Tom wrote: Hi, > now i am designing a colorimeter, as a hobby work, with a TCS34725 color > light sensor > from AMS(former TAOS). constant current led drive is implemented to provide > stable > illumination(even though it's a bad idea to use a colorimeter to measure the > characteristics of reflective materials). It's quite difficult to transform a colorimetric reading using one particular light source (i.e. a white LED) into one for a standard light source such as D50 in a way that is independent of the spectral characteristics of what you are measuring. Reflective colorimeters have traditionally used an incandescent light source, because it has some similarity to D50, or can be filtered to be like D50. > 1) how can that little thing talk to argyllcms? Short answer: you would have to write a driver. There are numerous examples, from the simple to complex in the ArgyllCMS spectro/ source (i.e. ColorHug, Huey, DTP51, etc). > does argyll have some kind of measure to communicate with colorimetry devices > that are > not directly supported by the binaries? like loading a custom dll, or making > a tcp > request to some port? or do i have to provide you with a datasheet of the > device, and > the device itself, in order to let you write some code about it? There are some hooks for using an external program, but they are limited, and more suitable for display measurement, not reflective measurement. Sorry, I'm not prepared to write a driver for a one off instrument, so it would be up to you. > i am not very familiar with native windows C programming(with serial > communication and > all those api stuff), nor do i have much experience to code in a pro style > like argyll. > for me it could be a fun of trial and error, or it could be a waste of time. Most of the very low level aspects are already taken care of in terms of accessing USB or serial instruments. But a good understanding of C programming would make it easier. > 2) what USB-TTL(serial) chip should i choose, to avoid trouble for linux > users? i > myself is a dedicated windows user. but i want my work to be useful for > everyone. here > in china a lot of options are available: > > > FTDI FT232 : stable, known to work with linux, Use this one. There is already code for recognizing instruments that use this chip (Klein K-10 & JETI 1211). > 3) a question: if the filters built inside an RGB sensor have exactly the > same spectrum > sensitivity characteristics to the LMS sensitivities of human eye, can i > obtain > accurate XYZ readings from it? Yes, but to be standard graphic arts XYZ, the illuminant needs to be D50 spectrum. Graeme Gill.