Support compression level in i/o dispatcher backend

Differential Revision: D50367721

Pull Request resolved: https://github.com/pytorch/audio/pull/3662

src/torchaudio/_backend/backend.py

@@ -3,6 +3,7 @@ from abc import ABC, abstractmethod
 from typing import BinaryIO, Optional, Tuple, Union
 
 from torch import Tensor
+from torchaudio.io import CodecConfig
 
 from .common import AudioMetaData
 
@@ -37,6 +38,7 @@ class Backend(ABC):
         encoding: Optional[str] = None,
         bits_per_sample: Optional[int] = None,
         buffer_size: int = 4096,
+        compression: Optional[Union[CodecConfig, float, int]] = None,
     ) -> None:
         raise NotImplementedError
 

src/torchaudio/_backend/ffmpeg.py

@@ -228,6 +228,7 @@ def save_audio(
     encoding: Optional[str] = None,
     bits_per_sample: Optional[int] = None,
     buffer_size: int = 4096,
+    compression: Optional[torchaudio.io.CodecConfig] = None,
 ) -> None:
     ext = None
     if hasattr(uri, "write"):
@@ -250,6 +251,7 @@ def save_audio(
         format=_get_sample_format(src.dtype),
         encoder=encoder,
         encoder_format=enc_fmt,
+        codec_config=compression,
     )
     with s.open():
         s.write_audio_chunk(0, src)
@@ -304,7 +306,13 @@ class FFmpegBackend(Backend):
         encoding: Optional[str] = None,
         bits_per_sample: Optional[int] = None,
         buffer_size: int = 4096,
+        compression: Optional[Union[torchaudio.io.CodecConfig, float, int]] = None,
     ) -> None:
+        if not isinstance(compression, (torchaudio.io.CodecConfig, type(None))):
+            raise ValueError(
+                "FFmpeg backend expects non-`None` value for argument `compression` to be of ",
+                f"type `torchaudio.io.CodecConfig`, but received value of type {type(compression)}",
+            )
         save_audio(
             uri,
             src,
@@ -314,6 +322,7 @@ class FFmpegBackend(Backend):
             encoding,
             bits_per_sample,
             buffer_size,
+            compression,
         )
 
     @staticmethod

src/torchaudio/_backend/soundfile.py

@@ -2,6 +2,7 @@ import os
 from typing import BinaryIO, Optional, Tuple, Union
 
 import torch
+from torchaudio.io import CodecConfig
 
 from . import soundfile_backend
 from .backend import Backend
@@ -35,7 +36,11 @@ class SoundfileBackend(Backend):
         encoding: Optional[str] = None,
         bits_per_sample: Optional[int] = None,
         buffer_size: int = 4096,
+        compression: Optional[Union[CodecConfig, float, int]] = None,
     ) -> None:
+        if compression:
+            raise ValueError("soundfile backend does not support argument `compression`.")
+
         soundfile_backend.save(
             uri, src, sample_rate, channels_first, format=format, encoding=encoding, bits_per_sample=bits_per_sample
         )

src/torchaudio/_backend/sox.py

@@ -56,7 +56,13 @@ class SoXBackend(Backend):
         encoding: Optional[str] = None,
         bits_per_sample: Optional[int] = None,
         buffer_size: int = 4096,
+        compression: Optional[Union[torchaudio.io.CodecConfig, float, int]] = None,
     ) -> None:
+        if not isinstance(compression, (float, int, type(None))):
+            raise ValueError(
+                "SoX backend expects non-`None` value for argument `compression` to be of ",
+                f"type `float` or `int`, but received value of type {type(compression)}",
+            )
         if hasattr(uri, "write"):
             raise ValueError(
                 "SoX backend does not support writing to file-like objects. ",
@@ -68,7 +74,7 @@ class SoXBackend(Backend):
                 src,
                 sample_rate,
                 channels_first,
-                None,
+                compression,
                 format,
                 encoding,
                 bits_per_sample,

src/torchaudio/_backend/utils.py

@@ -5,6 +5,7 @@ from typing import BinaryIO, Dict, Optional, Tuple, Type, Union
 import torch
 
 from torchaudio._extension import lazy_import_ffmpeg_ext, lazy_import_sox_ext
+from torchaudio.io import CodecConfig
 
 from . import soundfile_backend
 
@@ -229,6 +230,7 @@ def get_save_func():
         bits_per_sample: Optional[int] = None,
         buffer_size: int = 4096,
         backend: Optional[str] = None,
+        compression: Optional[Union[CodecConfig, float, int]] = None,
     ):
         """Save audio data to file.
 
