mirror of
https://github.com/home-assistant/home-assistant.io.git
synced 2025-07-19 15:26:59 +00:00
Merge pull request #939 from home-assistant/ffmpeg_update
Ffmpeg component
This commit is contained in:
Pascal Vizeli
GitHub
commit
7a9c442973
@ -16,10 +16,6 @@ ha_iot_class: "Local Polling"
|
||||
|
||||
The `ffmpeg` platform allows you to use every video or audio feed with [FFmpeg](http://www.ffmpeg.org/) for various sensors in Home Assistant. Available are: **noise**, **motion**. If the `ffmpeg` process is broken, the sensor will be unavailable. To restart the instance, use the service *binary_sensor.ffmpeg_restart*.
|
||||
|
||||
<p class='note'>
|
||||
You need the `ffmpeg` binary in your system path. On Debain 8 you can install it from backports. If you want Hardware support on a Raspberry Pi you need to build it from source. Windows binary are avilable on [FFmpeg](http://www.ffmpeg.org/) homepage.
|
||||
</p>
|
||||
|
||||
### {% linkable_title Noise %}
|
||||
|
||||
To enable your FFmpeg with noise detection in your installation, add the following to your `configuration.yaml` file:
|
||||
@ -31,7 +27,6 @@ camera:
|
||||
tool: noise
|
||||
input: FFMPEG_SUPPORTED_INPUT
|
||||
name: FFmpeg Noise
|
||||
ffmpeg_bin: /usr/bin/ffmpeg
|
||||
peak: -30
|
||||
duration: 1
|
||||
reset: 20
|
||||
@ -68,7 +63,6 @@ camera:
|
||||
tool: motion
|
||||
input: FFMPEG_SUPPORTED_INPUT
|
||||
name: FFmpeg Motion
|
||||
ffmpeg_bin: /usr/bin/ffmpeg
|
||||
changes: 10
|
||||
reset: 20
|
||||
# group feature / default not in use
|
||||
@ -95,4 +89,4 @@ For playing with values (changes/100 is the scene value on ffmpeg):
|
||||
$ ffmpeg -i YOUR_INPUT -an -filter:v select=gt(scene\,0.1) -f framemd5 -
|
||||
```
|
||||
|
||||
If you are running into trouble with this sensor, please refer to this [Troubleshooting section](/components/camera.ffmpeg/#troubleshooting).
|
||||
If you are running into trouble with this sensor, please refer to this [Troubleshooting section](/components/ffmpeg/#troubleshooting).
|
||||
|
||||
@ -15,10 +15,6 @@ ha_release: 0.26
|
||||
|
||||
The `ffmpeg` platform allows you to use every video feed with [FFmpeg](http://www.ffmpeg.org/) as camera in Home Assistant. The input for ffmpeg need to support that could have multiple connection to source (input) in same time. For every user in UI and all 10 seconds (snapshot image) it make a new connection/reading to source. Normally that should never be a trouble only in strange selfmade constructs can be make mistakes.
|
||||
|
||||
<p class='note'>
|
||||
You need a `ffmpeg` binary in your system path. On Debain 8 you can install it from backports. If you want Hardware support on a Raspberry Pi you need tobuild from source by yourself. Windows binary are avilable on the [FFmpeg](http://www.ffmpeg.org/) website.
|
||||
</p>
|
||||
|
||||
To enable your FFmpeg feed in your installation, add the following to your `configuration.yaml` file:
|
||||
|
||||
```yaml
|
||||
@ -27,7 +23,6 @@ camera:
|
||||
- platform: ffmpeg
|
||||
input: FFMPEG_SUPPORTED_INPUT
|
||||
name: FFmpeg
|
||||
ffmpeg_bin: /usr/bin/ffmpeg
|
||||
extra_arguments: -q:v 2
|
||||
```
|
||||
|
||||
@ -42,18 +37,6 @@ Configuration variables:
|
||||
|
||||
You can control the `image quality` with [`extra_arguments`](https://www.ffmpeg.org/ffmpeg-codecs.html#jpeg2000) `-q:v 2-32` or with lossless option `-pred 1`.
