I'd argue a NAM capture is going to be more accurate than a model, you just don't get the flexibility of a model.like knowing that the real physics and interactions are all being captured.
Parametric NAM is a thing: https://www.neuralampmodeler.com/post/the-first-publicly-available-parametric-neural-amp-modelNAM coming on hard...but I still prefer modeling...I like knowing that the real physics and interactions are all being captured.
They need modeling augmented profiles or something like that.
It's hapenning!Parametric NAM is a thing: https://www.neuralampmodeler.com/post/the-first-publicly-available-parametric-neural-amp-model
I'd love to take a stab at it. I think Two Notes did it with their TSM-AI amps already:It's hapenning!
Who's going to be the first to deliver a full 6 knob JCM800 parametric NAM model?
Let the full amp capture race begin!
Parametric NAM is a thing: https://www.neuralampmodeler.com/post/the-first-publicly-available-parametric-neural-amp-model
It's just that, for now, it's not something too many folks have attempted. I expect we'll see more stuff like this in the next year or 2.
This is the only one Steve's published so far.Woah...didn't know that was a thing. That looks like a parametric OD model. Have they published any amp models? I'd like to try out!
Oh yeah! That's right - I forgot about those!Gigfast lite also has some parametric models - I've not tried yet, but people seem to be enjoying.
Ok, this sounds like something very useful for me and my weird amps:NAM just went another level above:
View attachment 31986
![]()
TONEZONE3000: Training Made Simple
NAM is the fastest, most accurate--and now (I'd argue), easiest--way to make a neural model of your gear.TONEZONE3000 is a third-party website that puts the standard snapshot training process in a streamlined, easy-to-use website. Like the GUI and Colab trainers I've developed, users can reamp a...www.neuralampmodeler.com
...And another thing that just occurred to me: Fuzz! Because this solution allows you to feed the device an actual guitar signal it should be possible to accurately capture proper fuzzes that rely on the guitar's impedance for gain control (the ones that don't work well with active pickups and buffers). I think this will be my first experiment.Ok, this sounds like something very useful for me and my weird amps:
"1. Imagine playing the model.
2. Imagine asking "but can it handle XYZ?" (E.g. cleaning up when you roll back the volume)
3. Do that so it gets to train on it and the answer will be "yes!"" - Steve Ack
I'm definitely going to give the Wet/Dry training a go as soon as I have the time. If it can accurately capture the interaction between guitar volume knob, cable and amp I will be very happy.
Now *this* will be interesting...And another thing that just occurred to me: Fuzz! Because this solution allows you to feed the device an actual guitar signal it should be possible to accurately capture proper fuzzes that rely on the guitar's impedance for gain control (the ones that don't work well with active pickups and buffers). I think this will be my first experiment.
16GB RAM standard is nice too.The M4 Mac Mini at the same price as the Nano Cortex looking mighty tempting for NAM.
I used it to train all of the reamps (111) in the Badlander pack with the LITE architecture.Did you guys try Tonezone3000 yet? I only briefly looked at it and didn’t sign up yet. Can you train more than 1 at a time? The way I worked around the slow speeds of the Colab page was by doing training in parallel. What are the speeds like for 700-100 Epochs?