PDS_VERSION_ID         = PDS3
RECORD_TYPE            = STREAM
OBJECT                 = TEXT
  PUBLICATION_DATE     = 2005-04-22
  INTERCHANGE_FORMAT   = ASCII
  NOTE                 = "Errata for Cassini ISS EDR and 
                          Calibration Files Archive Volume DVDs."
END_OBJECT             = TEXT
END


COMMENTS ON THE CASSINI ISS EXPERIMENT DATA RECORDS AND CALIBRATION 
FILES ARCHIVE COLLECTION

This document contains comments and errors concerning the Cassini
ISS raw experiment data records and calibration files archive 
published on this collection of DVDs.  

This errata file is modified as problems/issues are encountered 
during the active mission lifetime, and result in changes or 
corrections, or notes worthy to the end user. For the most 
up-to-date version of this file look on the internet web site 
of the PDS Imaging Node:

    http://pds-imaging.jpl.nasa.gov/

----------------------------------------------------------------------


------------------
DATA Volume Errata
------------------

1. PDS Standards Exceptions

As part of the ISS index.tab files, the keyword DETECTOR_TEMPERATURE
has been represented with measurements in units Celsius (C).  Per the 
PDS Standards Reference Document, this keyword is typically found as 
units measured in Kelvin (K). See Appendix A, Cassini ISS Archive 
Volume SIS, JPL D-27548 and Cassini ISS Tour VICAR Image Data 
File and Detached PDS Label SIS JPL D-24724) for more details 
on this keyword.



2. Cassini ISS Ground Calibration Files

The ground calibration image files acquired pre-launch from the 
Cassini ISS instrument were produced and retained on a collection 
of CD-ROMs intended to be PDS-compliant. The PDS Imaging Node 
converted the dataset from CD-ROMs to DVDs for inclusion in this ISS 
archive collection. They can be found on the first ten calibration 
volumes.  



3. SPICE-generated Keywords

Keywords calculated using SPICE may use a c-smithed c-kernel (as 
indicated in the SPICE_PRODUCT_ID keyword) in replacement of a 
reconstructed ACS c-kernel.  The c-smithed c-kernel is produced by
automated auto-navigation software on a subset of the images.  In most
cases it is more accurate than the ACS c-kernel.  The kernel takes 
the naming convention "<image name>.bc".  For example: 
"N1363625260_2.IMG.bc.  

This c-kernel is delivered to NAIF some time afterwards in a concatenated 
file containing a larger time range of c-smithed c-kernels. This file 
takes the naming convention "<yy><doy>_<yy><doy>c<version>_ISS.bc". 
For example: "00275_00285ca_ISS.bc".  
To find the original c-smithed c-kernel, convert the image SCLK time to 
a day-of-year date to find the concatenated c-smithed kernel it is 
contained within. 



4. TELEMETRY_FORMAT_ID keyword problem

The label keyword value TELEMETRY_FORMAT_ID will be set to UNK for all images 
from Saturn Tour.  This is due to a flight software upgrade in sequence C37 
(starting at SCLK 1431917802) that routed the channelized engineering data to 
another partition, which isn't downlinked unless there is an emergency, so 
the ability to determine the telemetry format was lost.  The 
INSTRUMENT_DATA_RATE keyword can be useful to help figure out how fast data 
was being sent to the CDS, and is somewhat an indicator of the telemetry 
mode, although it is not deterministic (i.e., some telemetry modes share the 
same instrument data rate.)



5. PDS Label pointer discrepancy

There is a problem with the PDS label generation portion of the telemproc.  
The pointers were inadvertently hardcoded in the PDS labels.  A fix will be 
issued in March of 2005 and efforts will be made to fix all affected labels
previous to this date.



6. BIAS_STRIP_MEAN label value

Before sequence C32 (before SCLK 1401927444), when FLIGHT_SOFTWARE_VERSION 
= 1.2, this uses one overclocked pixel value per line. This value should be 
ignored for LOSSY compressed data as it does not contain valid data. 

During sequence C32 (beginning at SCLK 1401927444), when 
FLIGHT_SOFTWARE_VERSION = 1.3, this uses six overclocked pixel values per 
line. This value should be ignored for LOSSY compressed data as it does not 
contain valid data.

When FLIGHT_SOFTWARE_VERSION = 1.4, for Lossy compressed data, this value is 
the mean of the overclocked pixel sum returned in the last compression block 
of the image.



7. DARK_STRIP_MEAN label value

For FLIGHT_SOFTWARE_VERSION_ID=1.2 or 1.3, this value should be ignored for 
LOSSY compressed data as it does not contain valid data.

When the FLIGHT_SOFTWARE_VERSION_ID = 1.4, for Lossy compressed data this 
value is the mean of the extended pixel sum returned in the last compression 
block for the image.



8. Filename case consistency

All filenames and directories on the DATA volume are in lower-case with the
exception of the image vicar and label files in the data directory.  There
are possibly some inconsistencies in the documentation where it may reference
the file by the wrong case.



9. Rings_Flag keyword fix

The Rings_Flag in the index.tab file was being set to true only when the A, 
B, C, or D rings were visible in the image.  This condition was updated to 
also include the E, F, and G rings in this test.  This affects archive disks 
COISS_2007 and later (image mid time 2004-272T00:48:55.974 and beyond).  
The Rings_Flag definition in the index.lbl file was also updated to reflect
this change.



