Re: Modernization of JPEG and JPEG 2000: suggestions...

> Modernization of JPEG and JPEG 2000: suggestions...

Here my comments. Please note that these are my personal impressions
and in no way "official ISO" statements.

> JPEG
> * Support for 12x12 quantization, separate from the 8x8 used now.

What do you mean by 12x12 quantization? 12 bit (which is available)
or larger DCT arrays? (which is unlikely to happen, and doesn't solve
the elementary problem of DCT in first place - just moves it away for
a - short - while.)

> * Allow 8x8 and 12x12 to be used in same image, for image optimization
> [overlapping and non-overlapping].
> * RLE compression for single bitplane images.

First, I think the approach is a bit wrong here. You'd like to
see the "all purpose" image compression method (german: "Eierlegende
Wollmilchsau", literaly "egg-laying wool-milk pick", which means:
Asking for too much at once, maybe "one size fits it all" in english),
which is, I think the wrong strategy. The right strategy is "pick the
right tool for the job". The tricks for bi-level image compression play
in a quite different league, and it requires much more than RLE to
be state of the art here. JBIG-2 is what you want if you want bi-level
compression.

> * Alternate parameter for compression ratio parameter: x.x bits/pixel.

This requires a different rate-allocation strategy. It is of course
possible to offer something like this by just running a control loop
around the codec and by that, picking the right quantization table.
Therefore, vendors could provide this feature while still creating
compatible codestreams. It doesn't magically turn JPEG-1 into an
embedded codec, though, should that be what you're asking for.
Thus, "ask your vendor for it if you need it", it's nothing the
standard would need to be extended for.

> JPEG 2000
> * RLE compression for single bitplane images.

That won't make it into the standard, because it's the wrong standard
to do it, see above. If you want to compress bi-level images, you're
better off with JBIG-2. You need to pick the right tool to do the job,
and bi-level compression is not the right job for JPEG-1 or JPEG2000.

> * ICER wavelet compression methods, IE: make 2 wavelet types available.

Part-1 has two wavelet types (9/7 and 5/3). If you need more types,
look into part-2. It allows you to use arbitrary wavelet filters and
arbitrary filter cascades. In other words: "It's already in, no need
to add it".

However, if I may add my two cents here: In my experience, it is not
worth it, unless you've very special requirements. The currently
available wavelet types work very fine for natural images. It only
makes sense to switch the wavelet if your images are in a different
image class. It makes sense, for example, for finger print
compression.

So long,
Thomas
.

Relevant Pages

  • Re: Microsoft HDPhoto algorithm exposed
    ... how a new image file format would be really 'new'? ... I'm not clear whether compression factors really matter too much. ... JPEG 2000 probably got like 10-20% better at ten times the complexity, JPEG-XR didn't get much better at about twice to three times the complexity. ... A considerably useful feature would be to provide a format that would allow me to store "relatively unprocessed" images, ...
    (comp.compression)
  • Re: The question probably asked a 100 times before..
    ... >edit them in Gimp and save as 100% JPEG file. ... my average file size at 5MP and no JPEG compression is about 4 MB. ... >quality images on my harddrive in case DVDs get lost / scratched. ...
    (rec.photo.digital)
  • Re: File Size Question
    ... In 8 bit JPEG format, it takes 3 bytes to represent one pixel. ... your 10 Megapixel camera makes IMAGES that are around 30 Megabytes. ... JPEG uses highly efficient lossey compression. ...
    (rec.photo.digital)
  • Re: Desperately needing direction with EncoderParameters and saving images under different compressi
    ... Save your images in png format instead of jpeg. ... and will preserve your pixel values exactly. ... That is the nature of lossy compression. ...
    (microsoft.public.dotnet.framework.drawing)
  • Re: Differential intra-interchannel filters for RGB images lossless compression
    ... >>> However IMHO many photographic images are even slightly more difficult ... don't look right, lighting doesn't look right, surfaces are almost always ... compression, it seems concievable to me that they could. ... would likely be able to fairly accurately predict the pixel. ...
    (comp.compression)