Re: Precision of #xywh=percent:...

On Thu, 07 Apr 2011 09:38:41 +0200, Silvia Pfeiffer  
<silviapfeiffer1@gmail.com> wrote:

> On Thu, Apr 7, 2011 at 5:22 PM, Philip Jägenstedt <philipj@opera.com>  
> wrote:
>> On Thu, 07 Apr 2011 02:54:24 +0200, Silvia Pfeiffer
>> <silviapfeiffer1@gmail.com> wrote:
>>
>>> Oh, it hadn't occurred to me that we may need fractions of
>>> percentages. I think we should allow that.
>>>
>>> Another use case may be: a video recording of a concert from the
>>> distance with a camera that won't zoom any further and you know you
>>> can focus in on a specific cropped region.
>>
>> In this situation, why not just use the #xywh=pixel:x,y,w,h grammar?  
>> This is
>> more straight-forward if there is only one resolution of the video. If  
>> you
>> have multiple resolutions of the video, then you can do the cropping  
>> while
>> encoding those multiple versions and save bandwidth at the same time.
>>
>> This line of argument actually works against my 16:9 cropping as well  
>> -- if
>> there is only one version then you don't need percentages, but if there  
>> are
>> several then you can do cropping while encoding.
>
>
> I actually think that because things can go full-screen now, the
> non-percent version is almost useless and an author would almost
> always want to use percent, just to make sure that whatever display
> size the user chooses, it will still display the same thing.

The pixel syntax operates on the intrinsic size of the video, not the  
display size. The result would be the same regardless of fullscreen,  
modulo scale of course.

>>> There are further use cases at
>>> http://www.w3.org/2008/WebVideo/Fragments/WD-media-fragments-reqs/,
>>> e.g.
>>> 4.1.4 Scenario 4: Image Region of video over time
>>
>> MF doesn't allow for changing the highlighted/cropped region over time.  
>> Even
>> so, it's not something you need the percent syntax for.
>>
>> I'm going to play the devil's advocate and suggest that we drop the  
>> percent
>> syntax completely.
>
>
> I think that would be a big problem.

What are the use cases where all of the following are true:

* You have multiple videos of different resolution.
* Those videos are not the same video encoded at different resolutions.
* You want to apply the same relative cropping to all of them.

?

-- 
Philip Jägenstedt
Core Developer
Opera Software

Received on Thursday, 7 April 2011 08:07:17 UTC