Clarification about RAW modes packed vs. unpacked

[cpixip]

The picamera2 manual has the following section:

However, the Pi 5 is different because it uses different raw stream formats. It supports:

• Compressed raw pixels. Here, camera samples are compressed to 8 bits per pixel using a visually lossless
compression scheme. This format is more memory efficient, but not suitable if an application wants access to the
raw pixels (because they would have to be decompressed).
• Uncompressed pixels. These are stored as one pixel per 16-bit word. It differs from the Pi 4 unpacked scheme in
that pixels are left-shifted so that the zero padding bits are at the least significant end (the earlier devices will pad
with zeros at the most significant end), and the full 16-bit dynamic range is used. These pixels have never been
through the compression scheme, and so give a bit-exact version of what came out of the sensor (allowing for the
shift that is applied).

So packed RAW formats are described as "visually lossless", and unpacked formats as "bit-exact". Does that mean that packed RAW formats are not bit-exact? That would be an important information for some uses of the picamera2 lib/Pi 5 hardware...

[cpixip]

Naushir at the RP camera forum answered basically my question:

The Bayer compression/decompression algorithm on Pi 5 hardware is visually lossless, i.e. it would be extremely difficult to pick out visual degradation when in use. This is on by default as it saves a large chunk of memory bandwidth and framebuffer allocation. However, it's easy enough to disable this compression entirely and generate the actual Bayer data coming from the camera, shifted left to make a 16-bit sample. This can be done by requesting an "unpacked" pixel format from rpicam-apps or picamera2.

The decompression algorithm itself can be found here:
https://github.com/raspberrypi/rpicam-a ... g.cpp#L218

So, including some own research: the RP5 uses by default the "compressed" format. This speeds up creating an output image via the libcamera or picamera2 software. But it slows down creating and saving raw image data. In fact, saving a full res raw of the HQ sensor takes about one second on a RP4 as well as on the RP5 in a default setup/picamera2. This is due to the RP5 working internally with the compressed format, but converting it to uncompressed format before saving to disk. If the RP5 commanded to work with uncompressed raw, it needs only 0.16 sec to save a .dng. And: this .dng will feature the full raw sensor information, whereas a .dng saved while working in compressed format will have a lesser quantization. The line

capture_config = picam2.create_still_configuration(raw={'format':'SRGGB12','size':(4056, 3040)})

will switch on the normal uncompressed raw format in case of the RP5.