marlowg01 Posted February 6, 2021 Share Posted February 6, 2021 This machine learning plugin is amazing. It is CPU heavy, but the tone is fantastic. Guitar plugin made with JUCE that uses neural network models to emulate real world hardware. This plugin uses a LSTM model to recreate the sound of real amps and pedals. You can record samples and train models from the plugin. Tone models are saved in .json format. Model training is accomplished using Tensorflow/Keras. The main improvement from the original SmartAmp is that training takes less than five minutes on CPU (vs. 8 hours on GPU) for comparable sound quality. Training has also been integrated into the plugin. https://github.com/GuitarML/SmartAmpPro 1 Link to comment Share on other sites More sharing options...
jude77 Posted February 7, 2021 Share Posted February 7, 2021 I'll be downloading this a little later, but it looks amazing. Link to comment Share on other sites More sharing options...
jude77 Posted February 7, 2021 Share Posted February 7, 2021 (edited) I just gave it a test drive and, IMHO, it sounds stellar and is well worth a look. For a freebie it's a must-have. FWIW I didn't find it hit my machine that hard. Also, for my DAW (Studio One) to find it I had to install it in a VST3 specific folder rather than a general VST folder. Edited February 7, 2021 by jude77 Link to comment Share on other sites More sharing options...
marlowg01 Posted February 7, 2021 Author Share Posted February 7, 2021 I spoke to the developer and he is gathering models to share. Anyone willing to share their .JSON models, I'd be interested too. I'll post some here for those interested. 4 Link to comment Share on other sites More sharing options...
marlowg01 Posted February 7, 2021 Author Share Posted February 7, 2021 Here are a couple of models that I made... DODFX86B.json FAB_Metal.json Tantrum_Pedal.json Marshall_VS100_OD.json 3 Link to comment Share on other sites More sharing options...
Bruno de Souza Lino Posted February 7, 2021 Share Posted February 7, 2021 Too heavy on the CPU. Had to pass. Link to comment Share on other sites More sharing options...
jude77 Posted February 8, 2021 Share Posted February 8, 2021 Thanks for posting the models! Link to comment Share on other sites More sharing options...
filo Posted February 8, 2021 Share Posted February 8, 2021 9 hours ago, Bruno de Souza Lino said: Too heavy on the CPU. Had to pass. Could you compare it to the Neural stuff? (the most hungry VST that I know) Link to comment Share on other sites More sharing options...
Bruno de Souza Lino Posted February 8, 2021 Share Posted February 8, 2021 4 hours ago, filo said: Could you compare it to the Neural stuff? (the most hungry VST that I know) I'd say that is as resource heavy as the pre-Archetype Gojira plugins. I don't see why guitar amp emulations need to be so heavy in resources for what is sometimes a marginal improvement in sound quality. 1 Link to comment Share on other sites More sharing options...
marlowg01 Posted February 9, 2021 Author Share Posted February 9, 2021 If anyone wants me to convert a model for them, I can. Just send me the direct guitar recording and the one from the plugin or hardware pedal or amplifier and I can convert and post or email the .JSON file for you. The recording has to be about 3.5min long to get the best model. The plugin was struggling to work for me for conversions, so I am doing it via command line syntax to create the models. I could also share how to do that for others if there is interest. Here are a couple of other models that I think turned out well, too, taken from a plugin and from the G1X4 pedal from Zoom: G1X4_DZ_DRV.json G1X4_ORG120.json G1X4_NYC_Muff.json Marshall_Super_Lead_1959_Plexi_100W.json 4 Link to comment Share on other sites More sharing options...
Xoo Posted February 9, 2021 Share Posted February 9, 2021 It was unusable here - as soon as it was inserted into an FX bin, CbB's GUI ground to a halt and a couple of times, I had to Task Manager kill CbB to close it. My PC isn't cutting edge, but it's no slouch (Ryzen 5, 6/12 core, 3.4GHz, 32GB RAM, nVidia 1030 graphics card). Link to comment Share on other sites More sharing options...
Esteban Villanova Posted February 9, 2021 Share Posted February 9, 2021 This thing is like a software Kemper, right? I’ll test tonight in Linux. 1 Link to comment Share on other sites More sharing options...
