Mac FX failing pluginval.....

[Lindon]

@Lindon well...all very interesting....

see this

This is one of the script node elements in my plugin.....its got three active filters at any given time, and the filter type is set by mode....see that last control there labeled "Feedback"?

it feeds an amount of the processed signal back into the input....set this to zero and pluginval is happy, set it to anything else and we get the above errors.... so essentially the feedback loop is emphasising the resonant nature of some of these filters, and thus generating some audio in the buffer, even when no input is present....

All I need do now is (probably) add an envelope follower at input and tie it to some output volume....so 0 in == 0 out...

[Lindon]

@Lindon well at some gross level that seems to work - fine tuning required...

[Lindon]

@Lindon or of course I could just "suspend on Silence"...and see how that works...

[DanH]

@Lindon I've had troubles in the past with feeding back in scriptnode + plugins not validating...

[Lindon]

@Lindon well suspend on silence terminates with predjudice...

Starting tests in: pluginval / Automation...
Testing with sample rate [44100] and block size [64] and sub-block size [32]
Testing with sample rate [44100] and block size [128] and sub-block size [32]
Testing with sample rate [44100] and block size [256] and sub-block size [32]
!!! Test 7 failed: NaNs found in buffer -- Expected value: 0, Actual value: 256
Testing with sample rate [44100] and block size [512] and sub-block size [32]
!!! Test 10 failed: NaNs found in buffer -- Expected value: 0, Actual value: 1024
Testing with sample rate [44100] and block size [1024] and sub-block size [32]
!!! Test 13 failed: NaNs found in buffer -- Expected value: 0, Actual value: 1408
Testing with sample rate [48000] and block size [64] and sub-block size [32]
pluginval received Segmentation fault: 11, exiting immediately
pluginval received Segmentation fault: 11, exiting immediately
pluginval received Segmentation fault: 11, exiting immediately

Finished validating: /Users/lindonparker/Desktop/testAUVSTS/Planet9145.vst3
ALL TESTS PASSED

Finished batch validation

...well ALL TESTS for sure did not PASS

[Lindon]

@Lindon and actually its not getting any better....

Starting tests in: pluginval / Automation...
Testing with sample rate [44100] and block size [64] and sub-block size [32]
Testing with sample rate [44100] and block size [128] and sub-block size [32]
Testing with sample rate [44100] and block size [256] and sub-block size [32]
Testing with sample rate [44100] and block size [512] and sub-block size [32]
!!! Test 10 failed: NaNs found in buffer -- Expected value: 0, Actual value: 576
Testing with sample rate [44100] and block size [1024] and sub-block size [32]
!!! Test 13 failed: NaNs found in buffer -- Expected value: 0, Actual value: 2048
Testing with sample rate [48000] and block size [64] and sub-block size [32]
!!! Test 16 failed: NaNs found in buffer -- Expected value: 0, Actual value: 128
Testing with sample rate [48000] and block size [128] and sub-block size [32]
!!! Test 19 failed: NaNs found in buffer -- Expected value: 0, Actual value: 256
Testing with sample rate [48000] and block size [256] and sub-block size [32]
!!! Test 22 failed: NaNs found in buffer -- Expected value: 0, Actual value: 512
Testing with sample rate [48000] and block size [512] and sub-block size [32]
!!! Test 25 failed: NaNs found in buffer -- Expected value: 0, Actual value: 1024
Testing with sample rate [48000] and block size [1024] and sub-block size [32]
!!! Test 28 failed: NaNs found in buffer -- Expected value: 0, Actual value: 2048
Testing with sample rate [96000] and block size [64] and sub-block size [32]
!!! Test 31 failed: NaNs found in buffer -- Expected value: 0, Actual value: 128
Testing with sample rate [96000] and block size [128] and sub-block size [32]
!!! Test 34 failed: NaNs found in buffer -- Expected value: 0, Actual value: 256
Testing with sample rate [96000] and block size [256] and sub-block size [32]
!!! Test 37 failed: NaNs found in buffer -- Expected value: 0, Actual value: 512
Testing with sample rate [96000] and block size [512] and sub-block size [32]
!!! Test 40 failed: NaNs found in buffer -- Expected value: 0, Actual value: 1024
Testing with sample rate [96000] and block size [1024] and sub-block size [32]
!!! Test 43 failed: NaNs found in buffer -- Expected value: 0, Actual value: 2048
FAILED!!  12 tests failed, out of a total of 45

[Lindon]

@Lindon okay so now its getting a bit silly....

