Capture your gear for free with TONEZONE3000

Technically yes, as there are no terms of service for the account. It's certainly something @staas should add just to avoid any misunderstanding.

It does make creating captures very easy as it moves all the technical knowhow to the actual setting up your gear for captures.

Wasn't so much curious about "taking" the Captures made .... if I'm loading my own Capture files to a Cloud I know nothing about to do the work for me ..... then I'm *expecting* it to be copied .... I wouldn't give a shit.

What I would want to know is (a) is the person doing and offering this service for free doing so on hardware that they own and have chosen to offer to users for free = very kind and generous of them ..... or ...... (b) is this person "accessing" the time and bandwidth some Cloud Servers of another person without their permission / covertly = not cool <- i.m.h.o

Just my 2c .... maybe I'll ask the OP directly.
 
People do it, but it always confuses me as to why.

You’d need to adjust the underlying structure to capture time based stuff well, which is essential for compression.

Mic preamps make little sense because you’re losing the impedance interaction with the mic, as well as the fact that you’d be adding it on top of an existing preamp.

EQ’s, you’d have to use fixed settings or make a parametric model. All to achieve something that plugins can already do easily. Similarly, a compressor at fixed settings isn’t particularly helpful, let alone with no ability to adjust the sidechain between dual mono and linked stereo.

Add in the fact that stereo usage is rarely very nice with most NAM plugins, and the fact that the whole premise of NAM is to accurately capture guitar amps. You’re also adding in a ton of extra CPU cycles, upsampling and downsampling, aliasing and god knows what other compromises all the achieve what?

The fact that it can do other things doesn’t mean it’s a good idea to. Black box modelling does have a place but NAM isn’t the answer to absolutely everything.
I was just thinking of something similar to STL ControlHub, which basically snags a capture of pre-set gear.
 
Has one used this/NAM to capture studio outboard signal chains? Like mic preamps/compressors/EQs.
I tried to capture the line circuits of my mic preamps (CAPI VP28, Chameleon Labs 7603) a while back to see if the saturation of pushing the input could be captured.

The VP28 has a unity gain mode that users like to run tracks through. The Chameleon Labs has a line in. I never really got much feedback on how mine turned out.

https://tonehunt.org/OutmodedElectronics/4d8566f3-2525-4043-8003-cb23c07a9d72
https://tonehunt.org/OutmodedElectronics/1d97a8ca-a804-43de-a874-607e73d2ffbf
 
I was just thinking of something similar to STL ControlHub, which basically snags a capture of pre-set gear.
I can’t really understand that concept either - how can you know what chain you want/need to use before you’ve even heard the sound? what relevance is someone else’s idea of a chain going to be to you? I have some bits of gear I often like to use in certain places but it’s never with the same settings, and it’s also very often I don’t use them at all.

There’s an abundance of options that do that kind of thing 10000% better IMO.
 
I built something! It's called TONEZONE3000 (https://tonezone3000.com) and it's built on top of Nueral Amp Modeler (NAM) the open source project by Steve Atkinson. With this tool you can capture a digital model of your amps, pedals and signal chains for free.

Training/capturing tones is hard - modeling pedals are expensive, colab is frustrating/expensive, local set ups are daunting. My goal is to change that by making training simple, free and accessible. With TONEZONE3000, you can easily train models online, fast and free, with a slick UX.

Here are a few reasons why I've enjoyed TONEZONE3000 coming from colab / my local setup:
  • Super fast and totally free training on RTX 4090s (no longer consuming my own resources)
  • Dead simple workflow
  • Cloud-based, so you can start training and close your computer
  • Train multiple models concurrently
  • An organized library of your tones
  • Easily share your models with friends
  • Preview your models with a range of DIs
Btw, you can train models with just a Dry/Wet Pair (DI and its matching stem). This has enabled me to create tones models from sessions I recorded years ago!

I’m sharing the beta to you now, and I'd love for you to train a model (or ten!) and share your feedback and models
:)


Here's a capture of a Klon clone I built with my buddy: https://www.tonezone3000.com/tonezone3000/tones/klon-centaur-silver-clone-130

staas

Could you please explain how this is a free service (?)

a => have you setup your own Server and own hardware which you fully own and control and are doing this as a free service to NAM makers and you are paying all the bills to run this service ?

or

b => are you "accessing" the time and bandwidth of some Cloud Server(s) of another person or entity .... and do you have their permission and their agreement for them to cover the running costs to offer this service to all for free ?

Ben
 
Wasn't so much curious about "taking" the Captures made .... if I'm loading my own Capture files to a Cloud I know nothing about to do the work for me ..... then I'm *expecting* it to be copied .... I wouldn't give a shit.

What I would want to know is (a) is the person doing and offering this service for free doing so on hardware that they own and have chosen to offer to users for free = very kind and generous of them ..... or ...... (b) is this person "accessing" the time and bandwidth some Cloud Servers of another person without their permission / covertly = not cool <- i.m.h.o

Just my 2c .... maybe I'll ask the OP directly.
Yea, I didnt see this and think "I better figure out what they are up to!" lol
 
Ha....just realized the name.....of the OP....so dumb went right over my head.
 
I can’t really understand that concept either - how can you know what chain you want/need to use before you’ve even heard the sound? what relevance is someone else’s idea of a chain going to be to you? I have some bits of gear I often like to use in certain places but it’s never with the same settings, and it’s also very often I don’t use them at all.

