summaryrefslogtreecommitdiff
path: root/src/ipa/rpi
AgeCommit message (Collapse)Author
2023-09-16ipa: rpi: agc: Implementation of multi-channel AGCDavid Plowman
The switchMode, prepare and process methods are updated to implement multi-channel AGC correctly: * switchMode now invokes switchMode on all the channels (whether active or not). * prepare must find what channel the current frame is, and run on behalf of that channel. * process updates the most recent DeviceStatus and statistics for the channel of the frame that has just arrived, but generates updated values working through the active channels in round-robin fashion. One minor detail in process is that we don't want to change the DeviceStatus metadata of the current frame, so we now pass this to the AgcChannel's process method, rather than letting it find the DeviceStatus in the metadata. Signed-off-by: David Plowman <david.plowman@raspberrypi.com> Reviewed-by: Naushir Patuck <naush@raspberrypi.com> Reviewed-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com> Signed-off-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com>
2023-09-16ipa: rpi: agc: Reorganise code for multi-channel AGCDavid Plowman
This commit does the basic reorganisation of the code in order to implement multi-channel AGC. The main changes are: * The previous Agc class (in agc.cpp) has become the AgcChannel class in (agc_channel.cpp). * A new Agc class is introduced which is a wrapper round a number of AgcChannels. * The basic plumbing from ipa_base.cpp to Agc is updated to include a channel number. All the existing controls are hardwired to talk directly to channel 0. There are a couple of limitations which we expect to apply to multi-channel AGC. We're not allowing different frame durations to be applied to the channels, nor are we allowing separate metering modes. To be fair, supporting these things is not impossible, but there are reasons why it may be tricky so they remain "TBD" for now. This patch only includes the basic reorganisation and plumbing. It does not yet update the important methods (switchMode, prepare and process) to implement multi-channel AGC properly. This will appear in a subsequent commit. For now, these functions are hard-coded just to use channel 0, thereby preserving the existing behaviour. Signed-off-by: David Plowman <david.plowman@raspberrypi.com> Reviewed-by: Naushir Patuck <naush@raspberrypi.com> Reviewed-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com> Signed-off-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com>
2023-09-16ipa: rpi: histogram: Add interBinMean()Naushir Patuck
Add a new helper function Histogram::interBinMean() that essentially replaces the existing Histogram::interQuantileMean() logic but working on bins instead. Rework the interQuantileMean() to call into interBinMean() with the appropriate convertion from quatiles to bins. Signed-off-by: Naushir Patuck <naush@raspberrypi.com> Reviewed-by: David Plowman <david.plowman@raspberrypi.com> Reviewed-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com> Signed-off-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com>
2023-09-04ipa: rpi: vc4: Make the StatisticsPtr construction clearerDavid Plowman
StatisticsPtr is a shared pointer, so the use of std::make_unique to create it was a bit confusing. Use std::make_shared instead. Signed-off-by: David Plowman <david.plowman@raspberrypi.com> Reviewed-by: Naushir Patuck <naush@raspberrypi.com> Reviewed-by: Kieran Bingham <kieran.bingham@ideasonboard.com> Signed-off-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
2023-09-04ipa: rpi: agc: Split AgcStatus into AgcStatus and AgcPrepareStatusDavid Plowman
The Agc::process() function returns an AgcStatus object in the metadata as before, but Agc::prepare() is changed to return the values it computes in a separate AgcPrepareStatus object (under the new tag "agc.prepare_status"). The "digitalGain" and "locked" fields are moved from AgcStatus to AgcPrepareStatus. This will be useful going forward as we can be more flexible about the order in which prepare() and process() are called, without them trampling on each other's results. Signed-off-by: David Plowman <david.plowman@raspberrypi.com> Reviewed-by: Naushir Patuck <naush@raspberrypi.com> Reviewed-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com> Signed-off-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
2023-09-04ipa: rpi: agc: Filter exposures before dealing with digital gainDavid Plowman
