Thanks to all those who replied... Some people, it seems, read the 1:50k tiles ok. Others convert them before compressing. In my own previous experience with MrSID, one published limitation was that the compressor could only accept 24bpp (true colour) images, hence we had to convert the original 256 colour images before compression. This reduced compression rates by a third, since moving from 8 to 24bpp increases the size threefold... In my recent adventures, I tried using Wang Image Edit ActiveX control in a VB app, to batch convert the tiles to 24bpp. The resultant images compressed OK in the freebie ECW compressor (2.2 beta). The target compression ratio was 50:1, the acheived ratio was 18:1, thus a net compression of 6:1 was acheived, reducing the 1:50k tiles from 15Mb to 2.5mb... which was nice! I thought I'd experiment further; I used the same Wang/VB app to convert the uncompressed 8bpp paletted tiffs to uncompressed 8bpp paletted tiffs(!). Funnily enough, the resultant images WERE accepted by ECW, but the compression procedure reported an actual compression of around 11:1, with a files size of 4Mb. It doesn't take a genius to figure that 4Mb x 11:1 = 44MB approx - 3 times the original image size. Hence it appears that ECW too works with 24bpp images... Food for thought... Thanks once again to those who helped. Chris PS. My 'experiments' on the 8bpp tiffs weren't conducted under lab conditions, and were done with only a small sample of files etc...so if I've gone wrong somewhere, don't tell me - this has been the most interesting thing to happen to me this week _______________________________________________________________________ List hosting provided by Directions Magazine | www.directionsmag.com | To unsubscribe, send e-mail to [EMAIL PROTECTED] and put "unsubscribe MapInfo-L" in the message body.
MI SUM: MI-L Compressing 1:50k Raster with ECW
Martin, Christopher (WSA Telford) Fri, 27 Oct 2000 08:00:51 -0700
