Incredible, free, video editing software

I found some software today that is so amazing I
just wanted to share it, in case anyone might
benefit:

Avidemux - Open source video editor. I wanted a way
to rotate a video. What I found was a program that
does to video what Photoshop does to images. I
resized, cropped, rotated and changed the format
of a MOV file, all in one operation. It took about 1
minute to process a 134 MB MOV and get a smaller,
rotated, cropped, 23 MB MP4.

The software can edit colors, audio elements, cut,
copy and paste, add subtitles, borders... And the
installer is only about 12 MB, with no special dependencies.


What I'm using on XP:

https://sourceforge.net/projects/avi...atest/download

A slightly later version that says it's not XP-compatible.
This page also has a 64-bit version:

http://fixounet.free.fr/avidemux/download.html

Basic instructions:

http://www.howtogeek.com/108584/how-...with-avidemux/

Mayayana wrote:
I found some software today that is so amazing I
just wanted to share it, in case anyone might
benefit:

Avidemux - Open source video editor. I wanted a way
to rotate a video. What I found was a program that
does to video what Photoshop does to images. I
resized, cropped, rotated and changed the format
of a MOV file, all in one operation. It took about 1
minute to process a 134 MB MOV and get a smaller,
rotated, cropped, 23 MB MP4.

The software can edit colors, audio elements, cut,
copy and paste, add subtitles, borders... And the
installer is only about 12 MB, with no special dependencies.


What I'm using on XP:

https://sourceforge.net/projects/avi...atest/download

A slightly later version that says it's not XP-compatible.
This page also has a 64-bit version:

http://fixounet.free.fr/avidemux/download.html

Basic instructions:

http://www.howtogeek.com/108584/how-...with-avidemux/


D/L it. Have not tried editing but tried it for changing formats.
Changed an AVI to MP4 and got a larger file size. Then changed
another MP4 to AVI and also got a larger file.

--
GW Ross

Time flies like an arrow; fruit flies
like a banana.

| D/L it. Have not tried editing but tried it for changing formats.
| Changed an AVI to MP4 and got a larger file size. Then changed
| another MP4 to AVI and also got a larger file.
|

I just tried that. Changing a 48 MB mp4 to
AVI came out at 77 MB. Changing that back
to an mp4 came out at 74 MB.

I tried a little testing. Switching the video
output quantization setting from H.263 to
MPEG and switching macroblock decision
from QPel8 to none produced an identical video
of 179 MB rather than 74. What did I change?
I really don't know.

Then I went into the video settings and
selected a target size of 40 MB and got a
video of 47 MB, about 2/3 the size of the
original.

I can't assess the quality. They all look the
same to me. But they were not high quality
video starting out. They're just MP4s.

This page seems to shed a little light:

https://en.wikipedia.org/wiki/Video_..._picture_types

There seems to be some very complex
compression technology, similar to the
combination of quality and compression
alogrythms that can be applied to images.

I'm not an expert on either audio or video,
so I don't understand most of the options.
I don't know the differences any more than
most people know a BMP from a JPG. I'd be
curious to hear what a professional would say.

In my case I have a friend who had a 134 MB
video of birds, apparently taken with a iphone.
It was sideways and giant. I found it almost
effortless to rotate it, crop it, resize it and
change it from MOV format to mp4 in Avidemux.
That allowed me to make two other versions:
One desktop size at 23 MB and one for emailing
at 4.5 MB. I now know I probably could have
shrunk those by half, only requiring a longer
processing time.

Prior to that I had been searching for hours
just to find a way to rotate the video. Windows
Movie Maker for XP wouldn't open MOV. Once
I got the MOV converted to AVI, Movie Maker
closed with no error when I tried to rotate the
video. The Vista version just showed a shaking
image and also failed to rotate. Another program
was a 20 MB download and needed .Net. (The
Win7 version of Movie Maker requires signing up
for Microsoft's online nonsense.) Handbrake
could convert but didn't do much else and had
limited format options. VLC has impressive options
but the GUI is quirky and I had a hard time
getting a rotated video without messed up
audio. Avidemux was, by far, the most functional
option I found, and seems to be comparable to
software costing hundreds of dollars.