@@ -283,8 +285,32 @@ def get_save_func():
 
                 .. seealso::
                    :ref:`backend`
+
+            compression (CodecConfig, float, int, or None, optional):
+                Compression configuration to apply.
+
+                If the selected backend is FFmpeg, an instance of :py:class:`CodecConfig` must be provided.
+
+                Otherwise, if the selected backend is SoX, a float or int value corresponding to option ``-C`` of the
+                ``sox`` command line interface must be provided. For instance:
+
+                ``"mp3"``
+                    Either bitrate (in ``kbps``) with quality factor, such as ``128.2``, or
+                    VBR encoding with quality factor such as ``-4.2``. Default: ``-4.5``.
+
+                ``"flac"``
+                    Whole number from ``0`` to ``8``. ``8`` is default and highest compression.
+
+                ``"ogg"``, ``"vorbis"``
+                    Number from ``-1`` to ``10``; ``-1`` is the highest compression
+                    and lowest quality. Default: ``3``.
+
+                Refer to http://sox.sourceforge.net/soxformat.html for more details.
+
         """
         backend = dispatcher(uri, format, backend)
-        return backend.save(uri, src, sample_rate, channels_first, format, encoding, bits_per_sample, buffer_size)
+        return backend.save(
+            uri, src, sample_rate, channels_first, format, encoding, bits_per_sample, buffer_size, compression
+        )
 
     return save

test/torchaudio_unittest/backend/dispatcher/dispatcher_test.py

@@ -105,7 +105,7 @@ class DispatcherTest(PytorchTestCase):
             f"torchaudio._backend.utils.{expected_backend.__name__}.save"
         ) as mock_save:
             get_save_func()(filename, src, sample_rate, format=format)
-            mock_save.assert_called_once_with(filename, src, sample_rate, True, format, None, None, 4096)
+            mock_save.assert_called_once_with(filename, src, sample_rate, True, format, None, None, 4096, None)
 
     @parameterized.expand(
         [
@@ -126,4 +126,4 @@ class DispatcherTest(PytorchTestCase):
             f"torchaudio._backend.utils.{expected_backend.__name__}.save"
         ) as mock_save:
             get_save_func()(f, src, sample_rate, format=format, buffer_size=buffer_size)
-            mock_save.assert_called_once_with(f, src, sample_rate, True, format, None, None, buffer_size)
+            mock_save.assert_called_once_with(f, src, sample_rate, True, format, None, None, buffer_size, None)

test/torchaudio_unittest/backend/dispatcher/ffmpeg/save_test.py

@@ -4,11 +4,13 @@ import pathlib
 import subprocess
 import sys
 from functools import partial
+from typing import Optional
 
 import torch
 from parameterized import parameterized
 from torchaudio._backend.ffmpeg import _parse_save_args
 from torchaudio._backend.utils import get_save_func
+from torchaudio.io import CodecConfig
 
 from torchaudio_unittest.backend.dispatcher.sox.common import get_enc_params, name_func
 from torchaudio_unittest.common_utils import (
@@ -45,6 +47,7 @@ class SaveTestBase(TempDirMixin, TorchaudioTestCase):
         self,
         format: str,
         *,
+        compression: Optional[CodecConfig] = None,
         encoding: str = None,
         bits_per_sample: int = None,
         sample_rate: float = 8000,
@@ -104,7 +107,15 @@ class SaveTestBase(TempDirMixin, TorchaudioTestCase):
         data = load_wav(src_path, normalize=False)[0]
         if test_mode == "path":
             ext = format
-            self._save(tgt_path, data, sample_rate, format=format, encoding=encoding, bits_per_sample=bits_per_sample)
+            self._save(
+                tgt_path,
+                data,
+                sample_rate,
+                compression=compression,
+                format=format,
+                encoding=encoding,
+                bits_per_sample=bits_per_sample,
+            )
         elif test_mode == "fileobj":
             ext = None
             with open(tgt_path, "bw") as file_:
@@ -112,6 +123,7 @@ class SaveTestBase(TempDirMixin, TorchaudioTestCase):
                     file_,
                     data,
                     sample_rate,
+                    compression=compression,
                     format=format,
                     encoding=encoding,
                     bits_per_sample=bits_per_sample,
@@ -123,6 +135,7 @@ class SaveTestBase(TempDirMixin, TorchaudioTestCase):
                 file_,
                 data,
                 sample_rate,
+                compression=compression,
                 format=format,
                 encoding=encoding,
                 bits_per_sample=bits_per_sample,
@@ -198,11 +211,27 @@ class SaveTest(SaveTestBase):
         # NOTE: Supported sample formats: s16 s32 (24 bits)
         # [8, 16, 24],
         [16, 24],
+        [
+            0,
+            1,
+            2,
+            3,
+            4,
+            5,
+            6,
+            7,
+            8,
+        ],
     )
-    def test_save_flac(self, test_mode, bits_per_sample):
+    def test_save_flac(self, test_mode, bits_per_sample, compression_level):
         # -acodec flac -sample_fmt s16
         # 24 bits needs to be mapped to s32
-        self.assert_save_consistency("flac", bits_per_sample=bits_per_sample, test_mode=test_mode)
+        codec_config = CodecConfig(
+            compression_level=compression_level,
+        )
+        self.assert_save_consistency(
+            "flac", compression=codec_config, bits_per_sample=bits_per_sample, test_mode=test_mode
+        )
 
     # @nested_params(
     #     ["path", "fileobj", "bytesio"],
@@ -212,12 +241,25 @@ class SaveTest(SaveTestBase):
     #     self.assert_save_consistency("htk", test_mode=test_mode, num_channels=1)
 
