[PATCH] gstreamer:Implement caps parsing for video/x-bayer

[Nicolas Dufresne]

Hi,

comment on the subject, you are missing a space after ":". Instead of saying
"caps parsing", I'd prefer to call this caps negotiation.

more comments below...

Le jeudi 31 octobre 2024 à 09:16 +0100, mailinglist1 at johanneskirchmair.de a
écrit :
> From: Johannes Kirchmair <johannes.kirchmair at skidata.com>
> 
> The parsing of video/x-bayer sources from string makes is possible to

                                                   makes *it* ?

> use cameras providing e.g SGRBG8 streams via gst-launch.
> 
> Like:
> gst-launch-1.0 libcamerasrc camera-name=<cam> ! video/x-bayer,format=grbg
> 
> Without this change the gstreamer plugin complains about "Unsupported
> media type: video/x-bayer".
> 
> Signed-off-by: Johannes Kirchmair <johannes.kirchmair at skidata.com>
> ---
>  src/gstreamer/gstlibcamera-utils.cpp | 90 ++++++++++++++--------------
>  1 file changed, 46 insertions(+), 44 deletions(-)
> 
> diff --git a/src/gstreamer/gstlibcamera-utils.cpp b/src/gstreamer/gstlibcamera-utils.cpp
> index 79f71246..472367f2 100644
> --- a/src/gstreamer/gstlibcamera-utils.cpp
> +++ b/src/gstreamer/gstlibcamera-utils.cpp
> @@ -254,54 +254,53 @@ gst_format_to_pixel_format(GstVideoFormat gst_format)
>  	return PixelFormat{};
>  }
>  
> +static struct {
> +	PixelFormat format;
> +	const gchar *name;
> +} bayer_formats[]{
> +	{ formats::SBGGR8, "bggr" },
> +	{ formats::SGBRG8, "gbrg" },
> +	{ formats::SGRBG8, "grbg" },
> +	{ formats::SRGGB8, "rggb" },
> +	{ formats::SBGGR10, "bggr10le" },
> +	{ formats::SGBRG10, "gbrg10le" },
> +	{ formats::SGRBG10, "grbg10le" },
> +	{ formats::SRGGB10, "rggb10le" },
> +	{ formats::SBGGR12, "bggr12le" },
> +	{ formats::SGBRG12, "gbrg12le" },
> +	{ formats::SGRBG12, "grbg12le" },
> +	{ formats::SRGGB12, "rggb12le" },
> +	{ formats::SBGGR14, "bggr14le" },
> +	{ formats::SGBRG14, "gbrg14le" },
> +	{ formats::SGRBG14, "grbg14le" },
> +	{ formats::SRGGB14, "rggb14le" },
> +	{ formats::SBGGR16, "bggr16le" },
> +	{ formats::SGBRG16, "gbrg16le" },
> +	{ formats::SGRBG16, "grbg16le" },
> +	{ formats::SRGGB16, "rggb16le" },
> +};
> +
> +#define ARRAY_SIZE(x) (sizeof(x) / sizeof(x[0]))
> +
>  static const gchar *
>  bayer_format_to_string(int format)
>  {
> -	switch (format) {
> -	case formats::SBGGR8:
> -		return "bggr";
> -	case formats::SGBRG8:
> -		return "gbrg";
> -	case formats::SGRBG8:
> -		return "grbg";
> -	case formats::SRGGB8:
> -		return "rggb";
> -	case formats::SBGGR10:
> -		return "bggr10le";
> -	case formats::SGBRG10:
> -		return "gbrg10le";
> -	case formats::SGRBG10:
> -		return "grbg10le";
> -	case formats::SRGGB10:
> -		return "rggb10le";
> -	case formats::SBGGR12:
> -		return "bggr12le";
> -	case formats::SGBRG12:
> -		return "gbrg12le";
> -	case formats::SGRBG12:
> -		return "grbg12le";
> -	case formats::SRGGB12:
> -		return "rggb12le";
> -	case formats::SBGGR14:
> -		return "bggr14le";
> -	case formats::SGBRG14:
> -		return "gbrg14le";
> -	case formats::SGRBG14:
> -		return "grbg14le";
> -	case formats::SRGGB14:
> -		return "rggb14le";
> -	case formats::SBGGR16:
> -		return "bggr16le";
> -	case formats::SGBRG16:
> -		return "gbrg16le";
> -	case formats::SGRBG16:
> -		return "grbg16le";
> -	case formats::SRGGB16:
> -		return "rggb16le";
> +	for (unsigned int i = 0; i < ARRAY_SIZE(bayer_formats); i++) {
> +		if ((uint32_t)bayer_formats[i].format == (uint32_t)format)

I'm curious why the choice of uint32 cast from enum and int, instead of typing
the format argument as PixelFormat ?

> +			return bayer_formats[i].name;

c++17 have this nice feature that let you do:

	for (const auto &format : bayer_formats) {
		....
	}

You can then drop ARRAY_SIZE. You map want to stay concistent with the
format_map naming and call this bayer_map.

>  	}
>  	return NULL;

Use nullptr, this one was missed during original review.

>  }
>  
> +static PixelFormat bayer_format_from_string(const gchar *name)
> +{
> +	for (unsigned int i = 0; i < ARRAY_SIZE(bayer_formats); i++) {

Same.

> +		if (strcmp(bayer_formats[i].name, name) == 0)
> +			return bayer_formats[i].format;
> +	}
> +	return PixelFormat{};
> +}
> +
>  static GstStructure *
>  bare_structure_from_format(const PixelFormat &format)
>  {
> @@ -407,9 +406,8 @@ gst_libcamera_stream_configuration_to_caps(const StreamConfiguration &stream_cfg
>  	return caps;
>  }
>  
> -void
> -gst_libcamera_configure_stream_from_caps(StreamConfiguration &stream_cfg,
> -					 GstCaps *caps)
> +void gst_libcamera_configure_stream_from_caps(StreamConfiguration &stream_cfg,
> +					      GstCaps *caps)

We have chosen to follow GStreamer coding style for this one, please refrain
from creating a mix and match through this file.

>  {
>  	GstVideoFormat gst_format = pixel_format_to_gst_format(stream_cfg.pixelFormat);
>  	guint i;
> @@ -469,11 +467,15 @@ gst_libcamera_configure_stream_from_caps(StreamConfiguration &stream_cfg,
>  		gst_structure_fixate_field_string(s, "format", format);
>  	}
>  
> +	printf("in the function!!!!!!");

Please cleanup your temporary debug traces before submitting patches.

>  	/* Then configure the stream with the result. */
>  	if (gst_structure_has_name(s, "video/x-raw")) {
>  		const gchar *format = gst_structure_get_string(s, "format");
>  		gst_format = gst_video_format_from_string(format);
>  		stream_cfg.pixelFormat = gst_format_to_pixel_format(gst_format);
> +	} else if (gst_structure_has_name(s, "video/x-bayer")) {
> +		const gchar *format = gst_structure_get_string(s, "format");
> +		stream_cfg.pixelFormat = bayer_format_from_string(format);
>  	} else if (gst_structure_has_name(s, "image/jpeg")) {
>  		stream_cfg.pixelFormat = formats::MJPEG;
>  	} else {

Your patch otherwise looks right to me, we clearly missed this while adding
Bayer support. Please send a V2 with the above fixes.

regards
Nicolas