-
@cplee1 has noticed that essentially for all MSPs with small-scale structures, we are not able to resolve them (whereas in the legacy beamformer, and even in the PSRFITS output we can at least see hin…
-
## ❓ Questions and Help
### Before asking:
1. search the issues.
2. search the docs.
#### What is your question?
In the paper "Transformers with convolutional context for ASR"…
-
HI,sir! Your open source code is very exciting, its features are powerful,I am currently doing end-to-end speech translation tasks(chinese speech to english text),I want to refer to your code.I have a…
-
Hi, I cannot make same result of librosa's mfccs. This maybe relate to #206
Librosa code:
```python
def getMfcc(file_name):
X, sample_rate = sf.read(file_name, dtype='float32')
mfccs = l…
-
When using dsa110-pol to generate polarization calibrated filterbank files, frequency channels that are assigned to be masked are instead overwritten with a constant fill value. Existing Level 3 calib…
-
Add the scattering transform by Stephané Mallat.
This should now be actually possible using _warpedfilters_ with logarithmic scale or with dual-tree complex wavelet transform _dtwfb_
-
At https://github.com/tyiannak/pyAudioAnalysis/blob/master/pyAudioAnalysis/audioFeatureExtraction.py#L169, there's a check on the sampling frequency `fs < 8000` and then the number of logistic filters…
-
Hello
This reaserch is interesting. I want to try to used the envelope_spectrogram.
But I can't find the NSGT_ERB function.
May u give me some help? thanks.
-
>I always find shortened parameter names a bit cryptic (`sr`, `n_mels`, `n_stft`, `f_max`, `f_min`). Do we want to follow librosa here? If we do, we should do it correctly; if we don't, we can choose …
-
Hi,
When I tried to extract fbank features I found missing frequencies. So, I checked the filter bank in your code and found this:
![image](https://user-images.githubusercontent.com/24974144/62096…