These functions handle screen and audio capture.
This part of FFscala is experimental, and support for your system can't be guaranteed. I can only guarantee it works under x11 Linux with Pulseaudio or Alsa.
Check the FFmpeg device documentation for more information on how these capture backends work and what you can do with them.
- x11grab
- (video) (x11)
- dshow
- (video and audio) (Windows)
- gdigrab
- (video) (Windows)
- pulse
- (audio) (pulseaudio)
- alsa
- (audio) (alsa)
- oss
- (audio) (oss)
- jack
- (audio) (jack)
- avfoundation
- (video and audio) (MacOS)
- android_camera
- (video) (Android)
x11, pulseaudio, alsa and oss are common backends among open source operating systems, especially systems of the Linux and BSD families. Wayland/pipewire screen capture is not yet available on FFmpeg, so Wayland users cannot make use of screen capture yet.
Remember to import ffscala.capture to use these functions.
For desktop and audio capture functionality, you need to give the functions the input device to capture. To know your input devices, check the function listSources().
def record(
output: String, captureargs: List[String], args: List[String] = List(), filters: List[String] = List(),
hwaccel: String = "", quiet: Boolean = true, exec: String = "ffmpeg"): IntThis executes FFmpeg and begins the desktop recording. output is for the output file's path.
captureargs should be the FFmpeg arguments generated by the functions seen here.
args are the encoding arguments used in functions like encode().
filters are the filter arguments used in functions like encode().
hwaccel sets a hardware acceleration encoding mode that needs extra configuration, like Intel QuickSync. Note that, if you use qsv hwaccel, you must use the qsv encoder with the nv12 or qsv pixel format. QSV is only supported by Intel GPUs.
Supported hwaccel values:
- qsv
To stop the recording, type Q. You can also stop the recording with ctrl + C, but this will also shut down your program. You can alternatively add a recording duration with setDuration() (see Base doc).
def takeScreenshot(
output: String, captureargs: List[String], args: List[String] = List(),
filters: List[String] = List(), quiet: Boolean = true, exec: String = "ffmpeg"): IntTakes a screenshot and saves it at path output. Behavior is identical to record(), but only 1 frame is captured.
captureargs should be the FFmpeg arguments generated by the functions seen here.
args are the encoding arguments used in functions like encode().
filters are the filter arguments used in functions like encode().
def takeScreenshot_auto(mode: String, i: String, output: String, showmouse: Boolean = false,
args: List[String] = List(), filters: List[String] = List(),
quiet: Boolean = true, exec: String = "ffmpeg"): IntTakes a screenshot and saves it at path output. This function sets the screen/window capture arguments automatically based on the mode you choose, as a more convenient alternative to takeScreenshot().
showmouse tells to either include the mouse in the screen capture or not
Mode is the screen capture method.
All the encoding parameter arguments can be added through args. This is very useful for most image formats that are not PNG and require some manual encoding configuration to suit well your necessities.
The i is input screen. This varies between capture modes.
Supported desktop capture modes:
- x11grab
- dshow
- dshow_video
- gdigrab
def x11grab_captureVideo(input: String, fps: Int, showmouse: Boolean = true, width: Int = 0, height: Int = 0, x: Int = 0, y: Int = 0): List[String]Uses x11grab for video capture, for any OS that uses an x11 server.
input sets the display and screen to capture. Use "0.0" to use display 0 and screen 0, probably the ones you want to capture.
fps is the capture frame rate.
showmouse tells x11grab to whether show the mouse cursor in the capture.
width and height set the dimensions of the screen capture. Default is the native size of the screen you are capturing, and lower values than that will crop.
The arguments x y set the position offset of the capture, starting at the top left.
def dshow_captureVideo(input: String, fps: Int, width: Int = 0, height: Int = 0): List[String]Uses DirectShow for video capture, for Windows systems.
input sets the input video device to capture. Check listSources() to know which you have in your system.
fps is the capture frame rate.
width and height set the dimensions of the screen capture. Default is the native size of the screen you are capturing, and lower values than that will crop.
def gdigrab_captureVideo(fps: Int, showmouse: Boolean = true, input: String = "", width: Int = 0, height: Int = 0, x: Int = 0, y: Int = 0): List[String]Uses Gdigrab for video capture, for Windows systems.
fps is the capture frame rate.
showmouse tells Gdigrab to whether show the mouse cursor in the capture.
input sets the input video device to capture. By default it is set to "", which will capture the whole screen (equivalent to "desktop"). Set it to the name of a program window to capture that window instead.
width and height set the dimensions of the screen capture. Default is the native size of the screen/window you are capturing, and lower values than that will crop.
