Patch synth1

Does it have to come from Daichi himself? I downloaded just one zip file of an AI bank and I am truly impressed. They are even named somewhat properly and grouped into similar instruments. I am impressed. Skip to content. Wait, what? Yeah, that should be a thing. They are automatically replenished each day. To gain more daily download tokens, become a contributor! Upload some of your own patches to the community, like some downloads you enjoy, and you'll rise in status and be a valuable member.

My Account Login. Join [FREE]. File Name synth1. Continue Cancel. Member Admin. Topic starter May 2, pm. Updated May 4, I've figured out a complicated way to unify presets for the incredibly popular free VST synth Synth1 by Daichi Laboratory , and I've started working through the or so patch-banks circulating on the Internet.

This is in regular. Eventually these must be found and fixed. Active Member. May 2, pm. Reply Quote. Trusted Member. May 4, pm. I have categorised some of the synth1 patches some while ago May 6, am. P liked. Topic starter May 6, pm. May 9, am. May 12, am. Eminent Member. Hi, there's an interesting plugin witch classify Synth1 sounds. Erik van Wees. May 21, pm. Topic starter May 22, am. This is the point where years of experience reigns supreme over me taking a single machine learning course in undergrad, browsing the theory on Wikipedia, and reading a few Medium posts.

This is where the aggressive searching and trawling through forum posts began. After some tests, instead of training on all of the parameters; I sat down and tried to figure out the essentials. I also did some more data viz in the feature engineering jupyter notebook and dropped vars which had a low amount of variance. In the end, my model was training on a much smaller subset of parameters, only about a dozen.

Finally, I did away with all pretense of separating versions and ended up throwing all 25k presets into one big batch. I knew this might not be completely accurate and there would be subtle errors, but at this point I was pulling out all the stops and wanted to get something workable. Probably the first thing I should have done, but it never occurred to me. ML models usually perform better if you normalize the data to a fixed range beforehand, since that allows the model to push gradients better without them vanishing or exploding.

My model was using a tanh activation function, so my research found that normalizing my data to [-1,1] would give the best results. However , not all of my data was numerical. In fact, a lot of my data was categorical: wave shape, filter type, delay type, etc. I learned that the way you usually remedy this is by using a one-hot encoding of the categorical variables.

Instead of having a single variable take on a range of N values, instead add N new parameters and have a binary number indicate which data parameter is present. In theory, this should work fine, but in practice there were some minor issues I had to overcome. When converting back from one-hot encoding to the standard categorical variable, my strategy was to round everything to the nearest integer since I already had that pipeline set up from the previous iteration.

For example, if the model spit out something such as [0. Then we can reverse the one hot encoding and get back our original categorical variable Square. However, sometimes the model gave back some bogus answers such as [0. In hindsight, I could have just fixed this by making the categorical variable the maximum value in the one hot encoding which would be 0. Now that my data was cleaned up, I looked into seeing if I could also tune up my model a bit more.

One thing I did take away with all of this is that I need to be more patient in training my models.