10. Target_Distance keyword could have been named Target_Center_Distance

The Target_Distance keyword in the index.tab file calculates the distance 
from the spacecraft to the center of the target body.  However, PDS defines 
Target_Distance as "to a point on the target body".  Target_Center_Distance 
could have been chosen as this keyword name to more accurately match the PDS 
Data Dictionary, but it was not.



11. UNITS keyword added to index.lbl keyword definitions

The UNITS keyword was added to each COLUMN object in the index.lbl keyword
definitions file.  This new index.lbl file will appear on volumes COISS_2007
and later.  Index keyword units are also defined in the Appendix A of the
archsis.pdf file.



12. Incorrect 'Not Applicable' keyword values

The following keywords are incorrectly set to the value of '0' (zero) 
instead of the correct value of 'N/A' in cases where the values are indeed 
'Not Applicable'.  This has not been fixed as of this archive delivery. 
(Target_Aimpoint_Emission_Angle, Target_Aimpoint_Incidence_Angle, 
Target_Aimpoint_Phase_Angle, Target_Aimpoint_Latitude, 
Target_Aimpoint_Longitude, Target_Aimpoint_Distance) 




----------------------------------------------------------------------




-------------------------
CALIBRATION Volume Errata
-------------------------

Calibration activities for the ISS cameras are on-going, and they are 
incomplete at the present time. Some calibration algorithms have yet to be 
implemented or are only partially implemented. In addition, many future 
improvements to the current calibration algorithms have been identified. Some 
of these outstanding items are listed below.

1. PDS Standards Exceptions

Many of the PDS labels for the data files on the calibration volume, as well
as for the sample calibrated images, have yet to be finalized. We expect some
errors to be present in these PDS labels, and will work to fix them as they
are made known.



2. Dark current

The current dark current modeling algorithm is divided into two parts: a 
code that examines a set of dark images and creates a parameter file from 
them, and another code that takes that parameter file and constructs a dark 
image appropriate for a specific set of camera parameters. Together, this
dark model is referred to in the CISSCAL calibration software as the
"interpolation model."

The first code has been written by Bob West, but it is still being developed. 
Once complete, it will be included in future versions of the calibration 
archive volume. For now, the calibration volume contains only the parameter 
files produced by this code for NAC and WAC images. These can be found 
within the darkcurrent subdirectory of the CALIB directory. 

The second code that actually generates the dark current files is still 
being developed as well, but the current working version has already been 
integrated into CISSCAL.

There is also an alternate dark subtraction method, which is referred to 
in CISSCAL as the "two-parameter model." This method is outdated and 
generally produces unsatisfactory results, but has been included for the
time being for completeness while the interpolation model is being 
finalized.



3. Flat fields

Each filter combination has a signature flat field pattern that needs to be 
divided out of each image during the calibration process. This pattern is 
captured in the "slope files" which were derived from ground calibration 
images, and they are included on the calibration volume. Unfortunately, the 
flatfield is not constant with time: its structure changes whenever a dust 
grain lands on one of the filters of the CCD window, or an already-present 
dust grain simply shifts position.

Analysis is currently underway to determine the various flatfield changes 
that have occurred since launch, so that the appropriate corrections can be 
made to the flatfielding calibration step. So far only two of these 
corrections have been determined, both of which apply to the NAC only: a 
correction for a large dust ring discovered during Venus flyby, and another 
correction derived from Titan flyby data that removes a residual "mottle 
map." Both of these corrections are supplied with the calibration volume. 
Similar filter-specific and camera-specific flatfield corrections will be 
available in the future. 

The slope file subdirectory of the CALIB directory contains three database
files which associate the various filter combinations with their appropriate
slope files. The first of these, slope_db_1.tab, is the original slope file
database. The third database file, slope_db_3.tab, has been altered to take
into account new slope files generated for the NAC that have been corrected
for anti-blooming pixel pairs. Technically this is the "correct" database 
file. Unfortunately, some of the slope files to which it refers have since 
been found to be unacceptably noisy. Thus, the database file that is used
by CISSCAL and currently considered "best" is the second, slope_db_2.tab.



4. Absolute Calibration

As detailed in [PORCOETAL2004], absolute calibration of the ISS cameras is 
currently incomplete. Quantum efficiency correction functions have been 
derived for both cameras, but filter-specific absolute correction factors 
have so far only been calculated for the NAC. Currently, absolute flux in 
the NAC is only known to within an error of about 10-20% depending on filter 
combination (better for single broad-band filters than for narrow-band 
filters or filter combinations), and 15-25% in the WAC.

Also note that some filters, such as the polarized, methane, and continuous-
band filters, require special attention for proper absolute calibration. This 
analysis is currently underway.



5. Temperature Dependency

Many of the phenomena detailed during ground calibration testing were found 
to exhibit temperature dependency. Examples are the uneven bit-weighting, 
flatfield, and shutter offset corrections. Unfortunately, the algorithms 
currently implemented for these calibration steps do not give a realistic 
treatment of this temperature dependency. Attemps will be made to correct
this shortfall in future versions.



6. Coherent Noise Sources

Calibration algorithms have yet to be developed for some types of coherent 
noise found in the image data, such as the vertical banding.