We now time-filter the exposure before sorting out how much digital gain is required. This is actually a little more natural and simplifies the code. It also prepares us for some future work where this arrangement will be helpful. Signed-off-by: David Plowman <david.plowman@raspberrypi.com> Reviewed-by: Naushir Patuck <naush@raspberrpyi.com> Reviewed-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com> Signed-off-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
2023-09-04ipa: rpi: agc: Fetch AWB status in process method, not prepareDavid Plowman
prepare() doesn't use the AWB status, so fetching it in process() is probably better. This change is preparatory to other changes, where we may find ourselves calling process() without having called prepare() previously. Signed-off-by: David Plowman <david.plowman@raspberrypi.com> Reviewed-by: Naushir Patuck <naush@raspberrypi.com> Reviewed-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com> Signed-off-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
2023-08-29ipa: rpi: imx290: Hide one frame on startupNaushir Patuck
The imx290 produces a single unusable frame on startup and mode switch. This is signalled to the IPA in the mode switch case, but not the startup case. Fix this. Signed-off-by: Naushir Patuck <naush@raspberrypi.com> Reviewed-by: David Plowman <david.plowman@raspberrypi.com> Reviewed-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com> Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
2023-07-28ipa: rpi: Fix the reporting of Focus FoMsDavid Plowman
The FocusFom metadata was no longer being reported back because the "focus.status" metadata was never being created. Additionally, the scaling of the focus FoMs was over-zealous, rounding just about everything down to zero. Fixes: ac7511dc4c59 ("ipa: raspberrypi: Generalise the focus reporting code") Signed-off-by: David Plowman <david.plowman@raspberrypi.com> Reviewed-by: Naushir Patuck <naush@raspberrypi.com> Reviewed-by: Kieran Bingham <kieran.bingham@ideasonboard.com> Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com> Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
2023-07-28ipa: rpi: vc4: data: Updated color matrices for RPi CamerasBen Benson
Altered the color matrices for the tuning files for various cameras in order to make them more color accurate. Signed-off-by: Ben Benson <ben.benson@raspberrypi.com> Reviewed-by: David Plowman <david.plowman@raspberrypi.com> Reviewed-by: Naushir Patuck <naush@raspberrypi.com> Signed-off-by: Naushir Patuck <naush@raspberrypi.com>
2023-07-25ipa: rpi: common: Handle AEC/AGC flicker controlsDavid Plowman
We handle the flicker modes by passing the correct period to the AEC/AGC algorithm which already contains the necessary code. The "Auto" mode, as well as reporting the detected flicker period via the "AeFlickerDetected" metadata, are unsupported for now. Signed-off-by: David Plowman <david.plowman@raspberrypi.com> Reviewed-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com> Reviewed-by: Umang Jain <umang.jain@ideasonboard.com> Signed-off-by: Naushir Patuck <naush@raspberrypi.com>
2023-07-10ipa: rpi: imx296_mono: Disable all colour shadingNaushir Patuck
Colour shading should be turned off for the mono variant of the IMX296 sensor. To do this, set all the colour gain coefficients to 1. Signed-off-by: Naushir Patuck <naush@raspberrypi.com> Reviewed-by: David Plowman <daivd.plowman@raspberrypi.com> Reviewed-by: Nick Hollinghurst <nick.hollinghurst@raspberrypi.com> Signed-off-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
2023-07-10ipa: rpi: imx708: Fix mode switch drop frame countNaushir Patuck
The imx708 must drop a single frame on startup - but only when in HDR mode. Non-HDR modes do not need to drop frames. Fix the logic in hideFramesModeSwitch() which currently unconditionally advertises to drop one frame. Unfortunately there is no clear way to tell if the sensor is in the HDR mode. So for now, look the resolution and framerate to deduce this. Additionally ensure we override hideFramesStartup() and return the same number as hideFramesModeSwitch(). Bug: https://github.com/raspberrypi/libcamera-apps/issues/524 Signed-off-by: Naushir Patuck <naush@raspberrypi.com> Reviewed-by: David Plowman <david.plowman@raspberrypi.com> Reviewed-by: Kieran Bingham <kieran.bingham@ideasonboard.com> Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com> Signed-off-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
