Images (where all the high-order bits are zero).
As one might expect, the best compression ratios are from blank I had a look at a few of the outlierĬases. The correlation coefficient between the Huffmann and "theoretical"Ĭompression ratio was 0.97. The values for HuffmannĮncoding ranged from 0.95 to 4.7 with mean 2.4 and standard deviationġ.0. Images, the "theoretical maximum compression ratio" ranged from 1.2 toĤ.8 with mean 2.7 and standard deviation 0.7. Image reflects the maximum possible compression ratio. Understand the algorithm, but I do understand that the entropy of the I used a feature of Andy Hammersley's program How good could lossless compression of diffraction images possibly be?Īt one time, I ran an entropy calculation on the 44968 images on /data at theĪLS 8.3.1 beamline. It would be great if we could losslessly compress our data a tremendous amount, but Mp3 files can be made at a compression ratio of 10:1 over CD-qualityĪudio and we all seem to still enjoy the music. Work very well at all at compressing sampled audio (about 1.3:1), but
I still remember the days before MP3 when it was lamented that sampled audioįiles could never be compressed very well. Yes, there are experts who can hear the difference (and some will even tell you to use vinyl), but if your model is never going to agree with the structure factors to better than 20%, then how much could adding 1% noise hurt? However, it would certainly be nice to have a way to keep some sort of "representation" of these raw data readily and rapidly available. In general, it is probably not advisable to mess with your raw diffraction image data in any way, but it does take up a lot of space! Personally, I highly recommend archiving your raw and uncompressed image data on some sort of stable, long term media, such as DVD-R, LTO tape, which currently have the lowest price/GB and are supposed to last 30+ years. Lossy compression of x-ray diffraction images Lossy compression of x-ray diffraction images Why lossy compression?
0 kommentar(er)
