feat(ios): append data on slice index change instead of save on end

cpp/rn-audioutils.cpp

@@ -3,41 +3,24 @@
 
 namespace rnaudioutils {
 
-std::vector<uint8_t> concat_short_buffers(const std::vector<short*>& buffers, const std::vector<int>& slice_n_samples) {
-    std::vector<uint8_t> output_data;
-
-    for (size_t i = 0; i < buffers.size(); i++) {
-        int size = slice_n_samples[i]; // Number of shorts
-        short* slice = buffers[i];
-
-        // Copy each short as two bytes
-        for (int j = 0; j < size; j++) {
-            output_data.push_back(static_cast<uint8_t>(slice[j] & 0xFF));         // Lower byte
-            output_data.push_back(static_cast<uint8_t>((slice[j] >> 8) & 0xFF));  // Higher byte
-        }
-    }
-
-    return output_data;
-}
-
-std::vector<uint8_t> remove_trailing_zeros(const std::vector<uint8_t>& audio_data) {
-    auto last = std::find_if(audio_data.rbegin(), audio_data.rend(), [](uint8_t byte) { return byte != 0; });
-    return std::vector<uint8_t>(audio_data.begin(), last.base());
+void append_wav_data(const short* data, const int n_samples, const std::string& file) {
+    std::ofstream output(file, std::ios::binary | std::ios::app);
+    output.write(reinterpret_cast<const char*>(data), n_samples * sizeof(short));
+    output.close();
 }
 
-void save_wav_file(const std::vector<uint8_t>& raw, const std::string& file) {
-    std::vector<uint8_t> data = remove_trailing_zeros(raw);
-
-    std::ofstream output(file, std::ios::binary);
+void add_wav_header_to_file(const std::string& file, const int data_size) {
+    std::ofstream output(file, std::ios::binary | std::ios::app);
 
     if (!output.is_open()) {
         RNWHISPER_LOG_ERROR("Failed to open file for writing: %s\n", file.c_str());
         return;
     }
 
-    // WAVE header
+    output.seekp(0, std::ios::beg);
+
     output.write("RIFF", 4);
-    int32_t chunk_size = 36 + static_cast<int32_t>(data.size());
+    int32_t chunk_size = 36 + static_cast<int32_t>(data_size);
     output.write(reinterpret_cast<char*>(&chunk_size), sizeof(chunk_size));
     output.write("WAVE", 4);
     output.write("fmt ", 4);
@@ -56,13 +39,10 @@ void save_wav_file(const std::vector<uint8_t>& raw, const std::string& file) {
     short bits_per_sample = 16;
     output.write(reinterpret_cast<char*>(&bits_per_sample), sizeof(bits_per_sample));
     output.write("data", 4);
-    int32_t sub_chunk2_size = static_cast<int32_t>(data.size());
+    int32_t sub_chunk2_size = static_cast<int32_t>(data_size);
     output.write(reinterpret_cast<char*>(&sub_chunk2_size), sizeof(sub_chunk2_size));
-    output.write(reinterpret_cast<const char*>(data.data()), data.size());
 
     output.close();
-
-    RNWHISPER_LOG_INFO("Saved audio file: %s\n", file.c_str());
 }
 
 } // namespace rnaudioutils

cpp/rn-audioutils.h

@@ -8,7 +8,7 @@
 
 namespace rnaudioutils {
 
-std::vector<uint8_t> concat_short_buffers(const std::vector<short*>& buffers, const std::vector<int>& slice_n_samples);
-void save_wav_file(const std::vector<uint8_t>& raw, const std::string& file);
+void append_wav_data(const short* data, const int n_samples, const std::string& file);
+void add_wav_header_to_file(const std::string& file, const int data_size);
 
 } // namespace rnaudioutils

cpp/rn-whisper.cpp

@@ -100,6 +100,7 @@ void job::put_pcm_data(short* data, int slice_index, int n_samples, int n) {
     for (int i = 0; i < n; i++) {
         pcm[i + n_samples] = data[i];
     }
+    pcm_data_size += n;
 }
 
 float* job::pcm_slice_to_f32(int slice_index, int size) {

cpp/rn-whisper.h

@@ -33,6 +33,7 @@ struct job {
     int audio_slice_sec = 0;
     float audio_min_sec = 0;
     const char* audio_output_path = nullptr;
+    int pcm_data_size = 0;
     std::vector<short *> pcm_slices;
     void set_realtime_params(vad_params vad, int sec, int slice_sec, float min_sec, const char* output_path);
     bool vad_simple(int slice_index, int n_samples, int n);

ios/RNWhisperContext.mm

@@ -232,10 +232,9 @@ void AudioInputCallback(void * inUserData,
 - (void)finishRealtimeTranscribe:(RNWhisperContextRecordState*) state result:(NSDictionary*)result {
     // Save wav if needed
     if (state->job->audio_output_path != nullptr) {
-        // TODO: Append in real time so we don't need to keep all slices & also reduce memory usage
-        rnaudioutils::save_wav_file(
-            rnaudioutils::concat_short_buffers(state->job->pcm_slices, state->sliceNSamples),
-            state->job->audio_output_path
+        rnaudioutils::add_wav_header_to_file(
+            state->job->audio_output_path,
+            state->job->pcm_data_size
         );
     }
     state->transcribeHandler(state->job->job_id, @"end", result);
@@ -284,6 +283,14 @@ - (void)fullTranscribeSamples:(RNWhisperContextRecordState*) state {
       state->nSamplesTranscribing == nSamplesOfIndex &&
       state->transcribeSliceIndex != state->sliceIndex
     ) {
+        if (state->job->audio_output_path != nullptr) {
+            rnaudioutils::append_wav_data(
+                state->job->pcm_slices[state->transcribeSliceIndex],
+                state->sliceNSamples[state->transcribeSliceIndex],
+                state->job->audio_output_path
+            );
+        }
+        // TODO: Clean up the previous slice
         state->transcribeSliceIndex++;
         state->nSamplesTranscribing = 0;
     }