[BUG]

[kourkoumpini]

Before you make a new bug report:

• Have you searched the Issues to see if your problem has already been reported? If it has, please follow that Issue instead of opening a duplicate!
• Have you checked the Rough edges for help?

Description

DI sound is audible along with profile sound.

Tested this on Cubase and Reaper, on different PCs.
I'm not sure if this only occurs or some profiles or it is a general issue and it just so happens that it is more audible on some profiles than others. Maybe it has something to do with the sound of the profile and some profiles mask the DI better than others.
With the profile in the link https://www.tone3000.com/tones/mesa-cab-28749 it is pretty apparent, if you hit the strings hard.

Also setting the input gain in the plugin to -20, does not mute the DI sounds completely. This leads me to believe that there is some sort of blending happening, when I increase the input gain in the plugin, and it is not just adding "gain" to the plugin. Hence on 0 the profile sound and the DI are both audible.
Regardless, when the input is on 0, you can easily hear the DI sound.
Also keep in mind this is not audible when previewing the profile on the tone3000 page, so I'd say it is not a profile issue. Again not entirely sure.

I also posted a question on the unofficial subreddit and another person verified the issue as well.

To reproduce

Steps to reproduce the behavior:

1. Go to https://www.tone3000.com/tones/mesa-cab-28749)
2. Download the profile
3. Load it on NAM
4. Every setting on default.
5. You can hear the clean DI sound along with the profile.
6. Setting the input gain on -20, you no longer hear the profile sound but you can hear the clean DI.

Expected behavior

I guess only the profile sound should be audible.

Computer & other info

• OS: Windows 10 and Windows 11
• Plugin version: Tested on multiple versions, for sure 0.7.12 and 0.7.13. Also tested it with Universal NAM loader if that makes any difference.
• VST3
• DAW: Cubase and Reaper
• Audio interface: tested on Focusrite Scarlett Solo 2i2 2nd Gen, and Scarlett Solo 2i2 3rd Gen.
• Sample rate: 44.1 kHz
• Buffer size: various.
• Provide a model that the bug happens with: Try https://www.tone3000.com/tones/mesa-cab-28749

[sdatkinson]

Thanks for the bug report.

Can you create a dataset with a train/test split where:

1. The test split is recorded in the same sweep as the training, but is held out from training,
2. Verify that the test split doesn't have the DI leak, and
3. Verify that the trained model's predictions do?

Then share the input/output audio as well as the .nam file and the rendered predictions that show the problem.

I'm willing to look into it, but I think it's highly unlikely that this is a problem with the plugin, so getting the above would be very helpful.