itak 4 → 5

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 3f5c18bc3e593097a8a4efac8f78432ccf2db6b7771ba740dae975f12b0f12f3
-  data.tar.gz: b8e1a1c9e26a8427116515243703e889171b6a5f49d4c42285a89b19b63749e9
+  metadata.gz: b561ff088fcc7f0280505863ba376e5d394c7cb470790141b0a799272b6db981
+  data.tar.gz: 761cb0d70df9bb1b9e9f6bee02b18540798639f054476221080f29fdb8595acd
 SHA512:
-  metadata.gz: d8fc23e257aa870f7337ef05e6f930f96713402881c2f3290dececdbf2792522a1d8f9f4394e1e00f43108bd13c5498ee4e141409cee65f470dfcd28d9c01f03
-  data.tar.gz: 45e7516e05441624c7871c900097f12472e450bf0d557e9a81e7ef0c0f8512f9cc5d59dad0a88d4090db3f31f36b3ba5fef1d32c53e8fffce4cb2cdd35b0c9ce
+  metadata.gz: 8ed97cece759284876c9dca21cc8077a9c1b9659d90ffd28953899d8f4565bbecb183999b115648a9fd554e58a0f5ecacaad1ddc686167d0a3bfa3a802634a7c
+  data.tar.gz: ab10d6c6bcb13ad0b29a1529447368364e92764d44943991de5f7a0fea77244ec1ff053d453648baa92aa9bb15be80f786d362d923e6de0ac72bb8174360480f

data/README.md

@@ -56,7 +56,7 @@ Though Itak is intended to be used as CLI, it provides some useful classes for e
 * `waveform`: Multi-dimensional array containing the waveform. Currently, it is a `Torch::Tensor`
 * `sample_rate`: Sample rate
 * `layout`: Audio layout. `:interleaved` or `:planar`
-* `channels`: The number of channels
+* `num_channels`: The number of channels
 
 *Note* that attributes are *not* kept consistent automatically. You are responsible for updating them when modifying the audio data:
 
@@ -65,6 +65,14 @@ audio.waveform = ...
 audio.sample_rate = ...
 ```
 
+It has attribute-like methods, too:
+
+* `num_frames`: The number of frames
+* `duration`: Duration in seconds
+* `mono?`: `true` when it's mono audio
+* `interleaved?`: `true` when it's interleaved audio
+* `planar?`: `true` when it's planar audio
+
 Also, it has methods to load and save audio:
 
 * `Audio.load(path)` loads audio from `path` and returns `Itak::Audio` instance.

data/itak.gemspec

@@ -1,6 +1,6 @@
 Gem::Specification.new do |s|
   s.name = "itak"
-  s.version = "4"
+  s.version = "5"
   s.authors = ["Kitaiti Makoto"]
   s.email = ["KitaitiMaokto@gmail.com"]
   s.summary = "Audio editing tool for podcasters"

data/lib/itak/audio.rb

@@ -19,18 +19,66 @@ class Itak
       @sample_rate = sample_rate
       @layout = layout
       @debug_info = debug_info
+
+      validate
     end
 
-    def channels
+    def num_channels
       shape = @waveform.shape
       return 1 if shape.length == 1
 
-      layout == :interleaved ? shape[-1] : shape[0]
+      interleaved? ? shape[-1] : shape[0]
+    end
+
+    def mono?
+      num_channels == 1
+    end
+
+    def num_frames
+      shape = @waveform.shape
+      return shape[0] if @waveform.ndim == 1
+
+      interleaved? ? shape[0] : shape[-1]
+    end
+
+    def duration
+      num_frames.to_f / @sample_rate
+    end
+
+    def interleaved?
+      @layout == :interleaved
+    end
+
+    def planar?
+      !interleaved?
     end
 
     def save(path)
       TorchAudio.save path.to_path, waveform, sample_rate
       self
     end
+
+    def inspect
+      str = "#<#{self.class}"
+      instance_variables.each do |iv|
+        if iv == :@waveform
+          str << " #{iv}=#{@waveform.class}(#{@waveform.shape})"
+        else
+          str << " #{iv}=#{instance_variable_get(iv).inspect}"
+        end
+      end
+      str << ">"
+    end
+
+    private
+
+    def validate
+      if @waveform.ndim > 2
+        raise ArgumentError, "Dimension of waveform must be 1 or 2, but #{waveform.ndim}"
+      end
+      unless [:interleaved, :planar].include? @layout
+        raise ArgumentError, "Layout must be :interleaved or :planar, but #{@layout.inspect}"
+      end
+    end
   end
 end

data/lib/itak/process/normalize_loudness.rb

@@ -5,13 +5,13 @@ class Itak
     class NormalizeLoudness
       prepend Processible
 
-      def initialize(target: -14.0, true_peak_ceiling: -0.1)
+      def initialize(target: -14.0, true_peak_ceiling: nil)
         @target = target
-        @true_peak_ceiling= true_peak_ceiling
+        @true_peak_ceiling = true_peak_ceiling
       end
 
       def run(audio)
-        channels = case audio.channels
+        channels = case audio.num_channels
                   in 1
                     [:center]
                   in 2

data/lib/itak/process/resample.rb

@@ -10,7 +10,7 @@ class Itak
 
       def run(audio)
         samples = NumoNArray(audio.waveform)
-        audio.waveform = Awaaz::Utils::Resample.read_and_resample(samples, audio.sample_rate, @to, audio.channels).to_ndav.to_torch_tensor
+        audio.waveform = Awaaz::Utils::Resample.read_and_resample(samples, audio.sample_rate, @to, audio.num_channels).to_ndav.to_torch_tensor
         audio.sample_rate = @to
       end
     end

data/lib/itak/process/vad.rb

@@ -22,8 +22,8 @@ class Itak
         chunks = []
         @vad.segments_from_samples(@params, src).each do |segment|
           segment => {start_time:, end_time:}
-          st = start_time * SAMPLE_RATE / 1000
-          en = end_time * SAMPLE_RATE / 1000
+          st = start_time * ML_SAMPLE_RATE / 1000
+          en = end_time * ML_SAMPLE_RATE / 1000
           chunks << src[st..en]
         end
         chunks.empty? ? src : Torch.cat(chunks)

data/lib/itak.rb

@@ -12,7 +12,8 @@ class Itak
   include NDAV::Converter
   include Process
 
-  SAMPLE_RATE = 16_000
+  ML_SAMPLE_RATE = 16_000
+  FINAL_SAMPLE_RATE = 44_1000
 
   def initialize(src, dest=nil)
     @src = Pathname(src)
@@ -33,11 +34,12 @@ class Itak
     Audio.load(@src, debug_info: {src: @src, dest: @dest})
       .(&NormalizeLoudness)
       .(&DownMix)
-      .(&Resample[to: SAMPLE_RATE])
+      .(&Resample[to: ML_SAMPLE_RATE])
       .(&Denoise)
       .(&VAD)
-      .>>(&Save[to: @dest])
       .>>(&Transcribe[dest: @transcription_dest])
+      .(&Resample[to: FINAL_SAMPLE_RATE])
+      .>>(&Save[to: @dest])
       .<<
 
     [@dest, @transcription_dest]

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: itak
 version: !ruby/object:Gem::Version
-  version: '4'
+  version: '5'
 platform: ruby
 authors:
 - Kitaiti Makoto