[fitsbits] Format for multi-readout, multi-amplifier data

Dear FITSbits-readers,

What would be reasonably standard formats for multi-readout,
multi-amplifier data?

In this case a 4-amplifier near-IR array in non-destructive read-out mode,
used both for imaging and spectroscopy?

The observers are unlikely to appreciate N_amp * N_readout extensions,
although this is the preferred format I guess from the FITS point of view.

Are there other formats in use, which allow for different amplifier
characteristics and are observer-friendly at the same time?


Best regards,

Saskia Prins

- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
Saskia Prins
Nordic Optical Telescope
Apartado de Correos 474 Tel: +34-922-425448
38700 Santa Cruz de La Palma Fax: +34-922-425475
Canarias, Spain
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -

Saskia Prins,

Are there other formats in use, which allow for different amplifier
characteristics and are observer-friendly at the same time?

I think the most observer-friendly format would be a data cube in
the primary header/data unit, with N_readout elements in the third
dimension. You could have a separate file for data from each of
the four amplifiers, or you could use four IMAGE extensions.

Phil

In article rao.edu,
Saskia Prins writes:
What would be reasonably standard formats for multi-readout,
multi-amplifier data?

In this case a 4-amplifier near-IR array in non-destructive read-out mode,
used both for imaging and spectroscopy?

Such arrays (and ones with even more readouts) are widely used in
astronomy. At the stage where you have reduced the data to a single
brightness number for each pixel, most astronomers will appreciate a
single primary image with x,y corresponding to pixel position on the
array. Then it's easy to apply calibrations and add a WCS in either
spectral or positional coordinates.

At stages of the reduction prior to the above -- for example if you
are using "sampling up the ramp" in post-processing -- do whatever is
most convenient for the data collection and reduction software. Few
if any humans are going to look at the data at that stage. I can
imagine using tables or a 3d image with the third axis being sample
number. For example, with IRAC on Spitzer, when we do pedestal-
signal sampling, the pedestal and signal frames are in separate
files. The point is not that this is a good idea or recommended
practice; it's that anything will work OK for the early stages. It's
only the later stages that matter to most users.

--
Steve Willner Phone 617-495-7123
Cambridge, MA 02138 USA
(Please email your reply if you want to be sure I see it; include a
valid Reply-To address to receive an acknowledgement. Commercial
email may be sent to your ISP.)

I assume the 4 amplifiers are reading quadrants of one physical
detector? If so, this is the same as the NICMOS and WFC3 IR
cameras on HST (well, *will be* in the case of WFC3). Since
the data from the 4 amps produces what amounts to one contiguous
image of the sky, that's the way we've always packaged them in
FITS files. You stitch the data from the 4 amps for a given readout
together into one image, stored in one image extension. So then
you have N_readout extensions, not N_amp*N_readout.

The same is done for our CCD instruments that use 2 amps to read
one physical chip: the data from the 2 amps is stiched together
into a single image, which is in turn stored in one image extension.
Only when there's a physical discontinuity, such as multiple chips
butted together into one focal plane, do we split the data into
separate FITS extensions.

-Howard Bushouse
WFC3 and ex-NICMOS calibration programmer


Saskia Prins wrote:
Dear FITSbits-readers,

What would be reasonably standard formats for multi-readout,
multi-amplifier data?

In this case a 4-amplifier near-IR array in non-destructive read-out mode,
used both for imaging and spectroscopy?

The observers are unlikely to appreciate N_amp * N_readout extensions,
although this is the preferred format I guess from the FITS point of view.

Are there other formats in use, which allow for different amplifier
characteristics and are observer-friendly at the same time?


Best regards,

Saskia Prins

- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
Saskia Prins
Nordic Optical Telescope
Apartado de Correos 474 Tel: +34-922-425448
38700 Santa Cruz de La Palma Fax: +34-922-425475
Canarias, Spain
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -