
Hybrid audio call, April 23, 2013
=======================================


Participants:

Peter, Savvas, Chulhee, Naeem, Nicolas, Romuald, Marcus, Roland, Margaret, 



Agenda:

1. Source scene pool: status

2. PVS generation: status

   - HD1, HD2 and use of High-4:2:2-profile

3. Common set: status

4. Other  



Call:
--------------------


1. Source scene pool

   HD05 : replacements, still under progress
   


2. PVS generation:
   
   
   - Use/drop High-4:2:2-profile:

     Some arguments:
     
     keep:   + already generated
             + Margaret thinks this is not excluded from test plan

     re-do:  + this profile is not within the scope of test plan (not suggested in test plan either)
             + no test vectors with this profile, therefore, some models do not support it
             + test design of HD01 uses the Tlabs decoder, cannot be used using this profile
             + including High-4:2:2-profile would make it very difficult for encrypted models, as the
               bitrate-quality relation is possibly very different (note: High-4:2:2-profile uses
               10 bits per pixel and colour)


     --> Decided: do NOT use High-4:2:2-profile.

     Should make a clear statment about this in the final report, e.g.
     --> models in recommendation should not be used for High-4:2:2-profile.

     --> this decision implies: re-process HD01, HD02 databases.


   - Overview:
     

     HD01 : to be re-processed
     HD02 : to be re-processed
     HD03 : ready for verification
     HD04 : finalized, adjusted filenames of pcap files to comply with test plan          
     HD05 : source replacements in progress (see 1.)
     HD06 : Ghent University offered help (as did Opticom before), no update yet.

     VGA01 : finalized     
     VGA02 : uploading, there is a problem with src04 if compressed (black frames at beginning), therefore as avi files
     VGA03 : probably starting uploading tomorrow, but there is an issue with HRC19, using the 3G network.
             
     Opticom proposes to use 'wild www DSL' instead for HRC19.
     --> accepted.
             

     WVGA01: uploading
     WVGA02: no update (rebuffering experiment)
     

     Note: decoding using definitelyLossless-tool adds '_dec' to the filename. 
     
     

    
3. Common set:

   Situation: the common set sources are in different formats. Format conversion need to be performed, first, to
   process the PVSs, and second, for inclusion of the videos in the subjective tests.

   --> accepted: Multi-media: use frame rate conversion. Check the resulting files. if there is a visual difference, 
   another conversion method could be used, but should be discussed. 
   

   HD: 
   Proposition (as done for the common set PVSs using an HRC of HD04): Convert videos first to corresponding frame rate
   (e.g. for HD04 this is 25fps) then process PVSs. Then a final conversion is performed for inclusion of the video
   in all subjective test datasets.
   

   Proposition: Nicolas will check, if Ghent can perform all final conversions. 
   E.g. a PVS is created using and HRC of HD01, i.e. 1080i60. The final conversion is then the conversion of this PVS
   to the format of all other databases, in which this PVS will be shown in the subjective test.


   --> Proposal: create the common set PVSs as quickly as possible, then the group can inspect them.


   Naming convention for the common set:

   --> proposition by Silvio, see email of April 23.

   


4. Other: --


   

For Information: evaluation data sets:

-----------------------------------------------------------------------------------------------------------------------------------
|        | Format                        | Scene     | Design                | Test vectors    | PVS generations | Subjective Test |
-----------------------------------------------------------------------------------------------------------------------------------|
| HD01   | 60fps (interlaced)            | IRCCyN    | DT                    | RT-RK           | RT-RK           | IRCCyN          |
| HD02   | 50fps (interlaced)            | AGH       | AGH                   | RT-RK           | RT-RK           | AGH             |
| HD03   | 30fps (progressive)           | DT        | DT                    | DT              | DT              | DT              |
| HD04   | 25fps (progressive)           | Ghent     | Ghent                 | Ghent           | Ghent           | Ghent           |
| HD05   | 60fps (interlaced)            | Yonsei    | Yonsei                | Yonsei          | Yonsei          | Yonsei          |
| HD06   | 50fps (progressive), 1280*720 | FUB       | FUB                   | FUB             | FUB             | FUB             |
|        |                               |           |                       |                 |                 |                 |
| VGA01  | 30fps                         | Swissqual | Swissqual/Rebuffering | Swissqual       | Swissqual       | Yonsei          |
| VGA02  | 30fps                         | Yonsei    | Yonsei                | Yonsei          | Yonsei          | Swissqual       |
| VGA03  | 25fps                         | Opticom   | Opticom               | Opticom         | Opticom         | Yonsei          |
|        |                               |           |                       |                 |                 |                 |
| WVGA01 | 30fps                         | Yonsei    | Yonsei                | Yonsei          | Yonsei          | Opticom         |
| WVGA02 | 25fps                         | Acreo     | Acreo/Rebuffering     | FUB/RT-RK/Acreo | FUB/RT-RK/Acreo | Acreo           |







