af_scaletempo2: calculate latency by center of search block

This changes the emitted pts values from the start of the search block
to the center of the search block. Change initial `output_time`
accordingly. Initial `search_block_index` is irrelevant, because it's
overwritten before the first iteration.

Using the `output_time` removes the rounding of `search_block_index`,
which also fixes the <20 microsecond gaps in timestamps between output
packets.

Rationale:

The variance in audio position was in the range `0..search-interval`.

With this change, the range is

    (- search-interval / 2)..(search-interval / 2)`

which ensures lower maximum offset.

2 files changed, 6 insertions, 6 deletions

diff --git a/audio/filter/af_scaletempo2_internals.c b/audio/filter/af_scaletempo2_internals.c
index eb2d4bb20b..a9d0fba136 100644
--- a/audio/filter/af_scaletempo2_internals.c
+++ b/audio/filter/af_scaletempo2_internals.c
@@ -707,7 +707,7 @@ int mp_scaletempo2_fill_buffer(struct mp_scaletempo2 *p,
             p->wsola_output_started = false;
 
             // sync audio precisely again
-            set_output_time(p, p->target_block_index + p->search_block_center_offset);
+            set_output_time(p, p->target_block_index);
             remove_old_input_frames(p);
         }
 
@@ -725,7 +725,7 @@ int mp_scaletempo2_fill_buffer(struct mp_scaletempo2 *p,
 
 double mp_scaletempo2_get_latency(struct mp_scaletempo2 *p, double playback_rate)
 {
-    return p->input_buffer_frames - p->search_block_index
+    return p->input_buffer_frames - p->output_time
         + p->num_complete_frames * playback_rate;
 }
 
@@ -749,7 +749,7 @@ void mp_scaletempo2_destroy(struct mp_scaletempo2 *p)
 void mp_scaletempo2_reset(struct mp_scaletempo2 *p)
 {
     p->input_buffer_frames = 0;
-    p->output_time = p->search_block_center_offset;
+    p->output_time = 0.0;
     p->search_block_index = 0;
     p->target_block_index = 0;
     // Clear the queue of decoded packets.
@@ -771,6 +771,7 @@ static void get_symmetric_hanning_window(int window_length, float* window)
 void mp_scaletempo2_init(struct mp_scaletempo2 *p, int channels, int rate)
 {
     p->muted_partial_frame = 0;
+    p->output_time = 0;
     p->search_block_center_offset = 0;
     p->search_block_index = 0;
     p->num_complete_frames = 0;
@@ -813,8 +814,6 @@ void mp_scaletempo2_init(struct mp_scaletempo2 *p, int channels, int rate)
         sizeof(float) * p->ola_window_size * 2);
     get_symmetric_hanning_window(2 * p->ola_window_size, p->transition_window);
 
-    p->output_time = p->search_block_center_offset;
-
     p->wsola_output_size = p->ola_window_size + p->ola_hop_size;
     p->wsola_output = realloc_2d(p->wsola_output, p->channels, p->wsola_output_size);
     // Initialize for overlap-and-add of the first block.
diff --git a/audio/filter/af_scaletempo2_internals.h b/audio/filter/af_scaletempo2_internals.h
index f1b74ec64f..622d6e20da 100644
--- a/audio/filter/af_scaletempo2_internals.h
+++ b/audio/filter/af_scaletempo2_internals.h
@@ -64,7 +64,8 @@ struct mp_scaletempo2 {
     double output_time;
     // The offset of the center frame of |search_block| w.r.t. its first frame.
     int search_block_center_offset;
-    // Index of the beginning of the |search_block|, in frames.
+    // Index of the beginning of the |search_block|, in frames. This may be
+    // negative, which is handled by |peek_audio_with_zero_prepend|.
     int search_block_index;
     // Number of Blocks to search to find the most similar one to the target
     // frame.