JXL: JPG/ GIF Embed-Mode

[nji9]

JXL has the option of lossless embed JPG and GIF format.
The advantage compared to JXL-lossless is that
- it encodes about 30 times (!) faster and
- the file size is even smaller.

I took a 9 MPixel JPG (98%, 4-4-4)
www.lapseoftheshutter.com/printer-test-image/
File sizes
Orig: 7.2 MB
JXL lossless: 8.6 MB in 35s
JXL (JPG embed): 6 MB in 1s

The JXL embed option seems most useful for lossless;
but also if you're not too keen on small file sizes
(JXL -q90 (default) encodes to 1.6MB in 6s)

So my suggestion is to offer the "embed" option
in the JXL save options.

A checkbox should do like "[x] Embed mode"
beyond the other options.
It should be visible only if the prerequisites
are fulfilled (unchanged JPG or GIF)
If it is checked the both options above should be grayed.

So this would be most easy to handle
while the user has all the options JXL offers.

[Kadet]

+1

This should be added to the Export function.

[nji9]

Checked in v.1.4.3.
... still doesn't use the "embed mode" for lossless JXL save of JPG.
Due all of its advantages. Is there a reason for it?
The lib supports it.

[nji9]

I checked in newest XnViewMP v.1.7.1.
Still JXL lossless encoding of an (untouched JPG or GIF) produces the wrong format.

Please have a look at it, as it is one of the important advantages of JXL:

Encode JPG/ GIF lossless in about "zero time", and to a size of about 80% (of the orig file).

[masterjp]

I understand, that you need the loseless conversion of JPG files to JXL.
It is a very useful and HDD saving feature.

There are a few tools at the moment, which can do this.
The newest and best tool with JXL 0.10 Support is:

https://github.com/JacobDev1/xl-converter

It has several other features, which are useful, too.

I think XnViewMP will have this feature in the future.
But currently other bug fixes and changes are more important than full JXL support.
Furthermore you can write an small Batch cmd file to convert all JPG files within a folder or directory structure to JXL by yourself.

[nji9]

Thankyou for your valuable reply!