| Message ID | 20250113133405.12167-7-mzamazal@redhat.com |
|---|---|
| State | Superseded |
| Headers | show |
| Series |
|
| Related | show |
Hi Milan, Thank you for the patch. On Mon, Jan 13, 2025 at 02:34:05PM +0100, Milan Zamazal wrote: > Report the exposure+gain in metadata. > > The exposure value is especially dubious because it should be in > microseconds but it's handled using V4L2_CID_EXPOSURE control, which > doesn't specify the unit, see > https://www.kernel.org/doc/html/latest/userspace-api/media/v4l/control.html. Please convert it to the right unit. > Signed-off-by: Milan Zamazal <mzamazal@redhat.com> > --- > src/ipa/simple/algorithms/agc.cpp | 7 ++++++- > 1 file changed, 6 insertions(+), 1 deletion(-) > > diff --git a/src/ipa/simple/algorithms/agc.cpp b/src/ipa/simple/algorithms/agc.cpp > index 72aade14..ba7e7adc 100644 > --- a/src/ipa/simple/algorithms/agc.cpp > +++ b/src/ipa/simple/algorithms/agc.cpp > @@ -11,6 +11,8 @@ > > #include <libcamera/base/log.h> > > +#include "control_ids.h" > + > namespace libcamera { > > LOG_DEFINE_CATEGORY(IPASoftExposure) > @@ -99,8 +101,11 @@ void Agc::process(IPAContext &context, > [[maybe_unused]] const uint32_t frame, > [[maybe_unused]] IPAFrameContext &frameContext, > const SwIspStats *stats, > - [[maybe_unused]] ControlList &metadata) > + ControlList &metadata) > { > + metadata.set(controls::ExposureTime, frameContext.sensor.exposure); > + metadata.set(controls::AnalogueGain, frameContext.sensor.gain); > + > /* > * Calculate Mean Sample Value (MSV) according to formula from: > * https://www.araa.asn.au/acra/acra2007/papers/paper84final.pdf
Hi Laurent, thank you for review. Laurent Pinchart <laurent.pinchart@ideasonboard.com> writes: > Hi Milan, > > Thank you for the patch. > > On Mon, Jan 13, 2025 at 02:34:05PM +0100, Milan Zamazal wrote: >> Report the exposure+gain in metadata. >> >> The exposure value is especially dubious because it should be in >> microseconds but it's handled using V4L2_CID_EXPOSURE control, which >> doesn't specify the unit, see >> https://www.kernel.org/doc/html/latest/userspace-api/media/v4l/control.html. > > Please convert it to the right unit. But what's the original unit? It's unspecified in the link above: V4L2_CID_EXPOSURE (integer) Exposure (cameras). [Unit?] Is there any better source of information? >> Signed-off-by: Milan Zamazal <mzamazal@redhat.com> >> --- >> src/ipa/simple/algorithms/agc.cpp | 7 ++++++- >> 1 file changed, 6 insertions(+), 1 deletion(-) >> >> diff --git a/src/ipa/simple/algorithms/agc.cpp b/src/ipa/simple/algorithms/agc.cpp >> index 72aade14..ba7e7adc 100644 >> --- a/src/ipa/simple/algorithms/agc.cpp >> +++ b/src/ipa/simple/algorithms/agc.cpp >> @@ -11,6 +11,8 @@ >> >> #include <libcamera/base/log.h> >> >> +#include "control_ids.h" >> + >> namespace libcamera { >> >> LOG_DEFINE_CATEGORY(IPASoftExposure) >> @@ -99,8 +101,11 @@ void Agc::process(IPAContext &context, >> [[maybe_unused]] const uint32_t frame, >> [[maybe_unused]] IPAFrameContext &frameContext, >> const SwIspStats *stats, >> - [[maybe_unused]] ControlList &metadata) >> + ControlList &metadata) >> { >> + metadata.set(controls::ExposureTime, frameContext.sensor.exposure); >> + metadata.set(controls::AnalogueGain, frameContext.sensor.gain); >> + >> /* >> * Calculate Mean Sample Value (MSV) according to formula from: >> * https://www.araa.asn.au/acra/acra2007/papers/paper84final.pdf