There’s an abundance of options that do that kind of thing 10000% better IMO.
I mean all of those questions are relevant to guitar captures as well, no? I think it a similar idea. Flip through a library until something sounds 97% there and do the small adjustments.
 
I built something! It's called TONEZONE3000 (https://tonezone3000.com) and it's built on top of Nueral Amp Modeler (NAM) the open source project by Steve Atkinson. With this tool you can capture a digital model of your amps, pedals and signal chains for free.

Training/capturing tones is hard - modeling pedals are expensive, colab is frustrating/expensive, local set ups are daunting. My goal is to change that by making training simple, free and accessible. With TONEZONE3000, you can easily train models online, fast and free, with a slick UX.

Here are a few reasons why I've enjoyed TONEZONE3000 coming from colab / my local setup:
  • Super fast and totally free training on RTX 4090s (no longer consuming my own resources)
  • Dead simple workflow
  • Cloud-based, so you can start training and close your computer
  • Train multiple models concurrently
  • An organized library of your tones
  • Easily share your models with friends
  • Preview your models with a range of DIs
Btw, you can train models with just a Dry/Wet Pair (DI and its matching stem). This has enabled me to create tones models from sessions I recorded years ago!

I’m sharing the beta to you now, and I'd love for you to train a model (or ten!) and share your feedback and models
:)


Here's a capture of a Klon clone I built with my buddy: https://www.tonezone3000.com/tonezone3000/tones/klon-centaur-silver-clone-130
So I make a ton of Kemper and QC captures. A lot of my customers have asked about NAM models and I’ve honestly not bothered because I had heard it was a bit of a PITA.

A week ago or so one of them mentioned this. I’ve not tried it yet but I guess I may have to?
 
I mean all of those questions are relevant to guitar captures as well, no? I think it a similar idea. Flip through a library until something sounds 97% there and do the small adjustments.
IDK, I’d say a NAM capture is more like a synth preset. It’s a starting off point that you’ll shape further with an IR and write something with. Reverb presets have uses too for that reason.

Compression and eq presets aren't really too helpful - in the time you’ve flipped through 3 or 4 (drastically different sounding) presets you could just dial in what you need. A random NAM model that catches you off guard might be useful, a random eq or compressor setting probably won’t be.
 
IDK, I’d say a NAM capture is more like a synth preset. It’s a starting off point that you’ll shape further with an IR and write something with. Reverb presets have uses too for that reason.

Compression and eq presets aren't really too helpful - in the time you’ve flipped through 3 or 4 (drastically different sounding) presets you could just dial in what you need. A random NAM model that catches you off guard might be useful, a random eq or compressor setting probably won’t be.
That’s why I use modeling vs profiling, by the time I flip through enough profiles, find my IR, eq, etc i could have just dialed in a model. But there’s no reason this CAN’T capture a signal chain of just about anything that doesn’t include time based fx. Like ControlHub.
 
That’s why I use modeling vs profiling, by the time I flip through enough profiles, find my IR, eq, etc i could have just dialed in a model. But there’s no reason this CAN’T capture a signal chain of just about anything that doesn’t include time based fx. Like ControlHub.
I’m still not really seeing the benefit. With a guitar amp it saves you from having a load amp blasting, as well as time spent reamping. A guitar amp tone might be consistent across several songs or even on different projects.

With studio gear, IDK if the trade off with aliasing, up and down sampling, lack of stereo models, fixed settings etc is worth it over just using a plugin. The difference between most plugins and HW these days is way smaller than with guitar amps which have way more distortion and non linearities going on.

But each to their own, to me it makes more problems than it solves and the alternative approach of just using normal plugins has no downsides.
 
I've decided to give this a spin and capture a simple 3-block Helix Native preset with an OD (Horizon Drive), amp model (Badonk) and IR (my Dua Verde IR).

First pass was with the wet/dry training at 1000 epochs. The tone was spot on but with some grating high end, sounded like upper-range white noise that was clearly audible without tests, verified by a null test.

Second pass was with the provided sweep file and while this model didn't have the trashy noise in the high end, it didn't null with the original but the audible differences were very subtle.

Last night before Calling it Quits I set up my EVH 50W Stealth preamp out into Ignite's TPA-1, ran the sweep through it and let it train while I slept. Gonna crack it open this morning when I feel like moving.
 
I'm not the OP, but having read many of the FAQ answers elsewhere, eg. a post in the NAM Fb group...
this is my understanding of the situation:

1. How is this free?
TZ3k is a venture-backed studio, which allows and affords them to offer free GPU resources to the community.
Eventually, if demand grows too big then it would be time for them to figure out a business plan of sorts.

2. What happens to my uploaded files and generated models?
Your files are securely stored on Amazon S3, accessible only to you and those you share the link with. They are not visible to other TZ3 users, and are not used for any other purposes.

3. Will there be a written date policy / tos / eula etc...
Yep, I believe this is planned.
Until then, see (2).
 
staas

Could you please explain how this is a free service (?)

a => have you setup your own Server and own hardware which you fully own and control and are doing this as a free service to NAM makers and you are paying all the bills to run this service ?

or

b => are you "accessing" the time and bandwidth of some Cloud Server(s) of another person or entity .... and do you have their permission and their agreement for them to cover the running costs to offer this service to all for free ?

Ben
This is a direct question I asked in the FB group the day Tonezone was announced (~4 weeks ago or so). The answer we got was the guys hosting the platform are cofounders of a vc backed studio which means they have resources to offer for the compute requirements of Tonezone3000
 
Back
Top