@d-healey @Christoph-Hart Here's what I use:

// Get SkewFactor from MidPosition inline function getSkewFactor(start, midPosition, end) { return Math.log10(0.5) / Math.log10((midPosition - start) / (end - start)); } // Map from Skew inline function mapFromSkew(valueNormalized, skew, start, end) { return Math.pow(10, (Math.log10(valueNormalized) / (1/skew)) + Math.log10(end - start)) - start; }

@Christoph-Hart But who reads the docs? 😛

@Christoph-Hart Check at this, it appears that the glitch isn't from createPath but within the buffer itself.

Here I use getReadBuffer/getResizedBuffer with a sinus and you can see the spectral trace of the first harmonic as we should, plus the oscillo trace of the sine as a glitch at the same time. This means the buffer sometimes contains both sets of data so it is probably not clean as it should.

Another issue -> the data are spread only on the first half of the buffer (for instance the full spectrum goes from 0 to 128 in a 256 length resized buffer (or full read buffer), the remaining are zeros)

When using createPath, it isn't the sine that we see as a glitch, but it might be the imaginary part
The problem might not be a thread access issue after all...

HiseSnippet 2268.3ocyY0raajbDtojFJqYsMrALB1iSD1CzXkomg5GKua1Xp+nMgsjID0pcMLLLZMSSxFZloYlo4Jwrv.NuB41dKOAAHGyM+JjGibJ6aPRUcOCmYnHsnztqQDrAU+SU8WUc00WUTshDtr3XQDozRGMrOiT5lFsGFJ6sSOJOjzbWRoemQiFGYs8fNcXQV6FQOiG10Zcx1C6SiiYdjRkl+o3dKszBD0O+7S1l5SCcYYSQHGK3trWvC3xrYaU+4be+FTO1Q7fb6ds5McEg6H7EC.bMugMoO08TZW1ATbayYPJUdOOtTD0VRkrXRoE1V3MrcOwYg58eLOlehOCG3PZCJROcCguGhXbVxN839dsRs+XBnkVYdi40di6YrO2iOZ9LuxcTKXkIQd+Qo4JBu4K.Om7vyNG7l.jJkCRFZHcWi1tQ79xrUP77YFMCkrnNTvsmGJGLHXWdbee5P8MHfFiRyANYsRHy8Ol2XGAHZnrZ.8TViHXvHUUwols8JVaXae+u1D1VrvmU00mQip.SXB2RwRqefFY0hFx7cr9FqTc0kI2QDzWDBCprrd4kAYxDIFtccYfHp3MTfB.ssZ4JKqwoUiuWItoVSU8nRZ0STazFTgVWWPGUPXmWh9TYO6bnzMhA2PsfYQ6Ih00B0oMbNohEyjsf6b4ghARdHqRmAgtRtHrR26a9ilK0E2fNRsh9i3phHZXWFnOXUO38hR8xd73bXXEqWC+G9mZd.2eG2CAQ1DOiw61SV49uA1V00QC4cn8mGX3qlncn99m.OOxPlBXO7gJKAr0rSNwegZ+PVL+Oy7RbSN10VaEKGDxWhTzTQv8ZtzXVUVnwEVIVRijGvNq8fST9C6pn8W0VqmNhHqJ7uw9qs3+AEBp5yB6J6Ai+xu79lKAFzEznObcbjnB+g4E.rhGnF+Z9aPU+tQnLh0GuGqj3HS8hHrT9wJ0TA4Y+j5OsDgGHjrWl3XMemo03K0oyDWCixhD99n6ZBKqO1OhfUBGDbBKZE34h+.1nMBIpJldwX5oWxm8yUGzmaihvlgb4K6yBmVNQRxKE3291l6BNeLmTxbv95yhjbDBk1k8CPBdcFpkL1kEepTzWs2jj.PNaoZ0akl+BuAHbHK0MLz2FjyyyALL+fyvmGilv5uVum54wnYd+eoNIWpxEz.ARRK7F3SkEybizUIK.96BoKwLegwb4v7zY+pkNeVg3cMZwkt8lLFmaBXDtU9s.iIjf2xXO3MuqLCfKXz36+Dv3UVe9lIQLvYhG9MMx3DFmq6HJbJJNtRfoboje+q4LxwHExjPlu8B6BYVRnF1Mt+AL4YhnSqrbrRn214bESDrYLGbCQzYzHujWrwGIZQifpT.9y3Jcn9wrh7jGts83GwVgdGx5xiAQNDJsJGyUdwblMwbJdbACa2m4xo9YuUA87ilVVKqE3EprlK+UVa573Zq7vGtZsGswlqfq+c7POwY3E.r5xa6CTLAzvvmQih3wKq1xtfpOg4eHR2Aa50vbVVOvwFysq9c3WfOeiZyeaLqEidJHCcHrYkuIcgDE01k5iJRFMPuzqdJMHfBy3n03xoRCLhox9BQWZDW1Kf61Hh8m15b.cZUX9NvW.Nb7ZJyEk4IpLAuC39.e8UThwIL5ibMQ.4TKHxqRLMnuO6P3JaEqS7Etm1FXduXV+95v9swcTwsG3pY9wWGxgxy7KQmOdh.cIOc2mJi3PlYC30jtnrcRPW5qojw13XLWTaVnmZv+E9IYQmz2j3hNoKlQuTpbhcAvQljYT+JEXPNQrs3bEUwcLNIWb6NmPTVeVhvOuNYlzysMNaT78E0xctnVtchVZ6y8XQJcbSCUgduP77vKngW8gqfFdFeBZHg68RzvmYneiLIELSFwhFd3apBB+29oe5e+DRRdOLm7bI+NbsoS.mkJjbLjoCYn.MAO2qZS1x2WbFR6yS3tfLyp4ZI7G1umHj6hSo2QZrzVAhAXMB5.J3.Ed.ZLZPcgP4gXEiHqHVACPAwhp5NhJp.XxXCvXZTIocQpiLgCMPUKB1C4735p99RNMs5DwtPWoHgL4BGeDqZtkuZmFZv9rySKhZLhowmAF2LzicNx09.nxrLZE7sxnQvQsOO7X8kmAdCP1md9nwqAiaKY8wTN3XHOpxoUFq5fQJHGImRKmozxF0Jp0kfIrwb74z78.M3Xq+wYsGa63rAjb4T1YZOndC0p83Gau15aTq1pqs4i1b8j9Vwr1CXgtCSgycMbrWc8pah5ZcmU2XiBPqf85T.YkMb1XJF7MUGi0gXDIof3jY0W5LUe4SgLsSWoKbYJEUxhFs5Qi+H2HyjVJqJscbnjUlNgjKR2vnSGYwP7aYPCo9CiYUwk9zDdqHexGcOU.WFAry+OfXmKGwk02ppFhKB4aZD.epaUdLDm+6m5RP+bERHT5pGuomZr3s71wTsPUI53HnAj4zOtZFeLtJT+VpFg73aKDPIiJFDRQBjxFPoIxeanNJM6TGIv3ZQZ7K28bs5mtbt9ouwUoNobsV+E4as9KpGPwh6JUFFvkr.3v+8F3WHjYM601zbM6GugI1bfoyFqt4ZlpFDHEaG+8ud71wI+Qr3iYBoiUI1zv4Ao3bwLbV23PnqTnBJel4XMnX9LnpHn3U3yvPymS4wrHyih35M2.BuNBbmWQy31yVof4LAddSnY8.tmmOqk.ZdWEcmdvG8g3QDKoUf8OeRP9uT8W8Ak8Ox27gX39UUVtgk21j.UPOFnn5j5xsrl0krykXu0Gt0AOcOqW7RhGqCcfu73ITI6HWvslsZYmlKfPlhKfT6RcAu+3ht.B4WYWvyZRxapKMSEcmCPCGyRyC15Im0RFux5oas+9aMEuMYZwHjeNmyP4axi0Euz56mFNaVOCmu+d+8+ySl.tT8Cnw+hF6t2Na8pOMMGFPciDu0M4qVAxTdC0LPR3P0eHnkL1GGa4j18hlFC7f7255lV47jko10PlUuFxr10Pl0uFxrw0PlGcMjYyOpL3eSrsFHEAZ5aXhV6oX4JUZuP7qlS88zQ9eP5i+5d

