AWARE SYSTEMS
TIFF and LibTiff Mail List Archive

Thread

2000.03.30 10:53 "Complex Floating Point", by Antonio E. Scuri
2000.03.30 13:33 "Re: Complex Floating Point", by Frank Warmerdam
2000.03.30 13:56 "Re: Complex Floating Point", by Antonio E. Scuri
2000.03.30 15:07 "Re: Complex Floating Point", by Chris Hanson
2000.03.30 16:44 "Re: Complex Floating Point", by Antonio E. Scuri
2000.03.30 19:12 "Re: Complex Floating Point", by Cris Luengo
2000.03.31 11:34 "Re: Complex Floating Point", by Antonio E. Scuri
2000.03.31 15:14 "RE: Complex Floating Point", by Martinez, Max
2000.03.31 18:07 "RE: Complex Floating Point", by Martinez, Max
2000.03.31 17:05 "RE: Complex Floating Point", by Antonio E. Scuri
2000.03.31 18:12 "RE: Complex Floating Point", by Antonio E. Scuri
2000.03.31 18:31 "Fwd: RE: Complex Floating Point", by Antonio E. Scuri
2000.04.03 13:12 "Re: Complex Floating Point", by Frank Warmerdam
2000.03.31 18:48 "Re: Complex Floating Point", by Daniel McCoy

2000.03.30 13:33 "Re: Complex Floating Point", by Frank Warmerdam

I would like to store complex floating point images from FFT results in a TIFF file.

Although TIFF has support for single and double precision floating point, there is no definition for a complex pixel value.

If I use complex double precision floating point only, I can use SampleFormat=IEEEFloatingPoint + BitsPerSample=128. But if I use BitsPerSample=64, this is the same of real double precision, so can not use single precision for complex values using this approach.

In the case of non image Tags, also the definition of field type should receive two more options 13 and 14, for single and double precision floating points.

  1. Is there anyone using TIFF with complex images? If so, which were your solution to the problem?
  2. Do someone know if there is any complex floating point support in the TIFF 7.0 spec?

Antonio,

I would like to try and discourage you from producing files with 128 bits per sample to represent complex data. Couldn't you just produce files with two samples, and 64 (or 32) bits each? I realize there wouldn't be any explicit way of recognising them as a real/imaginary pair, but perhaps that could be accomplished by other means?

Best regards,

---------------------------------------+--------------------------------------
I set the clouds in motion - turn up   | Frank Warmerdam, warmerda@home.com
light and sound - activate the windows | http://members.home.com/warmerda
and watch the world go round - Rush    | Geospatial Programmer for Rent