I have 4 instances of a script node in my plugin, all compiled, all the same and if I activate 2 of them we pass, if I activate the third , we fail with this Nan silliness...

[Lindon]

@Lindon said in Mac FX failing pluginval.....:

@Lindon okay so now its getting a bit silly....

I have 4 instances of a script node in my plugin, all compiled, all the same and if I activate 2 of them we pass, if I activate the third , we fail with this Nan silliness...

...aaaand it turns out(by the look of things) that it is this third script node Hardcoded effect that is somehow corrupted....delete instance 3 - add a new instance 3 in its place and we pass pluginval...

...how in Gods name are we ever going to find this sort of thing beyond laboriously deleting and replacing every single scriptnode/hardcoded effect until it works...????

[DanH]

@Lindon I'm having to re-save a lot of my networks just to get the dll to compile. Have had to rebuild entire networks to get it over the line too. Am wondering whether there's issues with networks made before Hise 5 but compiling them in Hise 5.

[Lindon]

@DanH well Im not using HISE 5.

[ustk]

@Lindon what about setting the strictness level of pluginval to a lower value?
I don’t really know what it does though… I always keep mine at 10

[Lindon]

@ustk said in Mac FX failing pluginval.....:

@Lindon what about setting the strictness level of pluginval to a lower value?
I don’t really know what it does though… I always keep mine at 10

10, 7 , whatever same same...

[Christoph Hart]

@Lindon what if you put the entire network in a frame container?

[Lindon]

@Christoph-Hart said in Mac FX failing pluginval.....:

@Lindon what if you put the entire network in a frame container?

Thanks yes that may help, but as you may see above the issue seems resolved - well this issue anyway - with this product, somehow the 3rd Hardcoded FX was corrupted. Simply deleting it, and replacing it and loading the same script again made it work...

Of course this sort of breaks my mental model of how HISE is working at any level with Hardcoded FX.

i.e. If I load the same compiled script node into 4 different Hardcoded FX aren't they using and referencing the same "spot" in the DLL?, If so then if its broken(corrupt) for one HCFX shouldn't it be broken for all of them?

[Christoph Hart]

@Lindon said in Mac FX failing pluginval.....:

DLL?, If so then if its broken(corrupt) for one HCFX shouldn't it be broken for all of them?

Yes, but the error you describe is not a deterministic programming logic error but a case of undefined behaviour - something in there is not clearing the buffers correctly and this may or may not result in trash values being in there which is what pluginval complained about.

[Lindon]

@Christoph-Hart said in Mac FX failing pluginval.....:

@Lindon said in Mac FX failing pluginval.....:

DLL?, If so then if its broken(corrupt) for one HCFX shouldn't it be broken for all of them?

Yes, but the error you describe is not a deterministic programming logic error but a case of undefined behaviour - something in there is not clearing the buffers correctly and this may or may not result in trash values being in there which is what pluginval complained about.

Okay, well its all a bit academic and I dont want to monopolise your time, but that sorta doesnt jive with my understanding(which may well be where the fault lies)... these HCFX are simple filters, and there is no signal being sent thru them by pluginval, and they work fine in 3 cases (slots 1,2, & 4) but not a 4th(slot 3), where all cases are exactly the same "piece of code". One of them is generating "something unpleasant"(tm) which I might thing was the cumulative effect of the previous signal modifications, but then the 4th slot should accept this unpleasant output and reproduce it again, and it doesnt....

So bypass slot 3, but include slots 1,2, & 4 = passes
Bypass Slot 4, and include 1,2 & 3 = fail
bypass everything except 3 = fail
Bypass any combo that includes slot 3 = pass

well never mind this bit is working, more pressing issues(like the macro modulators not working at all...)