- 15 Apr, 2015 13 commits
-
-
Francois Cartegnie authored
Fixes alaw regression. /samples/A-codecs/aLaw/ Arbitrary limit still required for compressed format(AMR). Now max to read 25ms of samples.
-
Rémi Denis-Courmont authored
-
Thomas Guillem authored
If defined, getTimestamp (since Android 4.4) or getPlaybackHeadPosition + getLatency (since Android 4.3) will be used to get the AudioTrack position.
-
Thomas Guillem authored
-
Thomas Guillem authored
Since Android 4.3 there is a hidden method that retrieve the audio hardware latency: AudioSystem.getOutputLatency() (that should not be used according to the android comments). This latency should not be used when we use AudioTrack.getTimestamp since it already take into account the hardware latency. So why adding a new method ? Because AudioTrack.getTimestamp don't work with Bluetooth (it returns false). This fix lip sync with Bluetooth audio.
-
Thomas Guillem authored
It improves delay precision for old devices.
-
Thomas Guillem authored
Move AudioTimestamp variables into a separate struct. Move AudioTimestamp calculation into a separate function. Call getTimestamp every 500ms, keep the previous result in cache.
-
Thomas Guillem authored
Since it only reset PlaybackHeadPosition. Move i_samples_written initialisation out of this function.
-
Thomas Guillem authored
It allows to play more than 13h30min (@44.1kHz). Add more comments for the buggy getPlaybackHeadPosition method.
-
Francois Cartegnie authored
refs samples/A-codecs/Creative/
-
Francois Cartegnie authored
refs samples/A-codecs/CreativeADPCM8bit/
-
Francois Cartegnie authored
-
Christoph Miebach authored
-
- 14 Apr, 2015 18 commits
-
-
Francois Cartegnie authored
Some muxer is trowing incorrect rate/scale and then duration. As we're not computing pts from fully decoded samples, we need a correct pts for avcodec input. Fix is possible since ADPCM uses a fixed bits/sample and header.
-
Jean-Baptiste Kempf authored
-
Jean-Baptiste Kempf authored
-
Jean-Baptiste Kempf authored
-
Jean-Baptiste Kempf authored
-
Jean-Baptiste Kempf authored
-
Jean-Baptiste Kempf authored
-
Jean-Baptiste Kempf authored
-
Jean-Baptiste Kempf authored
-
Jean-Baptiste Kempf authored
-
Jean-Baptiste Kempf authored
-
Jean-Baptiste Kempf authored
-
Francois Cartegnie authored
Was going into infinite loop in that case. refs #14389
-
Jean-Baptiste Kempf authored
-
Francois Cartegnie authored
-
Jean-Baptiste Kempf authored
This mostly reverts commit f0380ff0 and fixes it for Win32, but also Linux
-
Francois Cartegnie authored
-
Jean-Baptiste Kempf authored
-
- 13 Apr, 2015 9 commits
-
-
Francois Cartegnie authored
Repicked pcr might be on a track that lacks of pcr field.
-
Francois Cartegnie authored
Avoids breaking with pcr on separate pid
-
Francois Cartegnie authored
refactorings... :/
-
Tristan Matthews authored
-
Francois Cartegnie authored
AMR is declared as V0, uncompressed chunk. Reading a max of 1024 samples in that case is too much.
-
Francois Cartegnie authored
also reverts dad98149
-
Thomas Guillem authored
-
Jean-Baptiste Kempf authored
It is present in the .zip and .7zip since forever It takes bandwidth and space that is not useful.
-
Thomas Guillem authored
Don't abort DecodeVideo when there is no p_block. It allows to fetch more than one output buffer consequently.
-