Add simulate_rir_ism method for room impulse response simulation #2880
Conversation
|
@nateanl has imported this pull request. If you are a Meta employee, you can view this diff on Phabricator. |
1 similar comment
|
@nateanl has imported this pull request. If you are a Meta employee, you can view this diff on Phabricator. |
nateanl
force-pushed
the
rir_ism
branch
2 times, most recently
from
7b5a3c1 to
15763cc
Compare
|
@nateanl has imported this pull request. If you are a Meta employee, you can view this diff on Phabricator. |
test/torchaudio_unittest/prototype/functional/functional_test_impl.py
Outdated
Show resolved
Hide resolved
| if e_abs.shape[0] > 1: | ||
| if center_frequency is None: | ||
| center = torch.tensor( | ||
| [125.0, 250.0, 500.0, 1000.0, 2000.0, 4000.0, 8000.0], dtype=room.dtype, device=room.device |
There was a problem hiding this comment.
where do these values come from? should they be generated programmatically or made customizable?
There was a problem hiding this comment.
It is mentioned in the note of docstring.
test/torchaudio_unittest/prototype/functional/functional_test_impl.py
Outdated
Show resolved
Hide resolved
nateanl
force-pushed
the
rir_ism
branch
2 times, most recently
from
a2e3285 to
b0e2061
Compare
|
@nateanl has imported this pull request. If you are a Meta employee, you can view this diff on Phabricator. |
torchaudio/_extension/__init__.py
Outdated
| _load_lib("libtorchaudio") | ||
|
|
||
| import torchaudio.lib._torchaudio # noqa | ||
|
|
There was a problem hiding this comment.
This part is now repeated multiple times, which is redundant. Can you merge the initialization logics around extension, kaldi, RIR as follow?
_IS_RIR_AVAILABLE = False
_IS_KALDI_AVAILABLE = False
_IS_TORCHAUDIO_EXT_AVAILABLE = is_module_available("torchaudio.lib._torchaudio")
if _IS_TORCHAUDIO_EXT_AVAILABLE:
_load_lib("libtorchaudio")
import torchaudio.lib._torchaudio # noqa
_check_cuda_version()
_IS_RIR_AVAILABLE = torchaudio.lib._torchaudio.is_rir_available()
_IS_KALDI_AVAILABLE = torchaudio.lib._torchaudio.is_kaldi_available()
|
|
||
| } // Anonymous namespace | ||
|
|
||
| TORCH_LIBRARY_IMPL(torchaudio, CPU, m) { |
There was a problem hiding this comment.
Can you put TORCH_LIBRARY_IMPL and TORCH_LIBRARY_FRAGMENT in anonymous namespace as well?
|
@nateanl has imported this pull request. If you are a Meta employee, you can view this diff on Phabricator. |
|
@nateanl has imported this pull request. If you are a Meta employee, you can view this diff on Phabricator. |
|
Hey @nateanl. |
replicate of #2644