Screen Recording 2022-05-16 at 01.50.36.gif

Below I generated two spikes (20Hz & 20kHz) + noise to demonstrate the buffer is going up to SR instead of 20kHz. Maybe a remaining of the real + imaginary part that necessitates doubling the Nyquist length? That isn't as bad as the glitch though because it is easy to limit the drawing to 20kHz...

SR = 48kHz

Screen Recording 2022-05-16 at 14.43.02.gif

SR = 96kHz

Screen Recording 2022-05-16 at 14.48.55.gif

@Christoph-Hart Any insight on this ? I really need the spectrum to work 🙂

Yes, nice video that outlines the process pretty well.

You don't have to do the "wrap into network" step though, you can just create a network directly and set its AllowCompilation flag.

@d-healey A server system from juce can be used for that kind of communication but it isn't easy to implement for what I've seen
I reckon it's InterprocessConnection with namedPipe

@Christoph-Hart it would be great to have this fixed!

@ustk It worked for me once, but I never used it since then 🙂

@d-healey yes

@d-healey Thank you so much for the help. 🙂

@Christoph-Hart Thank you!

@Christoph-Hart great!!

Should be good now.

@JL-LV This is one of the rare cases that actually require per sample modulation:

ed336c98-be86-46a7-b54e-4a88e7aacf39-image.png

Also, you definitely need to create a C++ class from it, the CPU usage will drop to about 25% of the interpreted graph.

@harris-rosendahl I just took a look and the JUCE oversampler actually only supports 16x:

https://github.com/juce-framework/JUCE/blob/2f980209cc4091a4490bb1bafc5d530f16834e58/modules/juce_dsp/processors/juce_Oversampling.cpp#L547

The exponent need to be smaller than 5, so 2^4=16x is all you can get here.

@Christoph-Hart - cool... well done...how does Daves fix for dynamic assignment of modualtors seem? I would really like that in there too... well actually I need that in there really...hint hint.. and he's already done the work... smiley face, sugar on top.. etc. etc. <fawning complements assumed here>......

@callybeat The problem I had was with the knob regulating the intensity of the LFO that in the module goes from 0 to 1. It worked fine inside hise but not when I exported it. Doing it as the Parce from Colombia suggested, I got it 🙂
I'm exploring script fx and for effects plugin projects I think it's the best.
Thank you very much
Saludos

@orange Thanks man! Appreciate.