Glitches and sync issues with Cakewalk and TASCAM US-600

[miltonmagalhaes]

I have a Tascam US 600 that I use in Cakewalk.
Although Cakewalk developers claim that it is not possible to use it with non-native ASIO drivers, I installed FlexAsio and have had very good results, as it is allowing me to "survive" my US 600 interface.
However, in some projects I experienced a slowness/latency when starting the recording when entering the metronome that I thought could be problems with the Windows 11 antivirus that I leave turned off.
Any tips? Thanks.

[dechamps]

This almost certainly has nothing to do with your antivirus. Especially if the slowness happens when it's disabled.

If you can provide a log while the slowness is happening, I can take a look at what may be causing it.

[miltonmagalhaes]

Hello, thank you for your quick response, but I don't know how to generate a log during this occurrence. Can you teach me how to do this? Milton Sica Magalhães

…

________________________________ De: Etienne Dechamps ***@***.***> Enviado: terça-feira, 17 de setembro de 2024 09:12 Para: dechamps/FlexASIO ***@***.***> Cc: miltonmagalhaes ***@***.***>; Author ***@***.***> Assunto: Re: [dechamps/FlexASIO] Slowness/Anti Virus/Windows 11 (Issue #245) This almost certainly has nothing to do with your antivirus. If you can provide a log<https://github.com/dechamps/FlexASIO?tab=readme-ov-file#logging> while the slowness is happening, I can take a look at what may be causing it. — Reply to this email directly, view it on GitHub<#245 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/ACCEFHITFFLK4BZJULSLHS3ZXAMEPAVCNFSM6AAAAABOLJYZC6VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDGNJVGU2TANJSGQ>. You are receiving this because you authored the thread.Message ID: ***@***.***>

[dechamps]

This is explained at the link I posted in my previous message.

[miltonmagalhaes]

Do you receive my log ?

[dechamps]

No, I did not. You should be able to simply attach it to a GitHub comment.

[miltonmagalhaes]

FleASIO.log

[dechamps]

I'm a bit confused because in your log it looks like you never actually play/record any audio. It also appears to be truncated (did you send the log while the app was running?).

Anyway, it looks like Cakewalk repeatedly initialized and deinitialized the ASIO driver 8 times(!) and at no point did it attempt to stream any audio. This looks like some kind of driver probing behavior from Cakewalk, but 8 times in a row seems a bit excessive... not sure if you did anything to cause that (did you repeatedly select/unselect the driver or something like that?).

Anyway, every time this happens it takes about 1.2 seconds for FlexASIO to initialize. This is kinda expected as PortAudio is not really optimized for initialization speed, especially when there are more than a couple of audio devices to enumerate.

If you send a log that actually shows streaming audio, I may be able to check what's happening right before streaming starts (this is when the problem occurs, right?)

[miltonmagalhaes]

FlexASIO.zip

1 - I created an empty FlexAsio.log.
2 - I entered Cakewalk.
3 - I created a new track to record with all the plugins working/connected.
4 - I changed the FlexAsio buffers.
5 - The recording had pops.
I only found a certain decrease in delays when I matched the FlexAsio buffers with the Cakewalk recording cache. 2048.
6 - I closed the project and Cakewalk.

[dechamps]

I'm confused. Are you complaining about pops, or about delays in starting the recording? These are very different problems.

[miltonmagalhaes]

Both.

There is delay and there are noises and pops.

[miltonmagalhaes]

I created a simple project and set it to record and executed the US 600 reconfiguration 2 or 3 times to see if it appears in the log.

In my ideal model, this US 600 restart process would be demanded by the Driver, being transparent to the user.
FlexASIO.log

[dechamps]

Okay so you are not actually complaining about a delay in starting the recording, i.e. the time it takes for streaming to start. Instead, you seem to be facing issues with (1) discontinuities/pops and (2) incorrect latency reporting causing sync issues.

To some extent, getting glitches/pop if you set the buffer size too low is normal, as explained in the FAQ.

As for the desync, this probably comes down to PortAudio not reporting correct timing information - there is a FAQ about this as well.

It's possible the behavior could be improved, but I wouldn't hold my breath.

[miltonmagalhaes]

For me, as an end user, it's all about clicks and delays. This is my perception of performance, because I want the application to record in sync with the click time.

I can only achieve this if I do as I said, reconfiguring the interface.

I think this is one of the main goals of your tool: to overcome difficulties brought by native/dedicated drivers on "obsolete" hardware.

So, I look forward to your efforts in helping the community overcome such problems and maximize the use of your approach.

Thanks,