Noel Borthwick

Hi Folks, as mentioned in the prior release that model is considered "beta" since I still have to do more testing with it. The latency problem is due to a weird race condition leading to audio buffers being lost. I just havent had time to track it down yet but will get to it. For now don't use that model until further notice.

@Maxim Vasilyev the problem with track names displaying garbage has been fixed. As mentioned above its actually caused by the way the VST3 SDK reported this so technically could have been easily addressed from the plugin side (most other VST3 plugins handle it). However I have made a change to resolve it from our end. As far as VST2 goes there is no way for a host to report track names to the plugin via the standard VST protocol so this won't work with Relay in any host. You must use the VST3 version to get this functionality. Cakewalk has supported VST3 channel naming from the get go for many years now. With regard to track ordering in Neutron, this is a deficiency in the plugin since it it appears to be sorting the track names alphabetically and ignoring the host specified track order. There is a specific VST3 interface that can be used by the plugin to identify the order of tracks. If they use that the tracks will be sorted in the same way that they appear in the host. This is how it works with the Softube Console One. BTW the same track ordering problem can be observed in other VST3 hosts such as Reaper.

I've fixed this issue as well for the next release. It had to do with some peculiarities with how Relay requests the host to do resizing via VST3.

Oh well I managed to sort out the string garbage by fixing some code in the VST3 SDK . The SDK code was not copying over the null terminator for the string ? I'll submit my change to Steinberg as well since any app that uses that would be susceptible to the same problem.

@Matthew Sorrels I finally had a chance to look at this. The problem is not in the Cakewalk code. The channel names are provided by a standard VST3 interface called IAttributeList and provided via ChannelContext::IInfoListener. In this the string is properly null terminated but VST requires the user of the string to use the string size provided and properly null terminate their target. This is not being done by the Izotope plugin hence the garbage appears. I will pass on this information to Izotope.

I cant quite follow what the problem the user is having other than the fact that something doesn't work right with REAC. Since we no longer have a REAC system to test and it is no longer supported by Roland its going to be hard to troubleshoot. One suggestion is to turn on "StopEngineOnASIOPanelOpen" since some Roland devices don't like the audio engine running when the ASIO panel is opened.

A fix for this is now available in the hotfix we just released.

After several hours of troubleshooting and a tip from @Steve Katsikas I managed to reproduce the issue again. Its a bizarre preexisting problem that happens when you seek on the timeline to a late measure. In my testing the app would appear hang for about 5-10 seconds and then vanish silently. I was able to find the root cause and fix it last night. The engine was incorrectly responding to a seek on the timeline and computing a bogus delay compensation value. In Steve's case this could compute a delay of up to 3 hours ? This value would then be used to send a HUGE MIDI buffer the next time playback started causing an overflow and crash. While most users wouldn't necessarily see a crash (since its dependent on the amount of MIDI data) this issue results in way more MIDI data being processed when playback starts at a high measure number. I've sent Steve a build to verify the fix and if all is well we'll be releasing a hotfix with this and a few other issues soon. I'm happy that this issue is fixed since it could have been the cause of many unknown random crashes in the field over the years - the problem has been latent at least since 2006 Fortunately his project was reliably able to reproduce this and allow us to fix it. Thanks for working with us to repro the issue.

@Steve Katsikas its the strangest thing. I tried for 30 mins today to reproduce this issue and it is no longer happening. I went back to 2020.01 and even SONAR and the project plays fine at 48K and I even the your markers while playing extensively. Both myself and a colleague were able to reproduce it earlier but the issue seems to have gone away. Can you please retest this after updating windows since its not impossible that there was some external update that fixed itself. If you can still reproduce it please see if you can get a sequence of steps that makes it more likely to happen. @Bob4u I think your issue may be something else based on the crash. Please ensure that your VC redists are up to date by downloading and updating the latest 64 bit VC redist from microsoft.

I posted an update for this issue here.

Hi @Robert Bone @FJ Lamela and others who had this issue. Please try out this version of VST scanner. I've updated the logic so that it is more specific about the VST3 folder rather than just checking for VST3, It will now only look at the standard VST3 folders under Program files or Program Files (x86). Also it wont get tripped up by names like VST32. You will need to copy that file into C:\Program Files\Cakewalk\Shared Utilities and then scan again. Let me know if it fixes the issue.

@FJ Lamela Some questions: Are the 7 plugins removed when you scan from cakewalk the same plugins from c:/vst32/? Do any of those plug-ins reside within a folder that has the name VST3? Are the removed plugins VST2 plugins?

There have been no changes to Plug-in manager recently. All it does is utilize the VST scanner for scanning but as has been mentioned numerous times, all scanning should be done from Cakewalk not plugin manager. We don't maintain or test scanning from plug-in manager and the scan functionality from there will be removed soon. There are many other improvements to scanning such as sandbox scanning that are only available from the Cakewalk preferences dialog. We'll look into why this happened but please use the standard preferences page for scanning, which has been the preferred way to do it for many years.

@mgustavoWe'll investigate it for a future release. Thanks for testing it.

Yes exactly. Please don't manually remove DLL's from those folders that you are not aware of since you will likely break existing VST3 plugins. Vendors may have necessary helper dlls there. My point was that you should not install VST2 plug-in's there.

