1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
|
Introduction
============
With home made videos or screencasts, we often find a constant noise in our recording due to electric wiring, fan, choke coil of fluorescent
lamp etc. This could be irritating.But fortunately, these noises can be easily detected and can be removed with GUI based tools, such as
Audacity.
To use any GUI based tool, we need to extract audio manually and then feed it to the software, once done we have to again join the noisefree
audio with the video. This is OK with 1 or 2 files. But to for automating each step and to handle multiple files we need a simple script.
One can simply use `sox` and `ffmpeg` commands shown below in given order to get the same result(see `Working`).
UPDATE for 14.04
----------------
Install **libav-tools** and **sox**.
Run as ::
$ bash noNoise.sh noisyVideo.mp4 noise-reduction-factor
This will create an backup file of original video in **/tmp**
and noisefreeVideo in **pwd**.
Example ::
$ bash noNoise.sh Kazam_screencast_00000.mp4 0.2
It will create **noisefree_Kazam_screencast_00000.mp4.mp4** in **pwd**
and a backup of original video in **/tmp/orig_Kazam_screencast_00000.mp4**
This is just an early fix, will add more features soon.
Please raise issues for any bugs or email me for feature requests.
Also, I recently came across **Kazam** package (available in repositories),
its a good screen recorder with most needed keyboard shortcuts, such as
* start recording: Super + Control + r
* pause recording: Super + Control + p
* finish recording: Super + Control + f
* show Kazam: Super + Control + s
* quit: Super + Control + q
Required packages (Upto Ubuntu 13.10)
-------------------------------------
* sox
* ffmpeg
* ffmpeg2theora
* libmp3lame0
* Linux machine with default python
Usage
-----
### Remove noise from a single file
$ python noNoise.py VideoWithNoise.ogv CleanVideo.ogv
(source file) (destination file)
(OR)
$ python noNoise.py VideoWithNoise.ogv CleanVideo.ogv 0.21
(source file) (destination file) (noise factor)
The third argument is optional(Noise factor). The scale spans from
`0.0` to `1.0`. Zero means no noise supression and 1.0 means full. The full
scale is avoided. Best optimum result is found between `0.2` to `0.3`. By default
script will take `0.26`. One can experiment with noise factor to get best noise
free video.
NOTE: Careful, destination file will be overwritten if exist in given path.
### Remove noise from all files inside a directory
$ python noNoise.py allNoisyFiles allCleanFiles
(source dir) (destination dir)
(OR)
$ python noNoise.py allNoisyFiles allCleanFiles 0.21
(source Dir) (destination dir) (Noise factor)
NOTE: Please don't use any '/' after directory name. It will spit error.
The fix is possible, but I don't want to spend time on it. This script is dirty
but useful(atleast for me). When I find time, I will surely modify it. Meanwhile
you all are welcome to add modifications.
Working:
-------
###1. Extracting video in less compressed format
ffmpeg -i 1.ogv -sameq -an 2.wmv
Extracting video in wmv format for easy editing(less compressed
than mp4,ogv,avi). We can leave the video intact and combine the
noiseless audio later to it, but it will hamper the video quality
of the newly joined video.
The size of this 'wmv' will be approximately 5 times than that of
original 'ogv' video.
###2. Extracting audio in less compressed format
ffmpeg -i 1.ogv -sameq 2.wav
Extracting audio in wav format for fast & easy editing.The size of the
`wav` audio file will be approximately 8 times larger than the original.
###3. Getting noise profile
sox 2.wav -t null /dev/null trim 0 0.5 noiseprof myprofile
Creating a noise profile of original audio at 0 to 0.5 second.
One can change this duration if required. In most cases the
standard noise is evenly distributed throughout the recording(eg:
fan, PC etc), so the default 0 to 0.5 value will do the trick.
###4. Converting audio according to noise profile
sox 2.wav 2-noisefree.wav noisered myprofile 0.26
Creating a noisefree audio based on our noise profile. The value
`0.26` is important. This is scale for noise removal. 0 means no removal
and 1 means full removal. The full removal will supress most of the
orginal audio too. So as per my R&D, I found `0.26` to be most optimized
one for noise removal.
###5. Combining back audio and video
ffmpeg -i 2-noisefree.wav -i 2.wmv -sameq vid.wmv
Merging new noiseless audio and old video together.
###6. Final conversion
ffmpeg2theora vid.wmv -o vid.ogv
Now converting wmv into our favorite ogv format. This will create a
`vid.ogv` of almost same size that of original video.
License
-------
GNU GPLV3
|