Bruno de Souza Lino Posted February 9, 2021 Share Posted February 9, 2021 2 hours ago, Kevin Perry said: My PC isn't cutting edge, but it's no slouch (Ryzen 5, 6/12 core, 3.4GHz, 32GB RAM, nVidia 1030 graphics card). I'm gonna guess the developer optimized this for Intel CPUs, since I also had issues and I also have an AMD CPU. 1 Link to comment Share on other sites More sharing options...
Tomgu Posted February 9, 2021 Share Posted February 9, 2021 Can it emulate NeuralDSP? 2 Link to comment Share on other sites More sharing options...
jude77 Posted February 9, 2021 Share Posted February 9, 2021 9 hours ago, marlowg01 said: The plugin was struggling to work for me for conversions, so I am doing it via command line syntax to create the models. I could also share how to do that for others if there is interest. I would be interested in learning how to do this. I downloaded the software but am completely baffled by it. Are there instructions anywhere? Link to comment Share on other sites More sharing options...
Grem Posted February 9, 2021 Share Posted February 9, 2021 How hard is it to install? I was reading about installing a Python library? Is this correct? 1 Link to comment Share on other sites More sharing options...
Bruno de Souza Lino Posted February 9, 2021 Share Posted February 9, 2021 4 hours ago, Grem said: How hard is it to install? I was reading about installing a Python library? Is this correct? That's only to creates profiles, since it uses TensorFlow 1 Link to comment Share on other sites More sharing options...
Grem Posted February 10, 2021 Share Posted February 10, 2021 2 hours ago, Bruno de Souza Lino said: That's only to creates profiles, since it uses TensorFlow But isn't creating profiles of stuff the whole idea behind the plugin? Maybe I am reading more into it than there is? Link to comment Share on other sites More sharing options...
Bruno de Souza Lino Posted February 10, 2021 Share Posted February 10, 2021 10 minutes ago, Grem said: But isn't creating profiles of stuff the whole idea behind the plugin? Maybe I am reading more into it than there is? 1 Link to comment Share on other sites More sharing options...
marlowg01 Posted February 10, 2021 Author Share Posted February 10, 2021 So I actually started working with these machine learning plugins with the PedalNet, where you download the git package and open the documents for setting up the software. This video describes the software setup and then you can skip and download the SmartGuitarPro part and use the commandline at the bottom of this post. Once I had completed that, I also installed the pip requirements for the SmartGuitarPro software. However, once I ran the plugin, a commandline window opened that said that I needed to install python and tensorflow, etc. Instead, I ran the terminal through Anaconda3 which I had used for the previous plugin and ran the following command lines for each stereo, hard-panned, un-effected on the left, effected on the right: (base) C:\Users\<username>\AppData\Roaming\GuitarML\SmartAmpPro\training>python train.py C:\Users\<username>\<location_of_stereo_wav_file(uneffected-Left, effected-Right)>\G1X4_MS800.wav G1X4_MS800 I kept the name of the wave file, but removed the *.wav extension and that is the name of the model saved. It saves in the models and tones folders in this location: C:\Users\<username>\AppData\Roaming\GuitarML\SmartAmpPro\ When you reopen the plugin, it re-scans the *.JSON models and you can access them from the pull-down. Each model takes about 5 minutes because it splits the data and using the power of your CPU to process the data more quickly than the GPU-accelerated PedalNet variant, which takes 8 hours to 2.5 days to create models depending upon your computer and hardware. I downloaded the non-cuda pytorch package because I do not have any GPUs that support either machine learning algorithms. I have spoken to the developer about my CPU woes with this plugin as well, and he suggested compiling for the specific system that you have to improve performance. I plan to do this and look at the code to see if I can see any optimizations that can reduce CPU. I have found that closing the GUI greatly reduces CPU, so I think that the GUI coding is probably trying to update too frequently, but I am not sure. Post what you create if you can! I'd love to see what others produce too. And here are a few new ones I made that I think also turned out pretty well: G1X4_HG_Throttle.json G1X4_RedCrunch.json G1X4_UK30A.json ENGL_Savage_120.json 2 Link to comment Share on other sites More sharing options...
Recommended Posts
Please sign in to comment
You will be able to leave a comment after signing in
Sign In Now