Hey 
after recently discovering and loving the Mod Dwarf the only thing that I was really missing was a place to share and discover tone models.
So I built something → https://www.tonelibrary.net
Please do keep in mind that it is very early days (every feedback is welcome though) and it currently only supports uploading and downloading models.
Features that should be added soon:
- Searching
- Filtering
- Icons for Amp/Amp+Cab/Pedal etc
- Previewing a tone model using the web based player
- Taking some time to actually work on the UX/design
In order to upload a model (and please do!) you have to sign up and verify your email address once.
As I said every bit of feedback/bug reports/feature requests helps.
The whole thing is open source and I will try to start outlining planned future work.
P.S. I only uploaded one model so far (due to spending most of my spare time on getting it together) of a JP2C lead channel, so if you think this could be a good thing for the community - please upload your files.
15 Likes
Great idea
IMVHO some kind of rating system would be necessary
1 Like
Agreed! It is on the list - as I said - the project was born roughly a week ago 
Would you rather have a star based systems or favorites (which can double as a rating).
And - if you created any models - please do upload them so there is more to discover for people
1 Like
Hey thanks @TobiasBales for the effort. We definitely need a place for all the models.
But we still need to change a few things(file extension, metadata, easier training) for catering to a upload platform.
That’s why we talked to the people from ToneHunt.org as an option - which makes also a lot of sense.
The project is mature enough and there are enough devs to help out.
With the option later one to integrate the API directly into the plugin. This would replace the constant downloading and uploading models which is a cumbersome process on the user side.
Not to say your platform can’t be an option but just to let you know what our approach is.
2 Likes
Hey,
thanks for the info, just for clarification:
You expect format, metadata etc to change over time and on top of that you already talked to the tonehunt folks about sharing the platform?
If so, that is great, and I’ll probably just stop here then tbh.
I talked to @gianfranco ahead of time to avoid work that ultimately will not bring a lot of value but I guess he might not not have been aware of those conversations?
format stays the same - only minor easy of use improvements.
And the Tonehunt just wait on the fixes on our side.
might be an overlap here - we only started to talk to them at the beginning of may.
Could you close/remove the thread?
So people don’t start using it, no hard feelings or anything just don’t want people to put effort into uploading models then.
My 2 cents here:
Adding the models to a mature platform like Tonehunt.org would be cool and is probably the way to go. However, it’ll need some time to allow their site to support NAM and AIDA-X side by side.
Until then I think @TobiasBales ’ site is already now better for sharing the models, than the forum.
2 Likes
100%. Forum is not great for sharing files. But you still have to consider other aspects with a more public facing site. Licensing and file security just to name a few.
this was already planned on there end from the get go - so only graphical stuff to sort out.
1 Like
I changed the topic but will keep it open for a general discussion. Which I think is good to have. Considering there are a lot of people with dev experience on the forum.
Just to elaborate a bit: I also have a backend (strapi, postgres) and frontend ready for tone models that integrates with our website.
But this becomes hard to maintain when opening to uploads. So I put it on hold for now.
Another option for a ready now solution is www.patchstorage.com . To whom I also talked and would be ready to accept model uploads.
4 Likes
Hi all,
It’s great to see so much interest in the platform.
From what I read, AIDA would currently be more viable on SBCs than NAM, as NAM’s algorithm albeit in theory more accurate, it’s more processing intensive.
I saw somewhere there was an intention to make NAM profiles readable by AIDA, in which all the current Tonehunt library would become usable by AIDA. Is that still the intention? Would anyone know roughly how far that is?
Thank you
Hi @chrisp250
@madmaxwell is the person for the precise reply, but from what I understood, even if AIDA loads the NAM models, it does not sort out the CPU issue.
The high CPU load needed to process NAM models is due to the network structure type they use. If AIDA can support the same model structure and is able to load NAM models, the CPU continues to be high. The issue is in the models themselves and not the model loader.
1 Like
Thank you very much for your reply.
I understand the situation. I guess the other alternative would be to use .nam files to train AIDA and produce an AIDA profile of that model. Not great, but even though we would have to go through the training process again, at least you could do this just with a computer without having to rent amplifiers, cabinets, etc.
Cheers
This would generate lesser quality models, like taking a photo of a photo. The most promising case is that people who did model training and still have the audio files can re-use this audio captures to train AIDA models.
By looking at their git repo, it seems like the NAM devs are adding LSTM support, which would make it possible to run in SBCs, but will bring a lot of confusion with two NAM model types, and will not sort out the frustration of not running “all the NAM models out there” in an SBC.
Let’s see. I’ve been looking with a lot of curiosity what is happening with NAM and I feel there still a lot of maturing required for the project to advance. There was really a lot of traction back in late March, when the FB group was growing more than 1000 users per week and members were all over their head with the growth (there was even a post of a guy somehow correlating the spike in NVidia stock prices to the surge of NAM users 
) but lately the growth has reduced a lot, in a typical S-shaped curve with 11k as a top ceiling, showing some sort of exhaustion on growth.
1 Like
Thank you @gianfranco for your insights.
I didn’t know it was possible to use the resulting wav from NAM’s training audio for AIDA.
I imagine as NAM keeps maturing, interest will increase in steps as you described. As new features get rolled out, NAM will become appealing to a wider user base. Still NAM has yet a long road ahead, and unlike AIDA, it’s not part of a complete solution yet. AIDA’s advantage is mod has been around for a while and has a good user base. Having said that, I tried to build mod on in my linux lab and failed miserably at getting mod-ui to work.
In a way, it feels that all this early effort going towards profiling amps for NAM might be at partially wasted as NAM evolves and future version introduce changes that might break existing profiles.
We’ll see what happens, certainly very exciting times!
Cheers
3 Likes
Hi!
I wanted to chime in my view.
Ive been using NAM and Proteus for a while now and making my own profiles (still learning).Since AIDA came out i tried the plugin and it was really good!It spiked my interest so i got myself a Dwarf and i have been getting really nice results with the captures!Not much difference in sound between NAM , AIDA and Proteus.Again the biggest difference is the cabinet IRs.Anyway AIDA is newer and needs a bit of polishing maybe but im thinking of playing live with just the Dwarf!Nice stuff!!
Cheers!
6 Likes
That’s cool.
I’ve been thinking about getting a Dwaft myself, but still not ready to spend the $$$ as I have a solution that works for me and this is more of a curiosity for me at the moment.
I’m eager to see the progress in the next 6 months.
Cheers
2 Likes