no-fast-bounce in attchment. What i can say that there is 2 instruments used in one for this track, there can be 4 for syntr and 16 for sampletank and maybe the problem is that some fast bounce engine code can't recognize that there is something else or work somehow different, so it's not sending some data or not taking the response (feedback problem ).
Yes. i can believe that they've made some errors in their code, but maybe there should be not-so-fast-bounce, but not real-time too that will render good, because i think that real-time rendering can cause some problems with high-cpu load, so it should be somewhat different.

There is no substitute for plugin vendors fixing their own code so you should contact them about these issues and send them a project to reproduce it.
Whether its fast bounce or realtime bounce there is no issue of CPU load since a dropout cannot occur during bounce. All fast bounce does is run the export as fast as the syths can render bypassing the audio driver clock. Realtime bounce does the export by running the transport. Thats the only difference.

[OeAi]

25 minutes ago, Noel Borthwick said:

There is no substitute for plugin vendors fixing their own code so you should contact them about these issues and send them a project to reproduce it.
Whether its fast bounce or realtime bounce there is no issue of CPU load since a dropout cannot occur during bounce. All fast bounce does is run the export as fast as the syths can render bypassing the audio driver clock. Realtime bounce does the export by running the transport. Thats the only difference.

ok, then maybe a new *intelligent-bounce mode? so if we know the start of notes and sustain pedal hard points - then we can render after note start with fast bounce and before the note start by taking realtime transport. maybe this can solve such a no-fast-bounce-synths. not sure if it is possible, but may be

Edited September 28, 2019 by oeai

[Noel Borthwick]

53 minutes ago, Noel Borthwick said:

@Jeremy JensenThanks for your report. I was able to reproduce this. This is caused by bugs in Arturia program management that I have reported to them.
I have worked around these issues for the other arturia plugins but apparently the workaround doesn't work for this specific synth. I'll look into it.

 

@Jeremy Jensen I was mistaken. I had an experimental build with some other stuff that isn't compatible with Arturia synths.
I am not seeing any issues with Piano -V in the early access build.  Which version of the synth are you running exactly. I have Piano V2 from VCollection 7 Version 2.4.1 and this is working fine here. I saved a project from an earlier release and opened it up in 09 and it plays fine.

If you are still seeing this please reply to this post or contact me with a link to a project file that has this issue and I can troubleshoot further with you. 

[Mark Morgon-Shaw]

Loving the Duplicate command !

[MarianoGF]

6 hours ago, Noel Borthwick said:

Thanks for sending the project file. Are you referring to the small spikes at the point where you do the crossfades? The stretches are performed per clip when bouncing but the crossfades are rendered post stretch (after the stretch is applied) so it might be that the sample accuracy of the crossfade positions is not accurate. 
Jut a guess though. I havent investigated it. In your project I found that the clicks go away in the bounce if you have wider crossfades. 
Will investigate further, but this wont be in the 09 release since we are almost done here.

Thanks Noel.

Yes, I'm talking about those spikes. They are not present when I edit in real time, but they appear after bouncing.

Hope you take it under consideration for another update.

[TheSteven]

Just started checking this out - first thing I've noticed is that this version of CbB starts up 10x faster than current release.

[marcL]

Yesterday I worked the whole day with this EA version. I am really pleased and happy with it! ? Together with the 2019-07 release it represents a milestone in the history of Sonar/CbB. It is definitly the best and most stable Sonar/CbB ever! There are so many helpful improvements and fixes in such a short time, incredibly!

These are my favorites:

• Issues with Focusrite interfaces fixed (gorgeous!)
• Take lane enhancements: Crossfades and cropping on single lanes/clips, crossfade type choice, audio file import (via menu) to lanes, a lot of fixes like clip copy, a.s.o.
• Transport issues fixed (stop playing jumping to project start or not to now time, play engine looping)
• Reduced starting time

Many thanks to Noel, the Bakers and the BandLab team! I almost have no wishes for CbB anymore, it is perfect! ?

@Noel Borthwick  I noticed that if you select 2 or more clips and then click on a single one of them, then the selection does not change to this one as it was in the earlier releases. Is this intended? Other then this I haven't found issues.

Edited September 29, 2019 by marled
invalid line break

[Wookiee]

18 hours ago, Jeremy Jensen said:

When I installed this EA release, it made it so my Arturia Piano V stopped making sound. How do I revert? I'm hoping I don't have to uninstall and reinstall.

@Jeremy Jensen What version of the V Piano are you using?

Piano V Version 1.1.0.205 ?
Piano V 2 Version 2.4.1.2810 ?

Can you not upgrade to Version 2 of the V Piano? 

[Andrew Pentecost]

Not sure if this is just an early access issue as I have not tried ACT Controller before but it seems a bit unstable. It stops responding after a while and needs to be rebooted (eg after setting up 8 sliders it refused to map any further parameters, and also stopped controlling the console, but after a reboot it carried on working perfectly).

Also, it is very slow when detecting a plugin (takes several seconds).

Is there any possibility that ACT Controller can have a user configurable number of faders etc (or at least make it 9 as it is very common to have a master fader which could be well used for an organ VSTi, which commonly has 9 drawbars, as well as any master fader duties).

[Noel Borthwick]

8 hours ago, marled said:

Yesterday I worked the whole day with this EA version. I am really pleased and happy with it! ? Together with the 2019-07 release it represents a milestone in the history of Sonar/CbB. It is definitly the best and most stable Sonar/CbB ever! There are so many helpful improvements and fixes in such a short time, incredibly!

These are my favorites:

• Issues with Focusrite interfaces fixed (gorgeous!)
• Take lane enhancements: Crossfades and cropping on single lanes/clips, crossfade type choice, audio file import (via menu) to lanes, a lot of fixes like clip copy, a.s.o.
• Transport issues fixed (stop playing jumping to project start or not to now time, play engine looping)
• Reduced starting time

Many thanks to Noel, the Bakers and the BandLab team! I almost have no wishes for CbB anymore, it is perfect! ?

@Noel Borthwick  I noticed that if you select 2 or more clips and then click on a single one of them, then the selection does not change to this one as it was in the earlier releases. Is this intended? Other then this I haven't found issues.

@marled I cannot repro your clip selection issue. Can you post a short video or upload a small project that exhibits this?
Is this with the smart tool or another tool?

[Noel Borthwick]

13 hours ago, TheSteven said:

Just started checking this out - first thing I've noticed is that this version of CbB starts up 10x faster than current release.

If you had tons of 32 bit plugins this is expected. There are optimizations to cache the inventory preventing too many registry accesses.

[Noel Borthwick]

19 hours ago, Noel Borthwick said:

@Jeremy Jensen I was mistaken. I had an experimental build with some other stuff that isn't compatible with Arturia synths.
I am not seeing any issues with Piano -V in the early access build.  Which version of the synth are you running exactly. I have Piano V2 from VCollection 7 Version 2.4.1 and this is working fine here. I saved a project from an earlier release and opened it up in 09 and it plays fine.

If you are still seeing this please reply to this post or contact me with a link to a project file that has this issue and I can troubleshoot further with you. 

Don't bother, I managed to work around this bug in older Arturia plugins. They go silent if we forward parameters to them when loading despite the plugin requesting the forwarding! I had to hack our code just for Arturia to avoid this.