cameraoverview.qdoc

Go to the documentation of this file.

// Copyright (C) 2021 The Qt Company Ltd.
// SPDX-License-Identifier: LicenseRef-Qt-Commercial OR GFDL-1.3-no-invariants-only
 
/*!
\page cameraoverview.html
\title Camera Overview
\brief Camera viewfinder, still image capture, and video recording.
\ingroup explanations-graphicsandmultimedia
 
The Qt Multimedia API provides a number of camera related classes, so you
can access images and videos from mobile device cameras or web cameras.
There are both C++ and QML APIs for common tasks.
 
\section1 Camera Features
 
In order to use the camera classes, a quick overview of the way a camera
works is needed. If you're already familiar with this, you can skip ahead to
\l {camera-tldr}{Camera implementation details}.
For a more detailed explanations of how a camera works, see the following YouTube
clip.
 
\youtube qS1FmgPVLqw
 
\section2 The Lens Assembly
 
At one end of the camera assembly is the lens assembly (one or
more lenses, arranged to focus light onto the sensor). The lenses
themselves can sometimes be moved to adjust things like focus and zoom. They
might also be fixed in an arrangement for a good balance between maintaining
focus and cost.
 
\image how-focus-works.gif "An animation of how focus works"
 
\image Zoom.gif "An animation of how zoom works."
 
Some lens assemblies can automatically be adjusted so that
an object at different distances from the camera can be kept in focus.
This is usually done by measuring how sharp a particular area of the
frame is, and then adjusting the lens assembly to find the peak sharpness. In
some cases, the camera will always use the center of the frame for this.
In other cases, a camera may also allow this target focus region to be specified.
Some examples of this feature include:
\list
\li Face zoom: Using computer vision to detect and use one or more faces as the
target.
\li Touch to zoom: Enabling the user to manually select an area via the preview
screen.
\endlist
 
\section2 The Sensor
Once light arrives at the sensor, it gets converted into digital pixels.
This process can depend on a number of things but ultimately comes down
to two things:
\list
\li The length of time conversion is allowed to take. Also known as exposure
time.
\li How bright the light is.
\endlist
 
The longer a conversion is allowed to take, the better the resulting image
quality. Using a flash can assist with letting more light hit the sensor,
allowing it to convert pixels faster, giving better quality for the same
amount of time. Conversely, allowing a longer conversion time can let you
take photos in darker environments, \b{as long as the camera is steady}. If the
camera moves while the sensor is recording, the resulting image is blurred.
 
\section2 Image Processing
After the image has been captured by the sensor, the camera firmware performs
various image processing tasks on it to compensate for various sensor
characteristics, current lighting, and desired image properties. Faster sensor
pixel conversion times may introduce digital noise, so some amount of image
processing can be done to remove this, based on the camera sensor settings.
 
The color of the image can also be adjusted at this stage to compensate for
different light sources - fluorescent lights and sunlight give very different
appearances to the same object, so the image can be adjusted based on the
white balance of the picture (due to the different color temperatures of the
light sources).
\image image_processing.png "5 examples of various image processing techniques."
 
Some forms of "special effects" can also be performed at this stage.  Black
and white, sepia, or "negative" style images can be produced.
 
\section2 Recording for Posterity
Finally, once a perfectly focused, exposed and processed image has been
created, it can be put to good use.  Camera images can be further processed
by application code (for example, to detect bar-codes, or to stitch together a
panoramic image), or saved to a common format like JPEG, or used to create a movie.
Many of these tasks have classes to assist them.
 
\target camera-tldr
\section1 Camera Implementation Details
\section2 Detecting and Selecting a Camera
 
Before using the camera APIs, you should check that a camera is available at
runtime. If there is none available, you could disable camera related features
in your application. To perform this check in C++, use the
\l QMediaDevices::videoInputs() function, as shown in the example below:
 
 
    \snippet multimedia-snippets/camera.cpp Camera overview check
 
Access a camera using the \l QCamera class in C++ or the \l Camera
type in QML.
 
When multiple cameras are available, you can specify which one to use.
 
In C++:
 
     \snippet multimedia-snippets/camera.cpp Camera selection
 
In QML, you can select the camera by setting the \l{Camera::cameraDevice} property.
You can also select a camera by its physical position on the system rather than
by camera info. This is useful on mobile devices, which often have a
front-facing and a back-facing camera.
 
In C++:
 
    \snippet multimedia-snippets/camera.cpp Camera overview position
 
 
In QML, you can set the \c Camera \l{Camera::cameraDevice}{cameraDevice} property.
Available cameras can be retrieved with MediaDevices.videoInputs
 
In QML:
 
\qml
Camera {
    position: Camera.FrontFace
}
\endqml
 
 
If both QCameraDevice and position aren't specified, the default camera
will be used. On desktop platforms, the default camera is set by the
user in the system settings. On a mobile device, the back-facing camera
is usually the default camera. You can get the default camera with
QMediaDevices::defaultVideoInput() or MediaDevices.defaultVideoInput
in QML.
 
\section2 Preview
 
While not strictly necessary, it's often useful to be able to see
what the camera is pointing at. This is known as a preview.
 