Hi Milan, On Mon, Jan 27, 2025 at 01:44:05PM +0100, Milan Zamazal wrote: > Laurent Pinchart writes: > > On Mon, Jan 13, 2025 at 02:34:05PM +0100, Milan Zamazal wrote: > >> Report the exposure+gain in metadata. > >> > >> The exposure value is especially dubious because it should be in > >> microseconds but it's handled using V4L2_CID_EXPOSURE control, which > >> doesn't specify the unit, see > >> https://www.kernel.org/doc/html/latest/userspace-api/media/v4l/control.html. > > > > Please convert it to the right unit. > > But what's the original unit? It's unspecified in the link above: > > V4L2_CID_EXPOSURE (integer) > Exposure (cameras). [Unit?] > > Is there any better source of information? The camera helpers in libipa provide functions to convert exposure time between time units and V4L2 control units. You can look at how rkisp1 handles that for instance. > >> Signed-off-by: Milan Zamazal <mzamazal@redhat.com> > >> --- > >> src/ipa/simple/algorithms/agc.cpp | 7 ++++++- > >> 1 file changed, 6 insertions(+), 1 deletion(-) > >> > >> diff --git a/src/ipa/simple/algorithms/agc.cpp b/src/ipa/simple/algorithms/agc.cpp > >> index 72aade14..ba7e7adc 100644 > >> --- a/src/ipa/simple/algorithms/agc.cpp > >> +++ b/src/ipa/simple/algorithms/agc.cpp > >> @@ -11,6 +11,8 @@ > >> > >> #include <libcamera/base/log.h> > >> > >> +#include "control_ids.h" > >> + > >> namespace libcamera { > >> > >> LOG_DEFINE_CATEGORY(IPASoftExposure) > >> @@ -99,8 +101,11 @@ void Agc::process(IPAContext &context, > >> [[maybe_unused]] const uint32_t frame, > >> [[maybe_unused]] IPAFrameContext &frameContext, > >> const SwIspStats *stats, > >> - [[maybe_unused]] ControlList &metadata) > >> + ControlList &metadata) > >> { > >> + metadata.set(controls::ExposureTime, frameContext.sensor.exposure); > >> + metadata.set(controls::AnalogueGain, frameContext.sensor.gain); > >> + > >> /* > >> * Calculate Mean Sample Value (MSV) according to formula from: > >> * https://www.araa.asn.au/acra/acra2007/papers/paper84final.pdf
diff --git a/src/ipa/simple/algorithms/agc.cpp b/src/ipa/simple/algorithms/agc.cpp index 72aade14..ba7e7adc 100644 --- a/src/ipa/simple/algorithms/agc.cpp +++ b/src/ipa/simple/algorithms/agc.cpp @@ -11,6 +11,8 @@ #include <libcamera/base/log.h> +#include "control_ids.h" + namespace libcamera { LOG_DEFINE_CATEGORY(IPASoftExposure) @@ -99,8 +101,11 @@ void Agc::process(IPAContext &context, [[maybe_unused]] const uint32_t frame, [[maybe_unused]] IPAFrameContext &frameContext, const SwIspStats *stats, - [[maybe_unused]] ControlList &metadata) + ControlList &metadata) { + metadata.set(controls::ExposureTime, frameContext.sensor.exposure); + metadata.set(controls::AnalogueGain, frameContext.sensor.gain); + /* * Calculate Mean Sample Value (MSV) according to formula from: * https://www.araa.asn.au/acra/acra2007/papers/paper84final.pdf
Report the exposure+gain in metadata. The exposure value is especially dubious because it should be in microseconds but it's handled using V4L2_CID_EXPOSURE control, which doesn't specify the unit, see https://www.kernel.org/doc/html/latest/userspace-api/media/v4l/control.html. Signed-off-by: Milan Zamazal <mzamazal@redhat.com> --- src/ipa/simple/algorithms/agc.cpp | 7 ++++++- 1 file changed, 6 insertions(+), 1 deletion(-)