Merge branch 'ggerganov:master' into new-android-app

examples/stream/README.md

@@ -10,6 +10,23 @@ More info is available in [issue #10](https://github.com/ggerganov/whisper.cpp/i
 
 https://user-images.githubusercontent.com/1991296/194935793-76afede7-cfa8-48d8-a80f-28ba83be7d09.mp4
 
+## Slidingwindow mode
+
+Setting the `--step` argument to `0` enables the sliding window mode:
+
+```java
+ ./stream -m ./models/ggml-small.en.bin -t 6 --step 0 --length 30000 -vth 0.6
+```
+
+In this mode, the tool will transcribe only after some speech activity is detected. A very
+basic VAD detector is used, but in theory a more sophisticated approach can be added. The
+`-vth` argument determines the VAD threshold - higher values will make it detect silence more often.
+It's best to tune it to the specific use case, but a value around `0.6` should be OK in general.
+When silence is detected, it will transcribe the last `--length` milliseconds of audio and output
+a transcription block that is suitable for parsing.
+
+## Building
+
 The `stream` tool depends on SDL2 library to capture audio from the microphone. You can build it like this:
 
 ```bash

examples/stream/stream.cpp

@@ -14,6 +14,7 @@
 #include <thread>
 #include <vector>
 #include <fstream>
+#include <mutex>
 
 //  500 -> 00:05.000
 // 6000 -> 01:00.000
@@ -633,7 +634,7 @@ int main(int argc, char ** argv) {
                     const int64_t t0 = std::max(0.0, t1 - pcmf32.size()*1000.0/WHISPER_SAMPLE_RATE);
 
                     printf("\n");
-                    printf("### Transcription %d START | t0 = %lld ms | t1 = %lld ms\n", n_iter, t0, t1);
+                    printf("### Transcription %d START | t0 = %d ms | t1 = %d ms\n", n_iter, (int) t0, (int) t1);
                     printf("\n");
                 }