Problems reported for the CGM*PIP Specification are entered as change requests in the POSC Request Tracker system. The problems appearing here are identified by the number assigned to them within that system.
Currently, there is only one change request that has been formally submitted against the CGM*PIP Specification since it was released by POSC in March 1995. It appears below.
For instructions for submitting a change request, see How to Submit a CGM*PIP Problem Report (Change Request).
Date entered: Tue May 9 09:22:46 1995
Status: Resolved
Submitter: Barry A. Tilton
Company: Schlumberger/GeoQuest
email: btilton@houston.geoquest.slb.com
Sample value handing not defined for background fill
9.2.2 Trace coordinates, scaling, and geometry (page 53):
The treatment of sample values when used to plot background fill is not defined.
Extensive treatment is given to how samples are handled when they are used to
generate coordinates for wiggle and VA fill plots. This section should at least
address whether or not the samples are scaled before being used to generate the
colour index or colour values used by background fill plots.
9.2.3.4 Background fill (page 57):
When sample values are used, are they scaled or not scaled? Again this section
does not define this issue and it should be addressed.
9.2.3.7 Variant colour specification (page 59):
The context in which "sample" and "sample value" is used can be implied to mean
"raw sample" and "raw sample value". The text in general leaves this somewhat
ambiguous by not specifying which. The definition of how seismic data is to be
rendered is core to this specification and how the sample values are used to do
this is key to the definition. Any ambiguities will lead to multiple definitions
by implementors and one of the key objectives of CGM*PIP was to eliminate divergent
implementations of seismic functionality in CGM.
General Discussion
As a member of the original CGM [PIP] project it is my recollection that the variant colour functionality in PIP II intended that the sample values be amplitude scaled prior to being used to determine the colour index. I have discussed this issue with [our other representative] on the PIP project and he concurs with my recollection. [My company] was one of the main promoters of and contributors to the variant colour functionality and our current usage of this functionality is consistent with the interpretation that the sample values be amplitude scaled prior to being used to determine the colour index.
One could put forward the argument that the colour scale value in the background fill variant linear element could be used as a sample value scaling factor. We see some basic problems with using the colour scale value for this purpose. First the amplitude scale factor in the trace scale factor was intended to be the value used for scaling the data. Our intent for the colour scale value was to support a "display range" scale value for this mode. The sample normalization extent value in the trace element is the "display range" scale value for wiggle and VA plots but the definition of value shows that it is to be used to scale VDC space and not colour space. Finally the background fill variant lookup table element does not support a colour scale value because the finite mapping of sample values to a finite sample range mapped to a specific set of colour indices does require "display range" scaling. In the lookup table mode the only appropriate defined sample scale value is the amplitude scale factor.
Our interpretation of the handling of samples for linear background colour fill mode is that each sample in the trace primitive be amplitude scaled, then scaled by the colour scale value, then offset and finally clipped to the min/max indices to produce the colour index. For the lookup table mode the sample values in the background fill variant lookup table element should be amplitude scale scaled by the renderer. The samples in the trace primitive element should also be scaled by the amplitude scale factor prior to being used to select the respective colour index from the lookup table.
Should an application that generates CGM*PIP/II wish to or need to do its own amplitude scaling the amplitude scale factor can be set to 1.0. Other applications that need the renderer to scale the sample data can set the amplitude scale factor and know that the renderer will scale the sample data appropriately.
In summary, the Plot/Hardcopy Work Group disagrees with the submittor's assertion that the suggested defect is in fact an error and misstatement of committee intent in the PIP document. The PIP document unambiguously states scaling rules in equations. However submittor's suggested interpretation is considered to be a useful option, and addition of such as an option would not comprimise existing implementations which interpreted the PIP document literally.
The defect resolution is in the form of a set of editing directives against the base document (CGM*PIP/II). The complete resolution and detailed editing requirements are on file at POSC and will be incorporated into a maintenance release of the document. A summary of the substantive technical changes is shown below. Note that the resolution is a substantive change affecting only PIP/II, therefore the PIP/II maintenance release number must be changed but the PIP/I maintenance release number need not be changed.
Substantive Technical Changes:
Page 59: replace the equation "Cn = (Sn)(Fc) + B" with
the new text:
Cn = SnFc + B or Cn = SnFaFc + B
Page 59, the table: Replace all occurrences of "sample" with "effective sample" and after the table add the text, "The effective sample value is either the sample value itself, or the sample value pre-scaled by the factor Fa, depending on the value of the parameter of the Variable-Area Fill Variant Colour Selection ESCAPE, or the Background Fill Variant Colour Selection ESCAPE (as appropriate). The factor Fa which is use to pre-scale the samples in the latter case is the Amplitude Scale Factor (see 9.2.2.3)."
Page 90, section 9.3.6.8: In the Description, replace "`GDP data', `linear', or `lookup table'" with "`GDP data', `linear', `lookup table', `linear prescale', or `table prescale'". Replace "If `linear' or `lookup table' then" with "Otherwise".
Page 91, table of enumerative values: add two lines:linear prescale 3 LinearPrescale table prescale 4 TablePrescale
Page 95, section 9.3.7.7: In the Description, replace "`GDP data', `linear', or `lookup table'" with "`GDP data', `linear', `lookup table', `linear prescale', or `table prescale'". Replace "If `linear' or `lookup table' then" with "Otherwise".
Page 95, table of enumerative values in 9.3.7.7: add two lines:
linear prescale 3 LinearPrescale table prescale 4 TablePrescale
| CGM*PIP Index Page | POSC Home Page |
Copyright © 1994, 1995, 1996 Petrotechnical Open Software
Corporation. All rights reserved.
POSC ® and the POSC Logo are registered trademarks and Epicentre™
is a tradmark of Petrotechnical Open Software Corporation.