Closed r9y9 closed 6 years ago
No, as I remember correctly it wasn't a ParameterError. Could you paste your code so that I can reproduce the issue? For now there is no plan to implement lsp2lpc from scratch.
Thanks for the quick response! I modified your test code and inserted it at the end of the speech analysis notebook. It now results in an AssertionError.
from pysptk.synthesis import LSPDF
order = 10
lpc = np.apply_along_axis(pysptk.lpc, 1, frames, order)
# make sure lsp has loggain
lsp = np.apply_along_axis(pysptk.lpc2lsp, 1, lpc)
lsp[:, 0] = np.log(lsp[:, 0])
# synthesis
synthesizer = Synthesizer(LSPDF(order), hop_length)
y = synthesizer.synthesis(source_excitation, lsp)
assert np.all(np.isfinite(y))
AssertionError Traceback (most recent call last)
<ipython-input-42-0aa71a6f3baa> in <module>()
10 synthesizer = Synthesizer(LSPDF(order), hop_length)
11 y = synthesizer.synthesis(source_excitation, lsp)
---> 12 assert np.all(np.isfinite(y))
AssertionError:
Yes, this is what I was trying to solve.
Ok, at least we're on the same page :) The problem seems to be for orders out of the range [20,25] as well.
On Tue, May 24, 2016 at 4:02 PM, Ryuichi Yamamoto notifications@github.com wrote:
Yes, this is what I was trying to solve.
— You are receiving this because you commented. Reply to this email directly or view it on GitHub https://github.com/r9y9/pysptk/issues/22#issuecomment-221261362
Is that the same issue you see before? This is not a ParameterError.
Yes, sorry. I got a ParameterError when I tried to plot the result using waveplot.
On Tue, May 24, 2016 at 5:01 PM, Ryuichi Yamamoto notifications@github.com wrote:
Is that the same issue you see before? This is not a ParameterError.
— You are receiving this because you commented. Reply to this email directly or view it on GitHub https://github.com/r9y9/pysptk/issues/22#issuecomment-221279879
When trying to use LSPDF after applying lpc2lsp, I keep getting "ParameterError: Audio buffer is not finite everywhere". Is this the bug you're referring to? If so, has there been any progress in fixing it? Or, alternatively, are there plans to implement an lsp2lpc function?
Thanks so much!