9776 2010-05-20 09:22:17 +0000 Percentages for text positioning don't seem better than using the actual video dimensions. If you require integer percentages, then it doesn't work well for videos larger than 100px high. If not, then it's just an arbitrary resizing of the dimensions (sli 2012-07-18 18:42:35 +0000 1 1 1 Unclassified WHATWG HTML unspecified Other other RESOLVED NEEDSINFO http://www.whatwg.org/specs/web-apps/current-work/#timed-track-cue-line-position P3 normal Unsorted 1 contributor ian astrange+w3 ian mike public-html-admin contributor oldest_to_newest 35774 0 contributor 2010-05-20 09:22:17 +0000 Section: http://www.whatwg.org/specs/web-apps/current-work/#timed-track-cue-line-position Comment: Percentages for text positioning don't seem better than using the actual video dimensions. If you require integer percentages, then it doesn't work well for videos larger than 100px high. If not, then it's just an arbitrary resizing of the dimensions (slightly less precise horizontally if you round off) and you're more likely to have pixel measurements on hand than percentage measurements (since Aegisub shows pixels). Posted from: 184.36.82.183 39352 1 ian 2010-09-25 15:51:25 +0000 What's the use case for precision to more than 1/100th of the video dimension? We can't use pixels because it would mean that you couldn't use the same subtitle file for different resolution versions of the same video (consider e.g. YouTube videos which are often available in several sizes, but all have just one subtitle track). 39399 2 astrange+w3 2010-09-25 19:56:11 +0000 This is mine - I meant that subtitle files could define their own pixel dimensions, and then coordinates could be resized to match the video. This is how SSA works and it's quite successful - explicitly positioned text uses pixels, but text that just needs to be "center lower left" just has the alignment set to "center lower left", avoiding having to think about pixels and the size of the text box entirely. Obviously none of this is a problem if you allow non-integer coordinates, since any size can be resized into 0-100% using floating point. And it would avoid needing to add a dimension header to the file as well.