     @nested_params(
+        [
+            None,
+            -1,
+            0,
+            1,
+            2,
+            3,
+            5,
+            10,
+        ],
         ["path", "fileobj", "bytesio"],
     )
-    def test_save_vorbis(self, test_mode):
+    def test_save_vorbis(self, quality_level, test_mode):
         # NOTE: ffmpeg doesn't recognize extension "vorbis", so we use "ogg"
         # self.assert_save_consistency("vorbis", test_mode=test_mode)
-        self.assert_save_consistency("ogg", test_mode=test_mode)
+        codec_config = CodecConfig(
+            qscale=quality_level,
+        )
+        self.assert_save_consistency("ogg", compression=codec_config, test_mode=test_mode)
 
     # @nested_params(
     #     ["path", "fileobj", "bytesio"],

test/torchaudio_unittest/backend/dispatcher/sox/save_test.py

@@ -40,6 +40,7 @@ class SaveTestBase(TempDirMixin, TorchaudioTestCase):
         self,
         format: str,
         *,
+        compression: float = None,
         encoding: str = None,
         bits_per_sample: int = None,
         sample_rate: float = 8000,
@@ -101,13 +102,16 @@ class SaveTestBase(TempDirMixin, TorchaudioTestCase):
         # 2.1. Convert the original wav to target format with torchaudio
         data = load_wav(src_path, normalize=False)[0]
         if test_mode == "path":
-            self._save(tgt_path, data, sample_rate, encoding=encoding, bits_per_sample=bits_per_sample)
+            self._save(
+                tgt_path, data, sample_rate, compression=compression, encoding=encoding, bits_per_sample=bits_per_sample
+            )
         elif test_mode == "fileobj":
             with open(tgt_path, "bw") as file_:
                 self._save(
                     file_,
                     data,
                     sample_rate,
+                    compression=compression,
                     format=format,
                     encoding=encoding,
                     bits_per_sample=bits_per_sample,
@@ -118,6 +122,7 @@ class SaveTestBase(TempDirMixin, TorchaudioTestCase):
                 file_,
                 data,
                 sample_rate,
+                compression=compression,
                 format=format,
                 encoding=encoding,
                 bits_per_sample=bits_per_sample,
@@ -134,7 +139,9 @@ class SaveTestBase(TempDirMixin, TorchaudioTestCase):
 
         # 3.1. Convert the original wav to target format with sox
         sox_encoding = _get_sox_encoding(encoding)
-        sox_utils.convert_audio_file(src_path, sox_path, encoding=sox_encoding, bit_depth=bits_per_sample)
+        sox_utils.convert_audio_file(
+            src_path, sox_path, compression=compression, encoding=sox_encoding, bit_depth=bits_per_sample
+        )
         # 3.2. Convert the target format to wav with sox
         sox_utils.convert_audio_file(sox_path, ref_path, encoding=cmp_encoding, bit_depth=cmp_bit_depth)
         # 3.3. Load with SciPy
@@ -175,15 +182,42 @@ class SaveTest(SaveTestBase):
 
     @nested_params(
         [8, 16, 24],
+        [
+            None,
+            0,
+            1,
+            2,
+            3,
+            4,
+            5,
+            6,
+            7,
+            8,
+        ],
+    )
+    def test_save_flac(self, bits_per_sample, compression_level):
+        self.assert_save_consistency(
+            "flac", compression=compression_level, bits_per_sample=bits_per_sample, test_mode="path"
         )
-    def test_save_flac(self, bits_per_sample):
-        self.assert_save_consistency("flac", bits_per_sample=bits_per_sample, test_mode="path")
 
     def test_save_htk(self):
         self.assert_save_consistency("htk", test_mode="path", num_channels=1)
 
-    def test_save_vorbis(self):
-        self.assert_save_consistency("vorbis", test_mode="path")
+    @nested_params(
+        [
+            None,
+            -1,
+            0,
+            1,
+            2,
+            3,
+            3.6,
+            5,
+            10,
+        ],
+    )
+    def test_save_vorbis(self, quality_level):
+        self.assert_save_consistency("vorbis", compression=quality_level, test_mode="path")
 
     @nested_params(
         [
@@ -254,9 +288,22 @@ class SaveTest(SaveTestBase):
         encoding, bits_per_sample = enc_params
         self.assert_save_consistency("amb", encoding=encoding, bits_per_sample=bits_per_sample, test_mode="path")
 
+    @nested_params(
+        [
+            None,
+            0,
+            1,
+            2,
+            3,
+            4,
+            5,
+            6,
+            7,
+        ],
+    )
     @skipIfNoSoxEncoder("amr-nb")
-    def test_save_amr_nb(self):
-        self.assert_save_consistency("amr-nb", num_channels=1, test_mode="path")
+    def test_save_amr_nb(self, bit_rate):
+        self.assert_save_consistency("amr-nb", compression=bit_rate, num_channels=1, test_mode="path")
 
     def test_save_gsm(self):
         self.assert_save_consistency("gsm", num_channels=1, test_mode="path")