The arguments x y set the position offset of the capture, starting at the top left.
def dshow_captureAudio(input: String, samplerate: Int, channels: Byte, depth: Byte): List[String]Uses DirectShow for audio capture, for Windows systems.
input sets the input audio device to capture. Check listSources() to know which you have in your system.
sampleraate is the audio sample rate. Minimum value accepted is 1.
channels sets the amount of audio channels. Minimum value accepted is 1.
depth sets the audio bit depth. Minimum value accepted is 8.
def avfoundation_capture(vinput: String, ainput: String, fps: Int, showmouse: Boolean = true, width: Int = 0, height: Int = 0): List[String]Uses Avfoundation for video capture, for MacOS systems.
vinput and ainput are respectively the input devices for your video and audio capture. Use index numbers, starting from 0. Leave them at "" to use the system's defaults. Set them to "none" to disable their capture.
fps is the capture frame rate.
showmouse tells Avfoundation to whether show the mouse cursor in the capture.
width and height set the dimensions of the screen capture. Default is the native size of the screen/window you are capturing, and lower values than that will crop.
def captureVideo(mode: String, i: String, fps: Int): List[String]Returns the configuration to capture your screen. It's preferred you use the functions above instead, as they are specifically tailored for their respective capture backend, while this function is more generic and supports less features.
fps sets the frame rate for the video capture
The i is input screen. This varies between capture modes.
You can specify a positive offset for the screen's position with x and y
Supported desktop capture modes:
- x11grab
- dshow
- gdigrab
For x11grab, the default input is "0.0", which will capture the whole main screen.
def pulse_captureAudio(input: String, samplerate: Int = 0, channels: Byte = 0): List[String]def alsa_captureAudio(input: String, samplerate: Int = 0, channels: Byte = 0): List[String]def oss_captureAudio(input: String, samplerate: Int = 0, channels: Byte = 0): List[String]def jack_captureAudio(input: String, channels: Byte = 0): List[String]Use pulseaudio, alsa, oss or jack to capture audio.
input is the input audio device.
samplerate and channels are the audio sample rate and the amount of audio channels, respectively. For both arguments, if the value is 0 or lower, FFmpeg assumes the sample rate and channel count.
The jack function doesn't take a sample rate argument.
def captureAudio(mode: String, input: String, ch: Byte = 2, rate: Int = 48000): List[String]Returns the configuration to capture audio. Just like captureVideo(), it's preferred you use the functions above instead, as they are specifically tailored for their respective capture backend, while this function is more generic and supports less features.
Mode is the screen capture method. Right now only alsa or pulse are supported.
The input is input audio device. This varies between capture modes.
ch is the channel count. By default, the capture is set to stereo with 2 channels, you can change this if you want.
rate is the audio sample rate, defaults to 48000.
Supported:
- alsa
- pulse
- oss
- dshow
For Linux, it's recommended to use pulse.
def androidcamera_capture(input: Byte, fps: Int, width: Int, height: Int): List[String]Uses android_camera to capture the video of your android phone's cameras.
input Is an index number. 0 tends to be your main back-facing camera.
fps is the capture frame rate.
width and height set the resolution of the capture. If the given dimensions are not available in your Android camera, then it will fallback to the default setting.
def addTracks(amt: Int): List[String]This maps multiple tracks to be included in the final file. If you capture multiple tracks of the same type (video, audio), you need to use this function so the tracks are properly set up.
Let's imagine you want to record your desktop screen, your desktop audio and your microphone audio. Since you are capturing 2 audio sources, being of the same type, you need to add 3 tracks with addTracks(3).
def listSources(mode: String, full: Boolean = false, exec: String = "ffmpeg"): List[String]This function calls FFmpeg to retrieve the currently available audio sources from the mode capture method.
full is set to false by default, which means the retrieved audio source strings are the source IDs that can be used directly on audio capture backends. If it's set to true, the full name of the source is displayed. This won't happen if mode is dshow or avfoundation.
Supported:
- alsa
- pulse
- oss
- dshow_video
- dshow_audio
- dshow
- avfoundation
- all
For all, all sources that can be listed with ffmpeg -sources will be listed and not parsed or processed.
def listSources_OSS(): Array[String]Lists all available devices for the OSS audio system. Use this instead of listSources() if it works better for you.