@rv729 How did you end up getting your VST2 plugins placed in the reserved VST3 folder? Actual plugins in that folder should always have the .VST3 extension. You should never install VST2 plugins in the VST3 folder since it will no longer work. As of 2019.01 we exclusively scan VST3 plugins in the Program Files\Common Files\VST3 folder. This was done because some vendors install child DLL files for their plugins there and scanning these dlls cause other problems.

Please send the dump file for analysis. Your issue may not be the same problem

@Steve KatsikasThere is something strange going on with that project leading to a MIDI buffer overflow. I'll have to spend some time debugging it on Monday. This happens in older versions of Cakewalk as well. I went back as far as September and also tried SONAR which hangs with that project.

@Terry Kelley your dump files indicate heap corruption. I see that you are running outdated Microsoft runtimes for the VC redistributables. Here are the time stamps you have: mfc140u.dll *C:\Windows\System32\mfc140u.dll 14.16.27024.1 11/7/2018 1:49 PM VCRUNTIME140.dll *C:\Windows\System32\VCRUNTIME140.dll 14.16.27024.1 11/7/2018 1:35 PM The ones we redistribute are from a later date so your system isn't up to date. The versions from the Cakewalk installer are as below: mfc14u.dll 14.20.27508.1 3/8/2019 vcruntime.dll 14.20.27508.1 3/8/2019 Please do an install of the latest VC redists by following this link. After ensuring that you have more current versions of these files please redo the test and see if it crashes. I'm not sure if your crash issues are related but having old redist files can definitely cause crashes and other incompatibilities.

This really looks like memory corruption. What plugins or synths are you using in the project? I see that TTS1 is loaded. Also there are a ton of different ASIO drivers on your system. If its basic can you share the project so we can try and repro? What happens if you remove TTS1 and instead use a different synth does it still crash? What sample rate are you running at?

Depends on what you mean by more efficient Having a single instance is more memory efficient obviously. However its not the most CPU efficient since the DAW has no way to process each channel on a different core. If you are running into bottlenecks splitting it up into individual synths will give you a lot more CPU headroom.

@bvideoYou can read more about our patented plug-in load balancing here. A fundamental requirement of load balancing is for there to be 2 or more serial plugins that can be processed in parallel. Synths in the synth rack are SOURCES. The synth audio processing must take place individually for each synth before any further signal processing in the downstream circuit is performed. i.e. synths get fed MIDI data not audio and produce audio which is then sent to the tracks sourced by the synth. Its no different from recording audio or input monitoring in that regard. So load balancing doesn't apply here. Any parallel processing for synths has to be done within the synth itself by parallelizing the processing of its voices. This cannot be done by the host since it requires the plug-in vendor to do this. Its a double edged sword even for the plug-ins that attempt to do this however, since the threads in the plug-in start to compete with the DAW's threads and this can lead to other performance problems. I had a proposal many years ago to solve this problem but it requires cooperation with plug-in vendors to implement. For now the best way to do this if the synth doesnt support parallel processing internally is to have per channel instances of the synth and make one instrument track per channel. This way the DAW will manage the parallel processing since each synth voice will be farmed out to its own core. Of course this is a trade off between RAM and CPU performance because inserting multiple instances can take more memory depending on what samples are loaded etc. Hopefully the synth vendor internally optimizes this and shares samples between multiple instances of the synth. For TTS1 the extra memory would be negligible. TTS1 is pretty lightweight so I don't think it would be necessary to split into multiple instances typically.

@paulo There are some more WASAPI improvements in 2020.01 (esp for exclusive mode) so please try it there as well. What output device are you when using the mic in Cbb? This is important because both the output and input have to support the same sample rate and bit depth for it to work. The fix I am referring to has to do with how some devices deal with 16 bit audio. We had a case where a device would only work when the bit depth was set to 24 bit manually. This has been fixed in 2020.01 and such devices should now connect properly. Let me know if updating fixes this issue. If you are still having a problem PM me and we can do some more troubleshooting.

The MIDI tracks themselves don't take any significant processing load. When processing the audio track is where the actual plugin processing takes place. If you have a single synth all the processing work gets done on one core. So your PC specs wont guarantee no glitches since only a single core is being utilized. If that one core cannot complete processing (the data from all 16 MIDI tracks) then you will get late buffers and glitches. There is no way around that other than the synth doing some work on their end. How you can utilize more cores is by creating separate SI tracks for of Sample that each handle a single part. Though this is not as memory efficient it will significantly reduce the load for any core. Also each instance of ST will be processed by a different core in parallel so it will be a lot more efficient. Try it and let me know if it solves the glitches.

@Skyline_UK This is caused by excess load being consumed by SampleTank. This has no bearing on your audio interface directly. If you have a single audio track feeding the synth this means that all the work is being done by a single core. If it takes longer than the deadline to process the buffer you will hear pops and clicks. Smaller buffer sizes mean there is a shorter time limit to process buffers and if the synth takes longer it will glitch. >>my engine load flickers between 70% and 90% followed by a figure in brackets of 102% to 105% once I start Play The figure in brackets is the MAX engine load since playback started. So if its > 100% it means that some buffers took longer than the max allotted time. You should be seeing late buffers reported for this case. One thing you can try in this use case is see if plug-in load balancing helps. Load balancing will use extra cores to process the plugin buffers so it should help in the case when you want to work at 512 or 1048 samples. Assuming that the plugin is well behaved with multiprocessing of course.