What are the channel order and bit width of the PIXEL_FMT_LegacyRGBA format

[suchao]

There is a pixel format: “PIXEL_FMT_LegacyRGBA”. I have following questions.

1. What is the channel order?
2. How many total bits is this format?
3. How many bits does each channel have?

[JerryChang]

hello suchao,

LegacyRGBA it is RGBA-8-8-8-8 planar

Argus it processed CSI camera with bayer raw formats, it could be Raw10, Raw12…etc. it’s processed by ISP and convert as YUV buffer for post-processing, display rendering.

are you going to develop camera use-case with RGBA?
you’ll need to convert the formats, such as YUV to RGBA for your test-pipeline.
please see-also Topic 165300 to do format conversion.

[suchao]

Thank you for your reply.

I am trying to get RGB888 from the output stream. The closest one I can find is this LegacyRGBA. Do you mean there is no way to get RGB888 and I have to convert it from YUV?

[JerryChang]

hello suchao,

please refer to Camera Architecture Stack.
let’s assume you’re using a normal bayer raw camera sensor.
if you go though Argus, the output format is YUV. you must to have format conversion to get RGB888.

[suchao]

If I use EGLStream, do I still need to convert the format from YUV to RGB?

In the sample “jetson_multimedia_api/argus/samples/demosaicOutput”, it seems that it is possible to set the pixel format to “PIXEL_FMT_LegacyRGBA” and then save the output to a file. In this example, the output is after demosiacing and in RGBA-8-8-8-8 format, right?

[JerryChang]

hello suchao,

VIC output format only support RGB/BGR with alpha channel.
so, it needs to transformation of YUV to RGB Formats.