View Full Version : question about rezising
Thunderbolt8
25th October 2006, 14:55
when keeping the orginal AR of a dvd 720x576 (PAL) and not resizing, the visible window I get from my encoded file has exactly the same size as the original version on dvd has.
now Im just asking myself why basically almost every movie you can find on the inet is made with resize instead of keeping that original AR. as an example, for a widescreen 1 : 2,35 movie you find lots of different "versions" with 672x288, 640x272, 800x336 etc. why do people do that ? there must be a reason or some advantage with resizing, but as it seems I cant figure out which it is. maybe it has something to do with the black widescreen bars, which remain when using 720x576, is there some "space" wasted in those black areas, which could otherwise be used for better encoding quality with another resized resolution ?
enlighten me please.
foxyshadis
25th October 2006, 18:54
Because gordian knot and auto gordian knot won't do soft AR, and 95% of dvd-rippers use one or the other. The only advantage is that compatibility problems are minimized, but if you're encoding for yourself, compatibility shouldn't be an issue.
What do you mean by black areas? A 2.35:1 movie can fill a full 720x576 frame with no black bars. That's what anamorphic means, different source and output AR.
GodofaGap
25th October 2006, 20:42
What do you mean by black areas? A 2.35:1 movie can fill a full 720x576 frame with no black bars. That's what anamorphic means, different source and output AR.
When talking about DVD and normal video, 2.35:1 can't fill the full frame. There is only 4:3 and 16:9. :)
Thunderbolt8
25th October 2006, 21:10
thats what I mean, with 1: 2,35 there are always black bars above and below the visible area, when not resizing and keeping 720x576. I was just curious if not resizing could have any disadvantages.
foxyshadis
25th October 2006, 21:21
Oh right, on DVDs it's a lot stricter. Well, in that case you can just crop off the borders, which is what most people do. When encoding, it doesn't waste a lot of space, but the edges will look worse than if you don't (and to a much lesser extent, the whole video).
valja
25th October 2006, 21:44
The reason, why rips are resized, is very clear. DVD frame has Aspect Ratio (AR) 720:576 = 1.25:1 (PAL) or 720:480 = 1.5:1 (NTSC). But both these frames represent 4:3 (or 16:9) frame.
DVD players "know" it and, when playing movie, change AR to described in headers - 4:3 or 16:9.
Think about simple thing - just the same, standard 4:3 movie will have PAL DVD frame 720x576, and NTSC DVD will have 720x480. These frames have different AR, but rips from both must have the same 4:3 frame.
That means, if one will just crop out black bars and encode (to avi etc), the result will have wrong aspect ratio (1.25:1 or 1.5:1 instead of needed 4:3). Rip from PAL DVD will be stretched by vertical, rip from NTSC will be stretched by horisontal - both will be wrong.
Take a look to http://lipas.uwasa.fi/~f76998/video/conversion/, there is very good explanation.
foxyshadis
25th October 2006, 22:23
Players have supported aspect ratio correction for years, however. PAR for mp4 (also how mpeg-2 stores it), DAR for mkv, and either one for avi or ogm.
Jay Bee
26th October 2006, 00:16
Also because ffdshow and xvid decoders ignore AR settings by default.
foxyshadis
26th October 2006, 04:15
Why shouldn't they? They decode, they don't display. The player (wmp, zoomlayer, mpc, etc) works with the renderer to display properly. All in ones, like VLC and Mplayer, do it correctly as well.
Jay Bee
26th October 2006, 04:31
If the players are fault then how come they all work with all other formats except XviD?
foxyshadis
26th October 2006, 06:11
I'm not psychic, I don't have any idea how your system is set up. If you mean PAR embedded in the mpeg-4, and not the container, that has never given me problems that I know of. But I avoid it in favor of container-based AR, for flexibility, whenever I can.
Jay Bee
26th October 2006, 07:03
I have no problems with stream embedded PAR on my system at all, thank you very much.
I was just saying that to make this work with ffdshow you need to enable an option in ffdshow (use overlay mixer) that is not set by default. This means that very few people are able to correctly playback xvid files with embedded PAR since fidling with DirectShow filter settings is not something that many average computer users do.
I think this is one of the answers to the OP's question. I'm not sure which part you disagree with. That a good player shouldn't need this setting? That you don't need this setting on your system? That it is in fact on by default? That it isn't an issue because users should be able to deal with it?
*.mp4 guy
26th October 2006, 15:04
There is another reason why people resize video, and that is because storing things anamorphicly is less efficient then storing them with square pixels, and the difference in efficiency can be important for low bitrate encoding, and usually justifies the minimal loss of detail from resizing. For transparent backups you should always encode the full anamorphic resolution, but for anything less then transparency you should resize to a square pel frame, because the loss of detail from compression will always be worse then what you loose with a good resizer.
bond
26th October 2006, 21:50
when keeping the orginal AR of a dvd 720x576 (PAL) and not resizing, the visible window I get from my encoded file has exactly the same size as the original version on dvd has.
now Im just asking myself why basically almost every movie you can find on the inet is made with resize instead of keeping that original AR. as an example, for a widescreen 1 : 2,35 movie you find lots of different "versions" with 672x288, 640x272, 800x336 etc. why do people do that ? there must be a reason or some advantage with resizing, but as it seems I cant figure out which it is. maybe it has something to do with the black widescreen bars, which remain when using 720x576, is there some "space" wasted in those black areas, which could otherwise be used for better encoding quality with another resized resolution ?
enlighten me please.- people crop the black bars as its a waste of space
- people downsize as for lower bitrates it can bring more quality than keeping the original resolution
- people resize to not have an anamorphic resolution as you dont have to bother with players/decoders/renderers that dont handle anamorphic resize (easily) and the idea of anamoprhic is kinda strange :D
sillKotscha
26th October 2006, 22:19
and the idea of anamoprhic is kinda strange :D
that's the reason for me I have never ever encoded an anamorphic backup :D
Jay Bee
27th October 2006, 09:24
Does anyone know why encoding is ever done with non-square pixels anyway? Is it something to do with the fact that horizontal motion is more likely to occur than vertical motion?
*.mp4 guy
27th October 2006, 15:57
Anamorphic encoding is done to get the most resolution out of a fixed resolution format. So if you have a 16:9 1280*720 source you could keep 66% of the vertical resolution with an anamorphicaly encoded ntsc dvd, but with a non-anamorphic ntsc dvd you can only keep 55% of the vertical resolution. Unless you don't have enough bitrate for the source, the higher vertical resolution will look better. However anamorphic encoding is only a good idea when you have extra bitrate and unused resolution in a fixed resolution format.
vBulletin® v3.8.11, Copyright ©2000-2026, vBulletin Solutions Inc.