-
Hi,
I compute the mel spectrogram on a time-domain signal that has `13230080` samples, like so:
mel_spectrogram = librosa.feature.melspectrogram(audio_data, sr=44100, n_fft=2048, hop_length=…
-
I find this sentence in code: window_size = 512 * (frames - 1)
I read the paper and found that the window size in this project is 1024, so why we have to use "512*(frames-1)" instead…
-
https://github.com/librosa/librosa/blob/c9c4254b540a8cf1091c6126c1be4890c86d298c/librosa/feature/spectral.py#L1350
I might be missing something, but is there a reason that melspectrogram should out…
-
Doc build currently has a few exceptions and formatting errors. These should be fixed up ahead of the 0.5 release.
TODO items:
- [x] Fix formatting errors
- [x] All functions are contained in …
-
Hi,
I've been trying to figure out how to use feature.mfcc with a window function applied on each frame, such as the hamming function. I see that it's implemented in core.stft, but as far as I can …
-
Using a combination of a low n_fft, high n_mels, and low fmax, one can obtain a mel spectrogram where some of the bins are always 0. I haven't examined the source code but my guess would be that this …
-
The original code is redundant, but it can be summarised as below.
```
...
Parallel(n_jobs=N_JOBS)(delayed(convert)(path) for path in paths)
...
def convert(path):
x, sr = librosa.load(path)
…
-
In your program , you have one wav file and you extract the MFCC feature 13 dimension => train_inputs. Second, you construct a label array like this [19 8 5 ...] and change it to Sparse representat…
-
(Questions in bold font)
I am trying to synthesize my own voice files.
As I did not find a documentation apart from `You can use other datasets if you convert them to the right format. See ljspeec…
-
- Create a generic onset detection module
- Move the onset strength function out of librosa.beat
- Implement peak-picking algorithms. Possibly in core? I could see this being useful for multiple app…