1. Generic questions

1.1 That's a lot of artefacts! Do you call that a succesful experiment?

Read the paper again, the answer is in it.

1.2 In page XXX, I don't understand your explanation about YYY

100% of the Trolls we asked found (after barely reading the paper while talking to somebody else at the same time) that "it was very clear, an impressive paper!".

However, you can email me should you have any question or comment.

1.3 Wow, you rock! Can I offer you a job in my company, with a big fat salary?

Sure, email me!

2. Comparison questions

2.1 Why did you not compare to Wavelet compression?

Simply because we could not find an implementation of wavelet compressions. Wavelet compression have been advertized for quite a while, but they do not seem to compete against JPEG: the gain of compression rate just does not seem to outweigh the need for more computational power.

Or, maybe the distribution of this particular format was too confidential.

2.2 What about JPEG2000?

JPEG2000 proposes a lot of enhancements, most of them being implementation of a wider set of images. From what I read, the compression rate is not that much better than JPEG, but the artefacts seem to be less visible.

3. Implementation questions

3.1 Why did you use gzip? bzip2 is much better

Gzip uses very little computational power compared to bzip2. Bzip2 is a good compression format for large files (>1MB), but the results are sometimes worse than gzip on small files.

If you find bzip2 better, go ahead and use it. If we did not include gzip in the definition of the VCF, this is exactly because we want everyone to use the lossless compression format best suited for his application.

3.2 When do you plan to release a usable version of the VCF?

Hopefully in a few months.


Last modification: March 2002, © Eric GAUDET, 2002