CloudDisplayFFmpeg

Using clouddisplayencoder

The encoder process must be spawned with the following parameters:

./clouddisplayencoder DEST_IP DEST_PORT WIDTH HEIGHT PIX_FMT [AUD_FMT SAMPLE_RATE]

WIDTH and HEIGHT are in pixels and must be even numbers, preferably multiples of 16 to enable asm optimizations in FFmpeg.

PIX_FMT must be one of:

• ABGR8888 Alpha Blue Green Red, 8 bits per pixel
• ARGB8888 Alpha Red Green Blue, 8 bits per pixel
• BGR888 Blue Green Red, 8 bits per pixel
• BGRA8888 Blue Green Red Alpha, 8 bits per pixel
• RGB888 Red Green Blue, 8 bits per pixel
• RGBA8888 Red Green Blue Alpha, 8 bits per pixel

AUD_FMT must be one of:

• PCMF32LE PCM 32-bit floating-point little-endian
• PCMS16LE PCM signed 16-bit little-endian

SAMPLE_RATE is number of samples per second.

Both AUD_FMT and SAMPLE_RATE can be omitted if audio encoding is not desired.

Feeding data to clouddisplayencoder

Once the encoder is spawned, one must feed data for it via the standard in pipe using one of the commands listed bellow:

---

Bytes	Format	Description
0 - 3	'FRM\n'	Command for video frame
4 - 11	uint64_t	Capture timestamp in microseconds
12 -	uint8_t	Video data

IMPORTANT: Video data size is BYTES_PER_PIXEL * WIDTH * HEIGHT.

---

Bytes	Format	Description
0 - 3	'AUD\n'	Command for audio frame
4 - 11	uint64_t	Capture timestamp in microseconds
12 - 15	uint32_t	Number of samples
16 -	uint8_t	Audio data

IMPORTANT: Audio data is always stereo, so audio data size is BYTES_PER_SAMPLE * NUMBER_OF_SAMPLES * 2 CHANNELS. Samples from channels are interleaved, so DATA[i] is left channel and DATA[i + 1] is right channel for i % 2.