Thanks to your comment I may do some research,
now knowing that a few settings changes can greatly
affect all sorts of things.
Maybe someone reading this can shed more
light on the technology.

Mayayana wrote:
| D/L it. Have not tried editing but tried it for changing formats.
| Changed an AVI to MP4 and got a larger file size. Then changed
| another MP4 to AVI and also got a larger file.
|

I just tried that. Changing a 48 MB mp4 to
AVI came out at 77 MB. Changing that back
to an mp4 came out at 74 MB.

I tried a little testing. Switching the video
output quantization setting from H.263 to
MPEG and switching macroblock decision
from QPel8 to none produced an identical video
of 179 MB rather than 74. What did I change?
I really don't know.

Then I went into the video settings and
selected a target size of 40 MB and got a
video of 47 MB, about 2/3 the size of the
original.

I can't assess the quality. They all look the
same to me. But they were not high quality
video starting out. They're just MP4s.

This page seems to shed a little light:

https://en.wikipedia.org/wiki/Video_..._picture_types

There seems to be some very complex
compression technology, similar to the
combination of quality and compression
alogrythms that can be applied to images.

I'm not an expert on either audio or video,
so I don't understand most of the options.
I don't know the differences any more than
most people know a BMP from a JPG. I'd be
curious to hear what a professional would say.

In my case I have a friend who had a 134 MB
video of birds, apparently taken with a iphone.
It was sideways and giant. I found it almost
effortless to rotate it, crop it, resize it and
change it from MOV format to mp4 in Avidemux.
That allowed me to make two other versions:
One desktop size at 23 MB and one for emailing
at 4.5 MB. I now know I probably could have
shrunk those by half, only requiring a longer
processing time.

Prior to that I had been searching for hours
just to find a way to rotate the video. Windows
Movie Maker for XP wouldn't open MOV. Once
I got the MOV converted to AVI, Movie Maker
closed with no error when I tried to rotate the
video. The Vista version just showed a shaking
image and also failed to rotate. Another program
was a 20 MB download and needed .Net. (The
Win7 version of Movie Maker requires signing up
for Microsoft's online nonsense.) Handbrake
could convert but didn't do much else and had
limited format options. VLC has impressive options
but the GUI is quirky and I had a hard time
getting a rotated video without messed up
audio. Avidemux was, by far, the most functional
option I found, and seems to be comparable to
software costing hundreds of dollars.

Thanks to your comment I may do some research,
now knowing that a few settings changes can greatly
affect all sorts of things.
Maybe someone reading this can shed more
light on the technology.

I can comment as a noobie :-) I've burned a total
of three dual layer DVDs, so that's some idea of
my "depth of experience" :-)

You can divide the techniques up.

1) Uncompressed. RGBA. Huge datarate. The format that
comes off a hardware capture device maybe. I might get
a 200GB video from this, from my ****ty WinTV card.

2) Lossless compression. Huffyuv or FFV1. Generally
achieves up to 3:1 compression. Same as (1), but
saves you some disk space. Tends to chew up CPU when
editing (i.e. you may regret your choice at that point).
Now, your (archived) original capture goes from 200GB to 70GB.

3) Lossy compression. Compression ratios of around 100:1 or so.

DCT - discrete cosine transform. Frequency domain analysis.
Discarding high frequency content and sharp edges.
Tossing detail humans don't notice, when there is a lot
of action.

- In addition to compression of a frame, there is a
possibility to do temporal compression, look at the
frames nearby, and come up with a difference used for
prediction of the next frame. Frame types then are IBP,
one of the frame types having a lot of bits, the others
representing differences. A repeating pattern of them
is a GOP or group of pictures. A GOP of 12 or 15 might be
typical, around 1/2 a second of video before the next
I frame. Snipping video can be seamless, if done on
GOP boundaries. An "open" GOP, one GOP can depend on an
adjacent GOP. With a bit of a bandwidth penalty, it's
possible to encode with a "closed" GOP, so each group of
pictures is independent of the next group. Which makes
fast forward and fast reverse work better.

