Dry/wet knob not working

[michaelgregorius]

The dry/wet knob on the "Impulse Responses Mono" plugin is not working as expected. I'd like to use the plugin with impulse responses of guitar speakers. Therefore I need to be able to fully disable the amp signal and to only use the processed wet signal. However, if I set the wet level to 100 % I get doubling effects.

Here are some further steps to reproduce that there's in fact a problem:

1. Set the "Dry" knob to -inf dbFS.
2. Set the "Dry/Wet" knob to 0%, i.e. fully dry.

Expected result: Nothing can be heard.
Actual result: the dry signal can still be heard.

I think this can the traced to the following lines of code: https://github.com/lsp-plugins/lsp-plugins-impulse-responses/blob/af5066bf43ae585f8d459a41d7b0fde34ca84e88/src/main/plug/impulse_responses.cpp#L475-L476

The current code looks as follows:

c->fDryGain         = (dry * drywet + 1.0f - drywet) * fGain;
c->fWetGain         = wet * drywet * fGain;

It needs to be adjusted as follows to correctly interpolate between dry and wet:

c->fDryGain         = (dry * (1.0f - drywet)) * fGain;
c->fWetGain         = wet * drywet * fGain;

[sadko4u]

Hello!
You're probably not properly understanding how 'Dry/Wet' knob actually works. Here it the chart:

[michaelgregorius]

Thanks @sadko4u!

So if I understand correctly I can only get a full wet signal if I turn "Dry" down to -inf dbFS and put "Dry/Wet" at 100%.

Out of curiosity: what's the intention of the design above? To me it seems that I could achieve the same results with a simply dry/wet knob as it is just mixing the dry and the wet signal in a more complicated way but still only via attenuation and summing.

I'd also propose to update the documentation with a link to the diagram above. Reading the documentation again I noticed that it only gives a rather subtle hint that "Dry/Wet" is not simply mixing the dry and wet signals but a dry and a mixed signal where it is not clear what the mixed signal is.

From a GUI point of view it's also not very intuitive as the controls just look like individual volume/gain knob for the dry and wet signal which would then get mixed via normal interpolation. It's not really clear that the complex mixing from the diagram is applied.

Thanks for providing all these plugins by the way!

[sadko4u]

Hello!
The idea was that we should handle both cases with at least possible amount of controls:

• the first case, which is mostly useful for reverb users: control Dry/Wet balance without affecting the level of dry signal (e.g. mix processed (reverberated) signal to the unprocessed one). In this case two separate Dry and Wet knobs work the best way.
• the second one, when we apply some sort of dynamics processing, we want to blend the dry (unprocessed) signal with wet one. In this case a single Dry/Wet knob that controls the balance between Dry and Wet signal works the best.

Anyway, yes, I think I need to add this chart to the documentation. The only one question is to find the right place to not to duplicate the manual across all plugins.

[michaelgregorius]

Thanks for the feedback! I can completely understand the potential need for both modes.

I'd like to give some feedback from the perspective of an end user who uses the software. My feedback mostly boils down to the fact that when working with the GUI, i.e. the three knobs, one must know the implementation details, i.e. the already somewhat complex diagram from above, to achieve the results one is looking for.

I'd like to propose an alternative presentation which would remove the need to show the diagram in the documentation in the first place. I'd also like to show how it would remove some surprising elements of the current implementation.

Proposal for an alternative GUI/presentation

As an end user I would find the following implementation easier to understand:

• Provide the user with a combo box that lets them switch between the two cases or modes. Let's call the options "Dry and wet leveling" and "Dry/Wet mixing" for now.
• If "Dry and wet leveling" is shown then only the "Dry" and the "Wet" knobs are shown and the "Dry/Wet" knob is hidden.
• If "Dry/Wet mixing" is selected then only the "Dry/Wet" knob is shown and the "Dry" and "Wet" knobs are hidden.

Behind the scenes it could still use the formula from above but it would just set the values of the hidden knobs accordingly to make it work.

You will find that this will exactly hide the surprising elements of the implementation that I will describe in the following.

Surprising elements of the current implementation

For me as an end user the current implementation has some surprising and rather non-intuitive behaviors/elements if I want to achieve the first or second case.

First case (leveling of dry and wet signal)

For the first case the following settings have to be made:

• Adjust "Dry" and "Wet" as needed
• Set "Dry/Wet" to 100%

Here it is surprising that I need to set the "Dry/Wet" to 100%. As a user I would think that I can use the "Dry" and "Wet" knobs to affect the individual channels and that they would then need to be mixed in a balanced fashion, i.e. I would expect that the "Dry/Wet" needs to be set to 50 % instead of 100%. Needing to set it to 100% gives the impression that the dry signal will be completely omitted.

The "Dry/Wet" button is hidden in my proposal so that as an end user I do not need to have that special knowledge anymore.

Second case (Dry/Wet mixing with a single knob):

For the second case the following settings have to be made:

• Set "Dry" to -inf dbFS
• Set "Wet" to 0 dbFS
• Set "Dry/Wet" according to taste

The surprising element of the second mode is that I have to set "Dry" to -inf dbFS. This again gives the impression that the dry channel will be completely omitted from the mix.

The "Dry" and the "Wet" buttons are hidden in my proposal so that as an end user I do not need to have that special knowledge about the "Dry" knob anymore.