Trying to create a random pitch detune effect in HISE

LeeC

@ustk Ok great, I'll give this a try later today when I'm back at the machine.

Big thanks again for putting something together

LeeC

Beautiful solution @ustk. Just gave it a try

Like how you've created randomness by combining different oscillators with different settings. Sounds ace

It's also helped me to see some of the possibilities available within script node.

Big thanks!

LeeC

Hey @ustk,

Are you finding scriptnode to be incredibly CPU hungry when doing certain things?

For example, the nice pitch detune snippet you created below uses 50-60% looking at the Hise meter on a Mac Pro.

HiseSnippet 2211.3oc6a0DaabbEdVRMxhzxJ+TkB2hdfHPnPFvQfjV1oAAshV+YHTSKZQYkDfz5NZ2gjC3xY1r6PKyDTfbnG5sdt8ROzBzdJnWyghHfdsAH.9Pt1VfhdHM.5btjLyN6xcVxkRqXhjrDzdQb9+aduu26M6rOUykYh87Xt.iba2yACLlFVuGk2ZkVHBErwp.iYfaQsVEy6RwqVuFX4dNHOOrEvvH68j8wH2D.+mCVZYjMhZhipB.1gQLw2mzgvipsVkeNw1dcjEdaRGsduXkMLYzUX1rtB7jEVD3fLaiZhe.R1sLPfwjqYQ3L25bDG6ALlXYlUu5sX6QU8eGhGYWarrPIPcwDopdclskDwxZAqzhXaUKbe6A.FvZQRgrJovrvpDKR+5ijFuneCEhFgt7vHygAuR5vqX5gmgF7lPAuWBV2zk3viZQhsqB2fxwtMPBUfNrT8EXrObElnCT9BcPswq6JJze.yemhEuYgaWr3MdyFcolbBiVfQe.ii2jN+Mx+A4yk+WmuvfM0nQhsIWFWlsM1MwlkZc2CafyS61YWr6MK7DjcWb+NJ19wkuvQKe0U+lpcsVGYzMnD9lNX5nj5f.QkT9GfJQW49R+YBj90sIVXW.QHjeAnxBovc6v5JFkOvCY123c+eKAdzFqh3nvoSLyhUyA6xIxMhwp3mHrSTJ2bh4xqMm4322NNLpDGFSl5E+o8Mo9U3J85W3C+kUbB2gaX4ao6OSEV+sKIrzbEFYBxfrkWNZBcrI7tVXgQ6PjPgo.ypqMhG29P5THnAgRMFQTx1ndDdOcmFmHFMoEtuDrFga1JY7lIA7JTfmz3MvEz0fq0nA1jGA1Ifq+1mM9axGPUDquDHwXNI4o4ifqQaRn3ELcwBnspmyCv78Xtsm+Uc8ClHp4Uuwads78s8cbwBJHdaVMaTu48PBhGdKwPuYgcsYlsqSde7vNLBnyKK6w7lsPTJ11ab7qLPXkroyuRBpwsXc4DZypHtKQXFBeP2N0EQyLwqDfNQcFYjtWTkKJKKYU0wTK+Bes3InwRxxFAMVJrwHOSf.gpuRK32heoBEzWNKaiYIUhqiLEaxd0P7VRluzsnfZgcWvrOEa5nwUBH.+fv99BIcrpMxJm7vX1p8hvoUGeWVhH1A+9sHVxEEZbSQOTLlsD5XYjfpd5m.HFaRnApE5WRtVYTqlZ+5umthBz64hb7jtpJAhsIut1lrgbhJ+Xe5zI1VyHSR.7JPevAFoBP07IFn5K0xn.0TJPoT6GAq3rfDLDf8kh4gcXVIf4Y0vb+dbxKKmPCZSASh8cVngCPUNkcAyyTbRaYHugQlKdAslOKTygwgDAxxFEAdajaSLOHlX3IjTQkjkDwhyJ2b4fOxCKOKNXG0YsLxVq6t1DS4u.ZtMD8ueIofg9DwrnV1pD5N5GTS9TE8zApapJ2msG1cfWhA.djiyP0NUk5briLZk9Qbp2FumR9qWqbWLo7PIXvPn.jFH+G98xmOdogg7bIB4Eqj.j+sOKFjCmScHmGVbgxkei2n3sCdMi0cwuWWL0rWDvW7VxmmsDH8x5CR.3ESD3GrTRx5hGSY8z93tvVRVFXHr.zLbtllgSoKsbtzx4nsblTXiTTb3vioYyGL6Gcu+86++u3Y1T9RylKMaNwLa9yN+z7e9e5yt3Y1bqKMatzr4Dyr48jVMN+2ywlM57kX7NcCpIgMG5kDupxRp4Y0qzF6Z6NZqnLCXEMMbKrGlqDHCwB0MnhOtWPZ8k1gpaKlI8Tj2Y+goHRI5vTj2Y+jnH.vQ8BH4l3mY92+zl+sWohR85+cm5uh+0q+O9pewOdg8AeqbeT42kd2Gh9dzP9q68Gq88V8YKorFq2gw3sHzlQ39S+mxm2sR.0MuxaoGoY4NLq3r2YfcD+cgv1NSCBjIETYcljw2Nu5GrT58pKbrnoVTms92bHNatBbDVEI4ZYJE2yAiZmnqE+FNatsrDzAFYFrRgSD45J+9N4fVXaTuGihzTxpuBbUY0fMc5eq+g9MVih10N7ihbbUgIGkatDcFLWLKqWakezW7etyObDpv0dpiK1yS9UajxMvgriyGtiKcwZKO53gGUjx7p6szzFiF7NNlV4swuoK80b55q4nTaSEn1jWxcomuu6d8u.xL9ehl91fiD2c5ZyImBecF8ODRj2vXv5ZPOd6EBaZ.HYbl4XezD97eW4lJYB+bIbRd.3vH7JWqQK4pe0W9E6+wexvtoEZ3QQ00bbeo144CsSbWTiQ.nQNAwDXSJgYBYhRZckmd2y6LXx53miNTre1C3EjxNAE0NSQvqYF67DAuahn+9Ja+2KeCKAsRlDIIbHii+aVM29em9lUvauvhkKW90ACcxBcM13j7R5ozxrpFuJ7sPOA6meg9WRwq3WtAysSg6gEt+UeukQmggeVZyvPmTmggaZxEPXaWD0yg4gKowK9jAZqr9ppnkkFtpX8Rt8VWr8z52KVIrxx5WGUMDM1jIJGalpRhxpKv+ZI0gSqiE6FqMitcPMRcst1dXemYkzGXT0k0q1HapyjnjSrxhm+RrxSuTZKa5f6QjKV4BwXchL6nVS35vVXEBB8S1.IN7TXswMipxnLmVLp7ldh3Tag4tjlMwwCNjzF5tbNxrcTMyVYKr3sC7hGrRbBNjaU08vMNxhRoWVbX5qqCUvsfz1uvEiTQL6EnTQ7X3.IA5+KCqwr6oHyqSrkwXUGhWU3vBe..Q2Itpx8u+U+KUdXPgJx7WHN885UdXWjslT+COX56tLwgYibSkFY7Y6SjRZy2O5h9OOw0kZlyO38G.CT5E7+5Dmav8LvGd9ArxS.7zmCf6oQN7dZrFcPltrGapR8YoKko7qQruo9++zjCVUVtPov+aEfxupHni3XXO1zT5i+0DxmjGS4wXL2ZLFyhiwXt8XLl6LFi40Giw7SNzwHOA7c6xYcT7eQE0Vy+EuLLhtM5rfuAIizt4