- It's possible to use MJPEG compression, and make all the
frames independent of one another. The compression is
better than (2), but less than a full blown DCT method
with temporal thrown in. One advantage of MJPEG as a
compressor, is every core on your CPU can work on
MJPEG compression of a frame at the same time. A
disadvantage, is JPEG tends to have mathematical
"ringing" and screwed up color fringes. Attempts to do
screen capture video with this, didn't look good.

Vector Quantization - Cinepak

5X bigger than modern lossy compression methods.

Compression is *very* slow (less than one frame a second).
Compressor is single threaded (split video in sections,
put each section on a separate CPU core for compression).
Video has excellent seek characteristics (fast forward
is like butter).
Decompression is possible on absolutely gutless processors.
Nobody used it but me :-)

https://en.wikipedia.org/wiki/Cinepak

I found a nice animation of how the compressor
works, but don't seem to have bookmarked it.

*******

When you use a DCT method, you can do things like
define a "quality" or Qfactor. At Q:2, that says you
want absolutely best quality, and the bitrate simply
leaps up to make that happen. If there is an explosion
in the video, maybe it takes 50Mbit/sec to full represent
it, without blocky artifacts.

The other constraint you can apply, is bandwidth. You
define

min bw
avg bw
max bw
buffer size

The average might be 3900Kbit/sec, the min 0, the max 8000Kbit/sec.
The buffer a bit less than 2MB (the size of some cheap DVD player
read-ahead buffer). If there is an explosion, the Qf leaps up
to 32, the quality is ****, but the segment of video stays
below 8000Kbit/sec. When the video scene content is still, the
Qf drops to 2 (excellent quality), the bitrate becomes minimal,
the read-ahead buffer can refill itself. So there is a model
based on player characteristics, with enough slop to handle
a variety of video content.

The compressor profile can work on either basis (Qf
constrained or bitrate constrained), based on the
preference of the person doing the video edit.

If you use FFMPEG, you get to set these things.
You can dial the knobs until you're happy with the
space versus quality tradeoff. In this example, the
ntsc-dvd specification, causes the usage of mpeg2video
codec for video. Not sure what the default audio
codec is.

ffmpeg -i G:\some.avi -target ntsc-dvd
-aspect 4:3 -g 15 -bf 2 -sc_threshold 1000000000
-b:v 3900k -maxrate 8000000 -minrate 0
-bufsize 1835008 -pass 1 -y NUL

ffmpeg -i G:\some.avi -target ntsc-dvd
-aspect 4:3 -g 15 -bf 2 -sc_threshold 1000000000
-b:v 3900k -maxrate 8000000 -minrate 0
-bufsize 1835008 -pass 2 F:\output.vob

ffmpeg -i F:\output.vob -vcodec copy F:\video.m2v
ffmpeg -i F:\output.vob -acodec mp2 -ac 2 -b:a 192k F:\video.mp2

That's two pass conversion, with the first two commands.
The first pass, generates a logfile in the working directory.
The second pass recognizes the logfile by name, and applies
the information during the actual final (second pass) compression.
By having Q or bandwidth info to work with, it knows a
"bad section" is coming or whatever. The logfile is relatively
tiny, so doesn't contain a representation of the video at all.

The last two commands split the video into two separate
files. One video, one audio. Which is sometimes required
by other authoring tools. If you want to put a French and
an English soundtrack on your output media, having separate
streams comes in handy before final re-assembly. The authoring
tool may allow you to build a menu, to select the language
before you press play.

Video can also be coded CBR or Constant BitRate, but
I've not discovered a reason to do that yet. The above
assumes VBR or variable bit rate.

*******

And these methods are designed for subsets of all video content.
The DCT compressors might be good for motion pictures, but
a poor choice for cartoons. Since I don't collect cartoon
video, I have no idea what is used there.

HTH,
Paul

| I can comment as a noobie :-) I've burned a total
| of three dual layer DVDs, so that's some idea of
| my "depth of experience" :-)
|

It's a lot more than I knew. Thanks. I realize
now, in thinking about it, that videos such as
Youtube videos are usually fairly poor quality.
Slight blurs are common when there's movement.