NAM: Neural Amp Modeler

Few technologies in audio are used in more places than NAM - the only that I can think of is impulse responses "

I think you completely misunderstood his point and went WAY off the rails. The way I read it, it was a 100% technical statement about the challenges of updating a product that is used on so many different hardware platforms in so many different software environments.

Like the difference between updating an app that just runs on your own dedicated hardware, vs updating an Android app that runs on dozens of different phones, tablets, etc.
 
Like the difference between updating an app that just runs on your own dedicated hardware, vs updating an Android app that runs on dozens of different phones, tablets, etc.

This is a long-solved problem on Android. Apps are built (and distributed) for all its supported platforms - including some truly ancient shit.

NAM being a very moving target currently is IMHO the main reason it hasn't gained widespread adoption yet. There's no guarantee that captures made in the future will run on current devices, even if software support is there.
 
Look. We'll just have to agree to disagree. I think his words and the intent of his words speak for themselves.

How good and great a guy he is - I have less than zero doubts.

How great NAM is I have less than zero doubts. How much he is committed to improving NAM I have less than zero doubts.

My comment was only directed at his - imho - patently false claim - as I pointed out with examples in my response.

As I said - I think his words and the intent of his words speak for themselves - as do mine.
 
1768787337494.png
 
More power to Steve ... period

I would however respectfully suggest a healthy reality check is needed or people need to stop pissing in his pocket:-


" A2 will be industry-informed Few technologies in audio are used in more places than NAM - the only that I can think of is impulse responses "

:facepalm

Just off the top of my head

- DAW's
- Midi
- Midi Instruments
- Ethernet Audio
- Dante
- Madi
- Samples
- Plugins all formats
- Software / Hardware Component Modelers
- FRFR Systems
- Kemper Profiles
- Tonex Capture
- QC Profiles
- ASIO Drivers
- CoreAudio

The list goes on. Dude needs to get out more.
Only a few of the things you named are "technologies". Many of them are well used, but far less common (MADI & DANTE are cool, but most smaller musicians will never work with it). Also I think that he's implicitly talking about a subset of technology that is meant for audio processing (which eliminates MIDI too as there's no audio involved - even if, he's not saying there was NO other technology, just few).

Say what you want, but I have to think hard to find a technology that inspired that many new products in the last 2 centuries. Maybe Modelling in general, but that started even earlier.
Profiling was influential, but only for few companies that came up with good algorithms which they kept to themselves. The nature of his project being open source has made it what it is. So if we don't give him any money for that anyways, maybe we should acknowledge what it has brought to musicians (mainly guitar players to be honest).
 
“When A2 launches, we'll retrain every single NAM profile uploaded to TONE3000 to support the new architecture. Hundreds of thousands of tones, from vintage Fender amps to modern high-gain rigs, classic bass preamps to boutique pedals, will be available as A2 models from day one.”

I can’t state enough how amazing that is. Making a massive update to the entire platform while also NOT rendering all previous efforts as v1 or “old shit”, 10/10. It’s almost like there was a thought process from the beginning there, well played to t3k.

Italian Kid GIF by Gogglebox Australia


Frustrated World Cup GIF
 
“When A2 launches, we'll retrain every single NAM profile uploaded to TONE3000 to support the new architecture. Hundreds of thousands of tones, from vintage Fender amps to modern high-gain rigs, classic bass preamps to boutique pedals, will be available as A2 models from day one.”

That's really cool for free stuff - but who's paying for all those compute hours to re-train?

And why bother training every single NAM profile - there are sure to be some real clunkers in that bunch. I'd just choose something like top 200 profiles (or some other number - don't know how many total profiles are on tone3000)
 
That's really cool for free stuff - but who's paying for all those compute hours to re-train?

And why bother training every single NAM profile - there are sure to be some real clunkers in that bunch. I'd just choose something like top 200 profiles (or some other number - don't know how many total profiles are on tone3000)
Yeah the retraining would be monstrous, maybe they have some nifty way of streamlining things. Who’s paying for it? The same service that’s been paying for it this whole time I guess (t3k). Is there some undertone of “why is this free” or is it a face value question

Yeah the count of profiles would be pretty insane. I guess they evaluated all that. I never really thought about the nuts and bolts or expected an in depth answer as to why all are being upgraded to a2. I would also just think if you have the ability to make them all a2 then just do it. But I guess if it’s a burning question to ask then you could tag and ask them directly
 
That's really cool for free stuff - but who's paying for all those compute hours to re-train?

And why bother training every single NAM profile - there are sure to be some real clunkers in that bunch. I'd just choose something like top 200 profiles (or some other number - don't know how many total profiles are on tone3000)

On the topic of 'top 200', there are countless packs of models that people won't have tried but are just as good.
Simply because it's not a fender or marshall or mesa or vox or peavey etc.

The current plan is the best choice, it raises the bar for everything and everyone.
 
Back
Top