|
||||
|
||||
### {% linkable_title Troubleshooting %}
|
||||
|
||||
In most of case, `ffmpeg` autodetect all needed options to read a video/audio stream or file. But it is possible in rare cases that's needed to set a option to help `ffmpeg`. Per default `ffmpeg` use 5 seconds to detect all options or abort.
|
||||
|
||||
First check, if your stream playable by `ffmpeg` with (use option `-an` or `-vn` to disable video or audio stream):
|
||||
|
||||
```
|
||||
$ ffmpeg -i INPUT -an -f null -
|
||||
```
|
||||
|
||||
Now you can see what going wrong. Following list could be help to solve your trouble:
|
||||
|
||||
- `[rtsp @ ...] UDP timeout, retrying with TCP`: You need to set RTSP transport in the configuration with: `input: -rtsp_transport tcp -i INPUT`
|
||||
- `[rtsp @ ...] Could not find codec parameters for stream 0 (Video: ..., none): unspecified size`: FFmpeg need more data or time for autodetect. You can set the `analyzeduration` and/or `probesize` option, play with this value. If you know the needed value you can set it with: `input: -analyzeduration xy -probesize xy -i INPUT`. More information about that can be found on [FFmpeg](https://www.ffmpeg.org/ffmpeg-formats.html#Description).
|
||||
If you are running into trouble with this sensor, please refer to this [Troubleshooting section](/components/ffmpeg/#troubleshooting).
|
||||
|
||||
|
||||
45
source/_components/ffmpeg.markdown
Normal file
45
source/_components/ffmpeg.markdown
Normal file
@ -0,0 +1,45 @@
|
||||
---
|
||||
layout: page
|
||||
title: "FFmpeg"
|
||||
description: "Instructions for how to integrate FFmpeg within Home Assistant."
|
||||
date: 2016-09-14 00:00
|
||||
sidebar: true
|
||||
comments: false
|
||||
sharing: true
|
||||
footer: true
|
||||
logo: ffmpeg.png
|
||||
ha_category: Hub
|
||||
featured: true
|
||||
---
|
||||
|
||||
<p class='note'>
|
||||
You need a `ffmpeg` binary in your system path. On Debain 8 or Raspbian (Jessie) you can install it from backports. If you want Hardware support on a Raspberry Pi you need to build from source by yourself. Windows binary are avilable on the [FFmpeg](http://www.ffmpeg.org/) website.
|
||||
</p>
|
||||
|
||||
To set it up, add the following information to your `configuration.yaml` file:
|
||||
|
||||
```yaml
|
||||
ffmpeg:
|
||||
ffmpeg_bin: /usr/bin/ffmpeg
|
||||
run_test: True
|
||||
```
|
||||
|
||||
Configuration variables:
|
||||
|
||||
- **ffmpeg_bin** (*Optional*): Default 'ffmpeg'. Set the ffmpeg binary.
|
||||
- **run_test** (*Optional*): Default True. Check if `input` is usable by ffmpeg.
|
||||
|
||||
### {% linkable_title Troubleshooting %}
|
||||
|
||||
In most of case, `ffmpeg` autodetect all needed options to read a video/audio stream or file. But it is possible in rare cases that's needed to set a option to help `ffmpeg`. Per default `ffmpeg` use 5 seconds to detect all options or abort.
|
||||
|
||||
First check, if your stream playable by `ffmpeg` with (use option `-an` or `-vn` to disable video or audio stream):
|
||||
|
||||
```
|
||||
$ ffmpeg -i INPUT -an -f null -
|
||||
```
|
||||
|
||||
Now you can see what going wrong. Following list could be help to solve your trouble:
|
||||
|
||||
- `[rtsp @ ...] UDP timeout, retrying with TCP`: You need to set RTSP transport in the configuration with: `input: -rtsp_transport tcp -i INPUT`
|
||||
- `[rtsp @ ...] Could not find codec parameters for stream 0 (Video: ..., none): unspecified size`: FFmpeg need more data or time for autodetect. You can set the `analyzeduration` and/or `probesize` option, play with this value. If you know the needed value you can set it with: `input: -analyzeduration xy -probesize xy -i INPUT`. More information about that can be found on [FFmpeg](https://www.ffmpeg.org/ffmpeg-formats.html#Description).
|
||||
Loading…
x
Reference in New Issue
Block a user