Hello,

I`m using the original rnnoise at a probject that runs on Linux- and WIndows-machines. So I would like zo build rnnoise for windows as a dll that has the exact same entrypoints as the original version... to be able to use the dotnet part unchanged on Linux and Windows.

Sadly, my c++-skills aren´t the best... can anyone help me?

Thanks alot.

Carl

Hi, how to make this work real-time, like input 480 (10ms), out 480 (10ms)

clean audio is silent audio, noise audio is pink noise audio, 175MB/48000hz/1channel/S16le

while exec python rnn_train.py feature.h5, it always stop here

Can someone please explain how to train it with a new speech and noise file

Hello,
I believe that this is a fairly simple question but since I'm very new to ML in general, it still baffles me. I just followed the training instruction and has successfully trained my model on one pair of data (a clean speech.wav and a noise.wav) now I want to ask how can you repeat this process for larger dataset, I'm currently having a set of data with 300 files for these 2 categories and I don't think repeating this process 300 times is the way I should go.

Thanks.

The original project has BSD-3-Clause license, but this repo doesn't mention any license.
If it is not open source, please clearly state it.

Thanks in advance.

Hello @jagger2048, thanks for your Windows adaption, I have successfully trained my own model and it is working really well with your adaption using 48KHz.

Now I need to add RNNoise to the Mozilla's DeepSpeech feeding pipeline, first I'm experimenting with your adaption and https://github.com/cpuimage/resampler to up-sample/down-sample. It is working but it generates a weird sound like static:
cleanspeech.zip

The code that I'm using with the resampler mentioned above :

https://gist.github.com/carlfm01/c8f49562b97a66fea1a72d23d1259b7d

Any idea of what's wrong?

Thanks.

I create new rnn_data.c through a README. I want rnnoise .wav file using rnn_data.c but I don't know how to make it. Thank you for your answer.

When I use src/complie.sh to constract a denoise_training tool.

I found that when I run the following in linux:

./denoise_training TRAIN_DR1_MRDD0_SI1680.WAV n95.wav 10000 feature.dat
There is an error:

but when I follow the instruction and run this:
./denoise_training TRAIN_DR1_MRDD0_SI1680.WAV n95.wav 10000
The date all directly outputs on the terminal like that

I try to use 16K wav file as program input, then transcoded to 16k output, tut the noise reduction effect is poor

I have built an executable using the original RNNoise code on Windows, but when executing it, it didn't work (I only obtain a noisy output). Your code on the other hand works. What are the changes that you made to the original code to make it Windows compatible? Was the problem in the src or the example code?