You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Supermassive loads, and I can set its parameters without too much issue, but there's another issue: It produces a chorus-y, "whoosh" effect at the start of the output file that isn't present when using it from a DAW in the same manner.
Repro code:
from pedalboard import Pedalboard
import pedalboard as pb
vsm = pb.load_plugin("./VST3s/ValhallaSupermassive.vst3")
setattr(vsm, "mix", 70)
setattr(vsm, "delay_ms", '15\u202fms')
setattr(vsm, "delaywarp", 100)
setattr(vsm, "feedback", 30)
setattr(vsm, "density", 50)
setattr(vsm, "width", 100)
setattr(vsm, "lowcut", 10)
setattr(vsm, "highcut", 10000)
setattr(vsm, "modrate", 0.01)
setattr(vsm, "moddepth", 0)
setattr(vsm, "mode", "Gemini")
with pb.io.AudioFile("velkommen.wav") as f:
with pb.io.AudioFile('output.mp3', 'w', f.samplerate, f.num_channels) as o:
while f.tell() < f.frames:
chunk = f.read(int(f.samplerate / 8))
effected = Pedalboard([vsm])(chunk, f.samplerate, reset=False)
o.write(effected)
Attached here is a zipped .mp3 demonstrating the issue:
I'm not sure if adding silence to the start of each file to be processed will fix this, as I don't know if it's triggered by the start of sound or the start of the file.
Changing the sample rate divisor in the above example may change something about it, but not the time it takes to "stabilize"-- it sort of changes its initial stability, but not the magnitude of the effect. Might be placebo though.
Running the latest versions of Pedalboard (0.8.6), Python (3.12.0) and Supermassive (2.5.0). Windows 11.
The text was updated successfully, but these errors were encountered:
Supermassive loads, and I can set its parameters without too much issue, but there's another issue: It produces a chorus-y, "whoosh" effect at the start of the output file that isn't present when using it from a DAW in the same manner.
Repro code:
Attached here is a zipped .mp3 demonstrating the issue:
output_trimmed.zip
I'm not sure if adding silence to the start of each file to be processed will fix this, as I don't know if it's triggered by the start of sound or the start of the file.
Changing the sample rate divisor in the above example may change something about it, but not the time it takes to "stabilize"-- it sort of changes its initial stability, but not the magnitude of the effect. Might be placebo though.
Running the latest versions of Pedalboard (0.8.6), Python (3.12.0) and Supermassive (2.5.0). Windows 11.
The text was updated successfully, but these errors were encountered: