Bug 303242

Summary:	File type should not create distinct results in fuzzy duplicate search
Product:	[Applications] digikam	Reporter:	Kyle Altendorf <kyle.altendorf>
Component:	Searches-Similarity	Assignee:	Digikam Developers <digikam-bugs-null>
Status:	RESOLVED NOT A BUG
Severity:	normal	CC:	caulier.gilles
Priority:	NOR
Version:	2.7.0
Target Milestone:	---
Platform:	Ubuntu
OS:	Linux
Latest Commit:		Version Fixed In:	2.8.0

Description Kyle Altendorf 2012-07-09 16:36:32 UTC

I understand that some people do not want RAW formats to even be shown in duplicate searches, but if they are going to show (as they do for me now) they should not be considered distinct from the JPG that they correspond to.

Reproducible: Always

Steps to Reproduce:
1. Shoot DNG next to JPG in camera
2. Duplicate both DNG and JPG to other names
3. Import to digiKam
4. Fuzzy search on album containing above images
Actual Results:  
Two entries will exist in the results.  One for the pair of JPG and one for the pair of DNG.

Expected Results:  
All four images should exist within one group.

This gets extra confused when combined with:

https://bugs.kde.org/show_bug.cgi?id=303239

Further, a net yet well defined or filed (at least by me) bug causes duplicates to end up in the groups of their parent images...  I think.  Still need to isolate what is actually happening before I officially report.

Comment 1 Marcel Wiesweg 2012-07-22 15:17:41 UTC

I assume that the (camera-generated) JPEG is in detail different from a RAW converted in fast mode, maybe the algorithm actually recognizes them as different. I could confirm if you could provide a pair for testing.

Comment 2 Kyle Altendorf 2012-07-24 01:26:23 UTC

duh! :[  I reduced the similarity constraint and they show up together.  The previews (SX40 with CHDK) are fairly different in both noise and exposure so this should have been the obvious answer.