[RFE]Implement the "sparkle" effect

RESOLVED WONTFIX

Status

()

Core
ImageLib
--
enhancement
RESOLVED WONTFIX
18 years ago
7 years ago

People

(Reporter: tenthumbs, Unassigned)

Tracking

Trunk
Future
Points:
---

Firefox Tracking Flags

(Not tracked)

Details

(Whiteboard: [imagelib], URL)

(Reporter)

Description

18 years ago
Mozilla currently uses a 2D tiling scheme for displaying interlaced PNG
images. The libpng docs describe a different approach, the "sparkle" effect,
where the image pixels are just display in their final locations as they
arrive. The image seems to just grow in place.

The libpng docs do not like the sparkle method but I don't like the tiled
method because the incoming image appears to move up and to the left as more
of it arrives.

Maybe mozilla should support both. A user pref would be nice. Maybe, even, a
web page could eventually select the method.

Just an idea.

Comment 1

18 years ago
Bah. Implementing the `sparkle effect' would take us back to the bad old days of 
1995, when Mozilla/0.91 and earlier did this for interlaced GIFs. And a user pref 
for it would incite rage from those who accuse Mozilla of being bloated, and 
they'd be right.

If you don't like the image appearing to move up and/or to the left with the 
tiled method -- which is a reasonable complaint -- then file an RFE for each tile 
(except the tiles on the edge of the image) to be centered around the downloaded 
pixel, rather than having its top left corner on the downloaded pixel.

Updated

18 years ago
OS: Linux → All
Hardware: PC → All
(Reporter)

Comment 2

18 years ago
Who mentioned GIFs? I didn't. This is about 2D interlacing.

You don't like it; I do. That's nice.

Comment 3

18 years ago
> Who mentioned GIFs?

I did. That's why the comment says `Additional comments from Matthew Thomas 
(mpt)' at the top.

> This is about 2D interlacing.

As used by both PNGs and GIFs. There's no good reason for them to behave 
differently.
(Reporter)

Comment 4

18 years ago
But a 1D sparkle would probably have a venetian blind effect. Hard to say if it
looks good. There's no reason why 1D and 2D interlacing has to be the same.

Updated

18 years ago
Status: NEW → ASSIGNED

Updated

18 years ago
Target Milestone: --- → Future

Comment 5

17 years ago
All pnunn bugs reassigned to Pav, who is taking over
the imglib.
Assignee: pnunn → pavlov
Status: ASSIGNED → NEW

Comment 6

17 years ago
The best method as far as I can tell is the interpolated approach. This has the 
effect of starting off looking blurry and coming into focus as the image loads. 
This method makes text in the image readable earlier in the download process 
than the other methods. O'Reilly's "PNG: The Definitive Guide" talks about it 
and shows a side-by-side comparison.

Comment 7

17 years ago
<URL:http://www.its.caltech.edu/~stl/png.html> has some nice images of PNGs
being showed with a) bicubic interpolation, b) bilinear interpolation, c) no
interpolation.

As you can see, it's much easier to see what the image is about when using
bicubic interpolation, even when only 1/64 of it is downloaded. I think
interpolation (preferably bicubic) would be the best way to display adam7 PNGs.

Comment 8

17 years ago
Bilinear/cubic interpolation should be filed as a separate bug (which would be 
much more likely to be fixed than this bug would).

Updated

17 years ago
Whiteboard: [imagelib]

Comment 9

17 years ago
Bicubic interpolation is now bug #75941.

Comment 10

17 years ago
> Bicubic interpolation is now bug #75941.

No, it's bug #75077.

Sorry for not mentioning that I had filed this bug.

Updated

16 years ago
Summary: Implement the "sparkle" effect → [RFE]Implement the "sparkle" effect

Updated

11 years ago
Assignee: pavlov → nobody
QA Contact: tpreston → imagelib

Comment 11

10 years ago
Swapping URL for archive.org
Talked with Joe, marking as wontfix.
Status: NEW → RESOLVED
Last Resolved: 7 years ago
Resolution: --- → WONTFIX
You need to log in before you can comment on or make changes to this bug.