NAM: Neural Amp Modeler
- Thread starter James Freeman
- Start date
Humbug
Roadie
- Messages
- 165
Searching T3000 for "TTSV10" got this result: https://www.tone3000.com/search?q=TTSV10
I'll assume that most captures using this input also uses a custom architecture, so maybe you'll find some more if you select that option in the filter box :)
DLC86
Roadie
- Messages
- 925
I've shared a couple pedal captures here:
TTS Cornish G2 (calibrated + anti-aliasing) NAM Profile · TONE3000
Check out thetonescientist's neural pedal capture for guitar, bass and recording: Clone of the Pete Cornish G-2 based on the schematic traced by Aion FX It has four D9E diodes with a Vf comprised between 0.21 and 0.23 2 files - "standard" was captured with v3.0.0 input file and standard...
www.tone3000.com
Diezel VH4 Pedal (calibration + anti-aliasing) NAM Profile · TONE3000
Check out thetonescientist's neural pedal capture for guitar, bass and recording: Diezel VH4 Pedal · djent, grunge, metal, rock, breakup, crunch, distortion, high-gain, saturated, analogue, guitar, pedal, preamp-only, solid-state and nam
www.tone3000.com
But I think there are other out there, try to search "aliasing" on tone3000
ArteraDSP
Groupie
- Messages
- 51
Artera DSP celebrates its first anniversary.
For a limited time, GigFast Lite and NAM xT are 75% off — until the end of July.
Both apps run on iOS and macOS, with full AUv3 support for seamless integration into your mobile or desktop setup. Windows and VST compatibility is also coming soon.
Available now on the App Store.
Thank you!
2dor
Shredder
- Messages
- 1,300
Steve's diving into NAM as his full-time gig:
www.neuralampmodeler.com
I'm now working on NAM full-time.
When I started NAM in 2018, the project was meant as a sort of "crossword puzzle" challenge for myself. I'd been picking up experience in machine learning and I wanted to see if I could successfully solve this application problem that was near and dear to me as an electric guitarist.What I...
www.neuralampmodeler.com
MirrorProfiles
Rock Star
- Messages
- 4,298
Just a quick bump/shoutout for how awesome the automatic level compensation is. I totally understand why some people don't really get why its helpful but here's an example that came up today.
Band I'm working with wants to reamp guitars. I'm making a NAM model of my Maxon OD820, but it's 18V and produces a really loud signal that easily clips my usual 11.4dBu input chain.
I've increased the headroom to 20.4dBu to capture it as a NAM model, and then I can use that model on my 11.4dBu input like normal, and the plugin just automatically corrects things for me, so its driving the model afterwards just as the real pedal would. To do this manually or with something like ToneX would be a massive faff and would lead me getting a different tone to what we're going for.
Band I'm working with wants to reamp guitars. I'm making a NAM model of my Maxon OD820, but it's 18V and produces a really loud signal that easily clips my usual 11.4dBu input chain.
I've increased the headroom to 20.4dBu to capture it as a NAM model, and then I can use that model on my 11.4dBu input like normal, and the plugin just automatically corrects things for me, so its driving the model afterwards just as the real pedal would. To do this manually or with something like ToneX would be a massive faff and would lead me getting a different tone to what we're going for.
Mongillo19
Rock Star
- Messages
- 6,149
I can't wait for someone to make a NAM player that has this ability in it. I'll probably snag one if/when it happens
MirrorProfiles
Rock Star
- Messages
- 4,298
There's no reason for any of the existing ones not to add it, really such an essential quality of life feature IMO.
paisleywookiee
Rock Star
- Messages
- 7,168
How well do the captures utilizing the Dry DI+amp recording work? Seems like a really easy way to make NAM captures on Tone3000
It depends on what the dry DI contains.
The super inputs that minimize aliasing use the wet-dry method, which produces the superior results.
The standard input file already has sine sweeps and a lot of varied playing/noise.
2dor
Shredder
- Messages
- 1,300
The more data / playing you put in them, the better.
Most important is to have them sample-aligned - that's going to lead to the best possible profile.
MirrorProfiles
Rock Star
- Messages
- 4,298
I’ve been retraining a load of old NAM models as xSTD models. Of the many things that are awesome about this (besides the improved accuracy, tone and feel) are:
- I can reuse the existing reamped files. I don’t have to recall settings and reamp everything one by one. I simply wouldn’t do it as the time involved would drive me nuts.
- I can upload these reamped files all to the cloud and they can train simultaneously, WAY faster than locally.
- Very easy to A/B the different training architecture against the standard one because the source files are the same
IMO it’s so much more robust of a way to catalog tones. I keep thinking that it would make sense for me to be making more ToneX packs (because honestly, they sell significantly more). But my real goal is making shit that sounds as accurate as possible and it’s just so much easier and more convenient to make NAM models. The QOL upgrades from Tone3000 have really made things so user friendly now, very excited to see them add support for Hyper Accuracy.
- I can reuse the existing reamped files. I don’t have to recall settings and reamp everything one by one. I simply wouldn’t do it as the time involved would drive me nuts.
- I can upload these reamped files all to the cloud and they can train simultaneously, WAY faster than locally.
- Very easy to A/B the different training architecture against the standard one because the source files are the same
IMO it’s so much more robust of a way to catalog tones. I keep thinking that it would make sense for me to be making more ToneX packs (because honestly, they sell significantly more). But my real goal is making shit that sounds as accurate as possible and it’s just so much easier and more convenient to make NAM models. The QOL upgrades from Tone3000 have really made things so user friendly now, very excited to see them add support for Hyper Accuracy.
I’ve been using the TTSV10 provided since the TONE3000 wet-dry volume issue was fixed. I’ve also used xSTD architecture to sample my gear. I think it’s nice to have all of these tweaks, especially since my goal is to replicate my own gear the best I can.
I thought the recent exploration of loads in recent packs by @2dor were interesting. I was inspired to experiment with a speaker load vs a Suhr Reactive Load on my Vox AC15. It was enlightening to hear the difference through NAM.
Aliasing seems to be a topic that is too inconsequential for some, but I thought that it was neat to stack 3 NAM instances (2 pedals + amp) and see minimal aliasing through the TTS Spectrogram Generator. I’ve been getting a lot of mileage out of the tools that @DLC86 has provided.
I’ll probably upload these experiments to TONE3000 at some point.
I thought the recent exploration of loads in recent packs by @2dor were interesting. I was inspired to experiment with a speaker load vs a Suhr Reactive Load on my Vox AC15. It was enlightening to hear the difference through NAM.
Aliasing seems to be a topic that is too inconsequential for some, but I thought that it was neat to stack 3 NAM instances (2 pedals + amp) and see minimal aliasing through the TTS Spectrogram Generator. I’ve been getting a lot of mileage out of the tools that @DLC86 has provided.
I’ll probably upload these experiments to TONE3000 at some point.
Last edited:
