Re: SVG12: rendering accuracy

Hello www-svg,


> On Thursday, December 29, 2005, 12:18:29 PM, Bjoern wrote:
> 
BH>> * Chris Lilley wrote:

>>>Yes, it did get overlooked. The wording about accuracy on initial
>>>zoom level from the message you cite did get added, but the change
>>>from device pixel to px unit was overlooked. Thanks to you both for
>>>catching this, it has now been changed in the spec to "px unit" as
>>>previously promised.

BH>> There are still a number of references to device pixels in the
BH>> draft, e.g. in the transformBehavior section; did the Working Group
BH>> review the entire draft and ensure that any occurence of the term
BH>> is precisely what is desired and that the draft draws attention to
BH>> the differences? There are various references to similar terms like
BH>> "device units", "pseudo device pixels", etc. which is at best
BH>> confusing.

> Thanks, we will check further.

On checking further we found this instance, already described as correct:

> The reference to device units in 7.7.4 The Current Transform Matrix is
> intentional and correct. At the end of the day, that is what is rendered
> and those pixels (if they are pixels) are not CSS px units but  are
> instead physical device pixels, or dot patterns for printers.

Also in multimedia.html, for untransformed video, again correct:

  The media shall be treated as a point, defined by x and y attributes.
  This point must be transformed to the nearest actual device pixel.
  Video at the native resolution given by the media shall then painted
  on a region whose center is the pin point and whose width and height
  are defined by the media. The pixels must be aligned to the device
  pixel grid and there shall be no resampling. The values of the width
  and height attributes shall have no effect on the rendering of the
  video.

because it really is untransformed video - not rotated and not scaled.
The MPEG people who reviewed this part of the spec were very insistent
that there had to be an option to put video onto device pixels without
resampling.

Also in painting.html

 crispEdges Indicates that the user agent shall attempt to emphasize the
 contrast between clean edges of artwork over rendering speed and
 geometric precision. To achieve crisp edges, the user agent might turn
 off anti-aliasing for all lines and curves or possibly just for
 straight lines which are close to vertical or horizontal. Also, the
 user agent might adjust line positions and line widths to align edges
 with device pixels.

Once again correct, device pixels are intended here. A printer, for
example, would not adjust line positions to px but to its native device
pixel grid.

Lastly in  svgudom.html

  Note that "pixel coordinate systems" are host-specific. Two possible
  approaches that hosts might use for pixel coordinate systems are
  actual device pixels or (particularly for high-resolution devices)
  pseudo device pixels which exactly match SVG and CSS's "px"
  coordinates.

Once more this is correct as described.

We therefore believe that all occurrences of 'device pixel' are
intentional and correct.


-- 
 Chris Lilley                    mailto:chris@w3.org
 Interaction Domain Leader
 Chair, W3C SVG Working Group
 W3C Graphics Activity Lead
 Co-Chair, W3C Hypertext CG

Received on Tuesday, 2 May 2006 14:20:40 UTC