2023-06-19ipa: rpi: agc: Do not switch to a default if a mode is unavailableNaushir Patuck
In commit 0ee9339331c6, a default metering/exposure/constraint mode is used if a control sets a mode that is not listed in the camera tuning file. Setting a default mode may be undesirable in these cases, so instead keep the agc mode unchanged. This also matches the behaviour for other IPA controls where no changes are made in error conditions. Fixes: 0ee9339331c6 ("ipa: rpi: agc: Gracefully handle missing agc modes") Signed-off-by: Naushir Patuck <naush@raspberrypi.com> Reviewed-by: David Plowman <david.plowman@raspberrypi.com>
2023-06-07ipa: rpi: agc: Gracefully handle missing agc modesNaushir Patuck
If a metering/exposure/constraint mode is not listed in the sensor tuning file, and a control for the missing mode is set on the agc, we terminate the application with a fatal log message. Instead of this fatal termination, log a warning message and switch to the appropriate default mode so that the application continues running. Bug: https://github.com/raspberrypi/libcamera/issues/59 Bug: https://github.com/ayufan/camera-streamer/issues/67 Signed-off-by: Naushir Patuck <naush@raspberrypi.com> Reviewed-by: Kieran Bingham <kieran.bingham@ideasonboard.com> Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com> Signed-off-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
2023-06-07ipa: rpi: agc: Use std::string instead of char arraysNaushir Patuck
Replace the char array strings in struct AgcStatus with std::string objects. This simplifies the string handling in the source code. Signed-off-by: Naushir Patuck <naush@raspberrypi.com> Reviewed-by: Kieran Bingham <kieran.bingham@ideasonboard.com> Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com> Signed-off-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
2023-06-07ipa: rpi: tunings: Add missing short and long exposure profilesNaushir Patuck
Some sensor tuning files were missing the short and/or long exposure mode profiles. Add the missing items to the relevant files. As a drive-by, rename the "sport" exposure profile to "short" for the IMX290 tuning. Signed-off-by: Naushir Patuck <naush@raspberrypi.com> Reviewed-by: Kieran Bingham <kieran.bingham@ideasonboard.com> Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com> Signed-off-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
2023-06-06ipa: rpi: Set lens position to hyperfocal on startupNaushir Patuck
On the first ipa->configure() call, set the lens position (if a lens is present) to the default position. Typically this would be the hyperfocal position based on the tuning data. Signed-off-by: Naushir Patuck <naush@raspberrypi.com> Reviewed-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com> Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com> Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
2023-06-06pipeline: ipa: rpi: Return lens controls from ipa->configure()Naushir Patuck
In addition to sensor controls, return lens controls from IpaBase::configure() back to the pipeline handler. If there are lens controls present in the ControlList, action them in the pipeline handler as part of the configure routine. Signed-off-by: Naushir Patuck <naush@raspberrypi.com> Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com> Reviewed-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com> Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
2023-06-05ipa: rpi: Handle controls for mono variant sensorsNaushir Patuck
Do not advertise colour related controls (i.e. [A]WB, colour saturation) in the ControlInfoMap of available controls returned out to the application. Silently ignore these controls in the control handler in case applications don't use the advertised ControlInfoMap to validate controls. As a drive-by fix, don't advertise controls::ColourCorrectionMatrix in the ControlInfoMap as it is not handled by the IPA. Signed-off-by: Naushir Patuck <naush@raspberrypi.com> Reviewed-by: David Plowman <david.plowman@raspberrypi.com> Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com> Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
2023-06-02ipa: rpi: imx296: Small refinements to the IMX296 mono sensor tuningNaushir Patuck
Update the noise profile and reference lux level with newly taken calibration pictures. Remove GEQ/cross-talk processing block as this is a mono sensor. Signed-off-by: Naushir Patuck <naush@raspberrypi.com> Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com> Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