Depending on whether you're using QML or C++, you can do this in multiple ways.
In QML, you can use \l Camera and videoOutput together to monitor a
captureSession.
 
\qml
Item {
    VideoOutput {
        id: output
        anchors.fill: parent
    }
    CaptureSession {
        videoOutput: output
 
        Camera {
            // You can adjust various settings in here
        }
    }
}
\endqml
 
In C++, your choice depends on whether you are using widgets, or QGraphicsView.
The \l QVideoWidget class is used in the widgets case, and \l QGraphicsVideoItem
is useful for QGraphicsView.
 
    \snippet multimedia-snippets/camera.cpp Camera overview viewfinder
 
For advanced usage (like processing preview frames as they come, which enables
detection of objects or patterns), you can also use your own QVideoSink and set
that as the videoOutput for the QMediaCaptureSession. In this case, you will need to
render the preview image yourself by processing the data received from the
videoFrameChanged() signal.
 
    \snippet multimedia-snippets/camera.cpp Camera overview surface
 
On mobile devices, the preview image is by default oriented in the same way as the device.
Thus, as the user rotates the device, the preview image will switch between portrait and
landscape mode. Once you start recording, the orientation will be locked. To avoid a poor
user experience, you should also lock the orientation of the applications user interface
while recording. This can be achieved using the
\l{QWindow::contentOrientation}{contentOrientation} property of QWindow.
 
\section2 Still Images
 
After setting up a viewfinder and finding something photogenic, to capture an
image we need to initialize a new QImageCapture object. All that is then
needed is to start the camera and capture the image.
 
    \snippet multimedia-snippets/camera.cpp Camera overview capture
 
\section2 Movies
 
Previously we saw code that allowed the capture of a still image. Recording
video requires the use of a \l QMediaRecorder object.
 
To record video we need to create a camera object as before but this time as
well as creating a viewfinder, we will also initialize a media recorder object.
 
    \snippet multimedia-snippets/camera.cpp Camera overview movie
 
Signals from the \e QMediaRecorder can be connected to slots to react to
changes in the state of the encoding process or error events. Recording
starts when \l QMediaRecorder::record() is called. This causes the signal
\l{QMediaRecorder::}{recorderStateChanged()} to be emitted. Recording is
controlled by the record(), stop(), and pause() slots of QMediaRecorder.
 
\section2 Controlling the Imaging Pipeline
 
Now that the basics of capturing images and movies are covered, there are a number
of ways to control the imaging pipeline to implement some interesting techniques.
As explained earlier, several physical and electronic elements combine to determine
the final images, and you can control them with different classes.
 
\section3 Focus and Zoom
 
QCamera allows you to set the general focus policy by means of the
enums for the \l {QCamera::FocusMode}{FocusMode}. \l {QCamera::FocusMode}{FocusMode}
deals with settings such as \l {QCamera::FocusModeAuto},
and \l {QCamera::FocusModeInfinity}.
 
For camera hardware that supports it, \l QCamera::FocusModeAutoNear allows
imaging of things that are close to the sensor. This is useful in applications
like bar-code recognition, or business card scanning.
 
In addition to focus, QCamera allows you to control any available zoom
functionality using \l{QCamera::setZoomFactor}{setZoomFactor()} or
\l{QCamera::zoomTo}{zoomTo()}. The
available zoom range might be limited or entirely fixed to unity (1:1). The
allowed range can be checked with \l{QCamera::minimumZoomFactor}{minimumZoomFactor()}
and \l{QCamera::maximumZoomFactor}{maximumZoomFactor()}.
 
\section3 Exposure, Shutter Speed and Flash
 
There are a number of settings that affect the amount of light that hits the
camera sensor, and hence the quality of the resulting image.
 
The main settings for automatic image taking are the
\l {QCamera::ExposureMode}{exposure mode} and \l {QCamera::FlashMode}{flash mode}.
Several other settings (such as: ISO setting and exposure time) are usually
managed automatically, but can also be overridden if desired.
 
Finally, you can control the flash hardware (if present) using this class. In
some cases the hardware may also double as a torch.
 
\target camera_image_processing
\section3 Image Processing
 
The QCamera class lets you adjust the image processing part of the pipeline.
These settings include:
\list
    \li \l {QCamera::WhiteBalanceMode}{white balance}
    (also known as color temperature)
\endlist
 
Most cameras support automatic settings for all of these, so you shouldn't need
to adjust them unless the user wants a specific setting.
 
\section3 Canceling Asynchronous Operations
 
Various operations, such as image capture and auto focusing, occur asynchronously.
These operations can often be canceled by the start of a new operation, as long
as this is supported by the camera.
 
\section1 Examples
 
There are both C++ and QML examples available.
 
\section2 C++ Examples
 
\annotatedlist camera_examples
 
\section2 QML Examples
 
\annotatedlist camera_examples_qml
 
\section1 Reference Documentation
 
\section2 C++ Classes
 
\annotatedlist multimedia_camera
 
\section2 QML Types
 
\annotatedlist camera_qml
 
*/