Compare video quality

[cd090580]

Hi,

I've done several captures of the same music video on several channels. Now is there a quick way (watching them is the best way but takes too much time) to determine the best video of all ?

No need of advanced analysis, just a way to get an average quality (ie: capture1: 90/100; capture2: 93/100,...) ?

Is there any programs that can do that ?

Thanks.

[redwudz]

Not really. If they all used the same capture/encoder settings, visual is still the best way to judge quality.

[noemi7]

I use Gspot to determine the highest average bitrate. Usually difference for the same tile/program is no higher than +- 500 kbps, so not much.

[redwudz]

Gspot or most any similar program will just tell you the quality of the capture or the encode, not the appearance quality of the video. You can have a out of focus video and a clear one that can show exactly the same in Gspot.

[schematic2]

What a good question. I really would be nice if someone made a software which displayed "zoomed in multi-windows" of 2 clips playing simultaniously. Even at lower framerate.

[jagabo]

If these are really the same videos recorded with the same codecs just looking at the file size will tell you as much as Gspot.

[cd090580]

It's the same videos (recorded with my standalone dvd recorder) but the source is not the same. Record1 come from channel1, Record2 come from channel 2,...

[Nelson37]

Once you got different sources bitrate and other "data" comparisons go out the window.

Would you prefer a hi-bitrate capture from Analog cable or a lower-bitrate capture from Digital Cable?

Have just done some similar comparisons of older vs newer captures and I guarantee you that simply comparing bitrates is not the answer.

Open two copies of Vdub and step thru the frames. Then compare full motion smoothness. Then watch for artifacts at scene changes or fast motion. Then compare audio quality. Then YOU decide based on YOUR opinion of which one YOU like best.

I have a program that says that hamburger tastes better than Filet Mignon. Want a free copy?

[noemi7]

It's the same videos (recorded with my standalone dvd recorder) but the source is not the same. Record1 come from channel1, Record2 come from channel 2,...

You wrote 'capture' hence I thought you have in mind a DVB-x card. Regarding a standalone recorder it doesn't matter as it re-encodes both video and audio. Thus the quality depends on your recorder setting.

[cd090580]

I know, but some satellite channels are more compressed than others.

[noemi7]

Here you will find bitrate and other quality parameters: http://www.digitalbitrate.com

[jagabo]

I know, but some satellite channels are more compressed than others.

If the frame size is the same then the one with the higher bitrate (the larger file) is proably the better one. But you can't be sure about this. If one was an all digital source and the other went through an analog step the all digital one may be better even if it has a lower bitrate. You might aslo find that one is better during slow scenes and another is better during fast scenes.

[vhelp]

Evening everyone.

The quesiton is, what are actually looking for in terms of determining which is the better quality of
the two videos ??

Maybe ask yourself, "what am I looking for, artifact'wise ??" in the videos.

Are you looking for..

mpeg artifacts; analog capture noise; analog capture of digital noise;
reception noise; antenna induced noise; digital-to-digital type noise;
and many more..

..and then compare which one has the least of these, I guess.

Anyway, when it comes to "digital" medium type sources (ie, digital cable/satellite) then you
prob have less work to rumble through. When you capture (record) direct digital type source
you have (or should have) no noise in the video. The video is crisp in terms of the means you
snipped it from.

So, that being said, what you might require comparing is the output source from the
other output source type -- you said, your source(s) were from two from the same unit
but different leads or (output) connectors, and you feel that there may be some difference..

(you could prob write a tool to do something "comparitably" of the two clip yourself)

..but you used the same unit, and the same program content (the video) though different times
you had obtained them. The difference could be from the result of two different output connect
setup in your equipment's setup screen. For instance, in my DVR-220 recorder, there are three
output sources to choose. And each one can have a separate configuration setting that is for
the users convenience. So, as an example, the difference could be in the brightness (levels) of
the video.

Given this last passage above, it is possible that there could be some minor differences in the two
videos, though hardly worth the effort, save mainly out of curiosity purposes. But, then you have
to separate into two catagories. 1) The program content could be the same but the provider airing
it could have different (player) equipment. 2) Then there could be "air times" that might have some
influence on the stream comming into your house and then your equipment, etc. Air traffice and
other such things influence the signal floating around. Your grounding, and other generator type
machines nearby can also influence the signal/video in some way while being received and/or played.

But, for the most part, and when the program content is digital based, and setups for each equip'mt
and so on are set the same, then there should prob be not much difference if any at all.

--- AVIsynth and plugin filters

There are several filters for manipulating images in a video that can be scripted to provide some
useful information. I don't know what avisynth all that much, so the names of them escape me.
But I have seen a few crude examples floating around from time to time on the internet. A google
search (using a few key phrases or words) might get some pointers in that direction...

--- writing your (crude/basic) image analysis tool

..or, if you have some programming experience, you could prob throw somethign up real quick
that just opens up a few BITMAP images and you could incorporate a routine of some sort that
will provide some information on the bitmaps.

Maybe, someone will be nice and write you up and quick n dirty tool that does only minor image
measuring and go from there. Anyway.

An example of usesful information could come in the form of the images brightness or darkness
levels, or color channels (components, r/g/b, or y/u/v) and measured for Averages or Means or
whatever you want to find out in your videos/images.

But, measuring for other more advance things like mpeg errors can more difficult a task. Like for
measuring which version (Fig A vs. Fig B) has lessor pixelation. Something I always wanted to
build for myself in my own image analysis projects I endeavor in my spare time.

---

I don't know.. I'm just throwing up ideas to spark anyone's imagination in hopes they might take
it further

-vhelp 4621