[Bug report] Decay rate are wrong

Hello,

when using Sigma Studio, all the elements which are using a Decay (compressors/peak, etc) seem to have the same bug:

the Decay rate is completely wrong.

To demonstrate the problem, I used the setup shown in this image:

What I did is simple: I use a RealTimeDisplay to monitor the output of a "Dynamics Processors>Envelop>Peak>Peak".

The input value of the Peak detector is initialy set to 1.0 and is changed to 0 at time==3000ms (as can be seen on the graph).

Here, the decay was set to 12 dB/s, so the value should have drop to 0.25 at time==4000ms.

The _profile_ of the decay is completely wrong: it is not a dB/seconds value, but rather something like a "linear gain"/second decrement.

Can we expect some feedback on this issue?

Parents
  • Hello,

    I think your discovery is correct, at least with regards to the envelope cells. However the compressors do this linear ramp before the dB conversion and the gain table so in effect the compressors are behaving correctly, but the envelopes are not.

    Here is an example of the compressors in action, where you can see the constant-dB release behavior:

    Based on some experimentation, I've found that the full-scale decay time in samples can be approximated like this:

    Decay (in samples) = 4.4*10^6 / x

    Where x = the setting in the control display.

    So, for a setting of 20, it will take about 220000 samples to decay.

    For a setting of 10, it will take about 4400000 samples to decay.

    I will check with the software team about getting constant-dB release versions of the envelope cells added in a later revision of SigmaStudio.

Reply
  • Hello,

    I think your discovery is correct, at least with regards to the envelope cells. However the compressors do this linear ramp before the dB conversion and the gain table so in effect the compressors are behaving correctly, but the envelopes are not.

    Here is an example of the compressors in action, where you can see the constant-dB release behavior:

    Based on some experimentation, I've found that the full-scale decay time in samples can be approximated like this:

    Decay (in samples) = 4.4*10^6 / x

    Where x = the setting in the control display.

    So, for a setting of 20, it will take about 220000 samples to decay.

    For a setting of 10, it will take about 4400000 samples to decay.

    I will check with the software team about getting constant-dB release versions of the envelope cells added in a later revision of SigmaStudio.

Children
No Data