Anybody else finding the same thing?

Cheers

ustk

@LeeC Mine stays at 2.50% max with this snippet... (i7 3.5GHz)

Although the CPU charge in Hise doesn't reflect what we are generally used to... I invite you to search about that in the forum, @Christoph-Hart gave some explanations to defend Hise's measuring system. (quickly, 50% doesn't mean the CPU is half used...)

LeeC

@ustk Just had a search around thanks.

Wow, 2.50% max with the same snippet :astonished_face:

Like I said mine shows around 60% easily and if I duplicate the 'ScriptFX' I start to hear glitching and dropouts in the audio.

Appreciate that we aren't exactly comparing apples to apples here with the different systems etc but, I thought that my Mac Pro would be apple to handle a ton of instances of the ScriptFx in your snippet.

If anyone experiences similar or has any ideas what's going on please share.

Cheers

alt text

LeeC

Duplicated the ScriptFX a few times and we're up around 250%

alt text

Christoph Hart

Quick question: do you use the Debug build of HISE? That might explain the huge CPU difference.

I've loaded the snippet, takes 3,5% on my CPU, but it's a fairly powerful i9 notebook.

As long as your MacPro is >2010, there's clearly something wrong.

Also be aware that scriptnode patches which use frame-based processing are supposed to be compiled into a C++ module for production builds - if you compile it to C++ that particular patch shouldn't take more than 0.5% - even in the HISE CPU meter.

LeeC

Hey @Christoph-Hart,

Yes I'm using a debug version of HISE 2.1.0
alt text

alt text

Guess it's time for me to get latest.
Could you point to the right/best version to be using?

Appreciate the response

Christoph Hart

This post is deleted!

Christoph Hart

Ahh, never use the Debug build and complain about CPU :)

It contains tons of boilerplate code that is supposed to help during debugging. Just recompile HISE with the "Build for Profiling" menu command and it will optimise HISE to run about 50x - 80x faster (expect it to take about 3x longer to compile though).

LeeC

@Christoph-Hart great, thanks for testing!

Ok, so I first need to get off the debug version I'm using and then I need to learn how to compile scriptnode patches into C++ modules.

Thanks for steering me in the right direction (I've been spinning in circles all day on this one).

Can somebody please point me to the latest version of HISE to be using?

Not sure how I ended up on debug anyway to be honest but It's been working fine so avoided updating until now.

Cheers

Christoph Hart

You don't need to get a different code, just open the HISE XCode project again that you've compiled to the Debug version and tell Xcode to compile it in a different setting (Product -> Build -> Build for Profiling)

LeeC

@Christoph-Hart nice.

Thanks again

LeeC

Yep, recompiled with "Build fo Profiling" and the CPU performance is night and day!
Bye bye debug mode

BIG thanks for the heads up @Christoph-Hart and thanks @ustk for testing earlier

Trying to create a random pitch detune effect in HISE

16

1.7k

11.8k

103.0k