2010.06.21 21:49 "Re: LibTIFF compression enhancements", by Stephan Busch

2010.06.21 21:49 "Re: LibTIFF compression enhancements", by Stephan Busch

Dear Frank,

thank you very much for your answer.

I tried to join the mailing list but the website tells me the list does not exist.

So far I haven't heard of any TIFF libraries that use jpeg2000 algorithms, but its sad that all those maintainers could not agree to use a unified version all libraries could read.

LOCO-I is the algorithm behind JPEG-LS. It was designed a decade ago by HP Labs, gives better compression than jpeg2000 in most cases and there are many implementations; some of them are open-sourced such as: www.stat.columbia.edu/~jakulin/jpeg-ls

In the case neither jpeg2000 nor jpeg-ls are implementable, I wanted to suggest another codec that will be free and open-sourced soon. The algorithm is called PackPNM at present and it performs like jpeg-ls.

Also, in my opinion only the codec that you will chose might survive. All high performance image codecs have been designed but were never implemented for practical use. Why should we still use LZW or LZ77 when we have the chances and source codes to do noticeably better.

Think of all the medical images and photography archives that would take less space.

Maybe, there is a chance to chose an alternate rather than solving compatibility issues.

Is there another way to discuss this with the whole mailing list?

Getting broader feedback would really be interesting. People can move much more in a group.

Looking forward to your answer

Best regards,

Stephan

_______________________________________________________________________________________________________________

> Stephan,

> I have cc:ed Andrey's working address. The remotesensing.org
> address no longer works. My comments are below.

> Squeeze Chart 2009 (Stephan Busch) wrote:
>> Dear LibTIFF team,
>>
>> I am interested in lossless data compression since 1993 and do file
>> compression benchmarks
>> in my spare time. And since I am interested in photography, one of my
>> topics is lossless image compression.
>> Please, feel free to visit www.squeezechart.com
>>
>> Like JPEG, TIFF is more than 20 years old and has gone a long way to
>> become what it is today.
>> It is the only lossless image format that support all kinds of metadata
>> - also the upcoming face tags,
>> which would be lost when you edit your digital camera image and save it
>> in other formats.
>>
>> But my question is: Why doesn't TIFF support stronger compression
>> algorithms such as
>> lossless JPEG2000 or the LOCO-I technology? Is there a conflict of the
>> licenses invoked
>> or is there no demand for better compression?
>>
>> In my opinion the only alternative to Adobe DNG is TIFF, because it
>> supports also 48 bit image
>> which only a few formats do support. TIFF seems to be used in pre press
>> environments,
>> by enthusiast photographers and professionals all over the world and I
>> want to suggest better
>> lossless compression to make TIFF files smaller and more competitive in
>> comparison to
>> old-style deflate and lzw formats.
>>
>> What do you think?

>> Is there a chance to propose better compression in .TIFF files?

>> And would you implement a newer algorithm as an additional choice?

> I have not personally pursued new compression options for libtiff because

> I have not had significant client interest. I have seen that some other

> tiff libraries have added non-standard codes for jpeg2000 and I would like > to see jpeg2000 support added in libtiff in a way that is compatible with

> those implementations. I would be very sad however to see it done in way > that was incompatible with other existing practice.

> I am not familiar with LOCO-I.

> I would encourage you to raise your questions on the mailing list for broader > feedback.

> Best regards,