2023-05-04pipeline: raspberrypi: Introduce PipelineHandlerBase classNaushir Patuck
Create a new PipelineHandlerBase class that handles general purpose housekeeping duties for the Raspberry Pi pipeline handler. Code the implementation of new class is essentially pulled from the existing pipeline/rpi/vc4/raspberrypi.cpp file with a small amount of refactoring. Create a derived PipelineHandlerVc4 class from PipelineHandlerBase that handles the VC4 pipeline specific tasks of the pipeline handler. Again, code for this class implementation is taken from the existing pipeline/rpi/vc4/raspberrypi.cpp with a small amount of refactoring. The goal of this change is to allow third parties to implement their own pipeline handlers running on the Raspberry Pi without duplicating all of the pipeline handler housekeeping tasks. Signed-off-by: Naushir Patuck <naush@raspberrypi.com> Reviewed-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com> Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
2023-05-04ipa: raspberrypi: agc: Move weights out of AGCDavid Plowman
The region weights for the the AGC zones are handled by the AGC algorithm. Apply them directly in the IPA (vc4.cpp) to the statistics that we pass to the AGC. Signed-off-by: David Plowman <david.plowman@raspberrypi.com> Signed-off-by: Naushir Patuck <naush@raspberrypi.com> Reviewed-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com> Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
2023-05-04ipa: raspberrypi: Introduce IpaBase classNaushir Patuck
Create a new IpaBase class that handles general purpose housekeeping duties for the Raspberry Pi IPA. The implementation of the new class is essentially pulled from the existing ipa/rpi/vc4/raspberrypi.cpp file with a minimal amount of refactoring. Create a derived IpaVc4 class from IpaBase that handles the VC4 pipeline specific tasks of the IPA. Again, code for this class implementation is taken from the existing ipa/rpi/vc4/raspberrypi.cpp with a minimal amount of refactoring. The goal of this change is to allow third parties to implement their own IPA running on the Raspberry Pi without duplicating all of the IPA housekeeping tasks. Signed-off-by: Naushir Patuck <naush@raspberrypi.com> Reviewed-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com> Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
2023-05-04pipeline: ipa: raspberrypi: Restructure the IPA mojom interfaceNaushir Patuck
Restructure the IPA mojom interface to be more consistent in the use of the API. Function parameters are now grouped into *Params structures and results are now returned in *Results structures. The following pipeline -> IPA interfaces have been removed: signalQueueRequest(libcamera.ControlList controls); signalIspPrepare(ISPConfig data); signalStatReady(uint32 bufferId, uint32 ipaContext); and replaced with: prepareIsp(PrepareParams params); processStats(ProcessParams params); signalQueueRequest() is now encompassed within prepareIsp(). The following IPA -> pipeline interfaces have been removed: runIsp(uint32 bufferId); embeddedComplete(uint32 bufferId); statsMetadataComplete(uint32 bufferId, libcamera.ControlList controls); and replaced with the following async calls: prepareIspComplete(BufferIds buffers); processStatsComplete(BufferIds buffers); metadataReady(libcamera.ControlList metadata); Signed-off-by: Naushir Patuck <naush@raspberrypi.com> Reviewed-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com> Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
2023-05-04pipeline: ipa: raspberrypi: Refactor and move the Raspberry Pi codeNaushir Patuck
Split the Raspberry Pi pipeline handler and IPA source code into common and VC4/BCM2835 specific file structures. For the pipeline handler, the common code files now live in src/libcamera/pipeline/rpi/common/ and the VC4-specific files in src/libcamera/pipeline/rpi/vc4/. For the IPA, the common code files now live in src/ipa/rpi/{cam_helper,controller}/ and the vc4 specific files in src/ipa/rpi/vc4/. With this change, the camera tuning files are now installed under share/libcamera/ipa/rpi/vc4/. To build the pipeline and IPA, the meson configuration options have now changed from "raspberrypi" to "rpi/vc4": meson setup build -Dipas=rpi/vc4 -Dpipelines=rpi/vc4 Signed-off-by: Naushir Patuck <naush@raspberrypi.com> Reviewed-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com> Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com> Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>