- Wed Aug 03, 2016 11:16 am
#2026
ISDCF Main Meeting Notes – May 19, 2016
Upcoming Meetings
Thursday July 21, 2016 Main meeting at Universal
Thursday September 22 Universal
* Traveling Plugfest - November 14-17 2016
Part 1: General Reporting
Housekeeping:
Antitrust Disclaimer – The official antitrust guidelines are posted on our website and are linked from the main ISDCF page. A short verbal overview of guidelines was given.
Thanks to Fox for lunch. Upcoming (note that the dates are messed up): Barco June ’16, August ’16 WB/Mike Z, Jim W/Wendy A (Consultants) Sept ‘16
Thank you to Universal for the facilities and parking. Universal will be the primary location for future ISDCF meetings - although Uni may loose the conference room
Thank you to Universal for support of the Chairman.
Thank you to Inter-Society for providing the funding for notes, travel, general expenses and admin support.
Thank you to Universal/Sony for the coffee and treats
Meeting notes from March ‘16 was reviewed and approved.
Legal reminder / press reminder
Chatham House Rule:
When a meeting, or part thereof, is held under the Chatham House Rule, participants are free to use the information received, but neither the identity nor the affiliation of the speaker(s), nor that of any other participant, may be revealed.
Attendance is at end of these notes.
Action Items from May 19, 2016
Start a thread on Forum on Event Cinema Automation needs / workflow solutions. (CJ and Jerry)
Eikon needs to get Jerry the variable Z test package for posting on ISDCF (the one that was used at the plugfest.
Distributors will ask their movie makers on the absolute need for continuous variable Z - is it good enough to just have a set of fixed Z subtitle placement - would that be acceptable for using this feature?
Get results from EDCF for ‘Stair step’ static Z positioning to emulate variable Z.
Subgroup to edit / recommend changes for Document 12 - RP for Ingest Behavior - Steve L to lead, Dean B, John H, Mike R, Jim W, Chris W, Bill E “volunteered” to participate.
Create an updated suite of well described test content on the ISDCF website (Jerry) Including SMPTE DPROVE content.
Try Dolby conference system - Dean will try to get a system for us to try.
From earlier meetings:
Wendy will check with John H and SMPTE to see if there is a list of SMPTE specs that could be posted on the ISDCF website.
Review the use of ContentTitleText in the PKL for situations where the PKL holds multiple CPLs. Language to improve the Digital Cinema Naming Convention site OR may become a new recommended practice document for ISDCF (Chris W)
Test content to show the difference between font rendering color and expectations. i.e. create a font color that should match the background and blend in - this would show the issue. Should be ready by plugfest in May.
Add to the SMPTE CPL metadata document the use of lower case subtitles indicating burned in subtitles and upper case is rendered. Dean [Post on Forum.]
Still looking for DCP that includes Atmos in test content B configuration (Dean) Should be ready by plugfest in May.
Still a need for a simplified description of FLMx for what it does and doesn’t do (to make sure it won’t be rejected due to lack of knowledge of what it does and doesn’t do.
ISDCF should have a recommended practice for zip/tar content to create drives (NOT for distribution to theaters.) Dean? Nick? Steve?
ISDCF should write description of submission aspect ratios to the Academy. (Ernie / Bill H. will take the lead.)
We had discussion on when the next ISDCF meetings will take place. We deleted the June meeting, we set a July 21 meeting date, we deleted the August 18 date and set the September 22 dates. We also set a traveling plugfest for November. We discussed the need to NOT have meetings closer than 4 weeks apart.
It was pointed out the meeting notes from 10 years ago (at end of last meeting notes).
Discussion of iPath Reflector - still needed? Should we kill this reflector or make it more known? Results: keep the iPath list. Send the current membership out once a year for updates. Use this for highly technical communications (the nerds list).
Part 2: Real Work
Plugfest Results! Thanks again for Deluxe for use of room for two full days!
Deluxe Hollywood - GREAT VENUE!
TMS to Server testing on Monday
Participating: Sony, Hollywood Software, Dolby, GDC
(Missing: Arts Alliance, Unique, others)
Plugfest “Light”
Most worked - but mainly latest version of software, not all the equipment in field is up to this level.
Next plugfest should be a “traveling roadshow” to test field systems.
(Field is about 50% success - mainly due to non-updated software)
Discussion of what should be the guidelines for ingest behavior - probably not ingest EVERYTHING. To talk about later.
Results:
Most everything worked!
(One issue with ingest of Barco Escape content).
Everything followed the guideline of “ingest everything” even if you don’t recognize it.
HOWEVER ingesting of “bad” (corrupted) content and/or malformed (not meeting spec) content.
WHAT SHOULD TMS BEHAVIOR BE? (Basic more to discuss about validation and ingest behavior. )
More details:
Some systems could talk to many servers.
Others have a restricted set of servers that are supported.
Some API’s have changed so communication broke down.
Test Content playback on Tuesday (Wednesday cancelled)
Goal was to test deployed SMPTE-DCP compatibility - didn’t happen. Still testing latest-and-greatest software versions.
Tested Servers: Dolby, Barco, GDC, Qube
Test content:
3D Rendered Subtitles
Static Z - fixed per subtitle - is this useful?
Variable Z - changing Z per subtitle (Eikon)
Definition:
Variable Z = each frame is different in Z for a single subtitle
Static Z = each SUBTITLE is at different Z (all frames of that subtitle is the same)
30% of released titles use variable Z, the rest are static Z (all are burned in today)
This might be emulated by shifting Z every 1 second? This is being tested at EDCF (stair step approach). We will wait to hear how it goes…
Distributors would like variable Z - some in the room believe variable Z is REQUIRED (not sure if the stair step Z without flash would be sufficient).
PNG placement (JW)
A test of color and placement - this may become part of the compliance test plan.
HFR in 2D and 3D
Fox Format Test (encrypted, all frame rates from 24 to 120, including captions and subtitles, 250/500Mbps)
Pretty picture demo material (JC)
Bottom Line:
ALL servers played ALL frame rates that are within spec!
Up to 2D60-2K 3D60-2K - including 24, 25, 30, 48, 50, 60fps
Some could play up to 3D60-4K (with extraction of 2K layer) and 2D120-4K (with extraction of 2K layer)
We did learn that the macro to set up projectors needs to be set prior to shifting frame rates. We found visible problems with some at HFR, but we could trace it to a flash set up that would have been solved with a proper macro. This might be very important for instructions to projectionists that are setting up for Ang Lee’s upcoming movie. (It took us 2 hours to figure out what was going on.)
Pretty picture playback was fun to watch.
It was impressive the similarity between 3D120 and 3D60. We don’t think anyone could tell the difference.
• Variable3D was in B2.1 format with the MCA metadata (no true labeling) in WTF format, no issues with that at all anywhere.
• Still having some issues with PNG color.
• We have 3 formats that are the highest common FPS, which is playable on ALL of the systems tested. Anything past this is either downscaled or completely unsupported on some systems, even if one or two are ok with it.
• 2D 2K 60fps (tested @ 250 mbs)
• 2D 4K 30fps (tested @ 250 mbs)
• 3D 2K 60fps (tested @ 500 mbs)
Future Plugfests, we NEED to get ALL content a full week prior to make sure all content conforms to specs. [Except for those pieces purposely put in to trip up and test how systems handle bad content.]
Notes: PNG Placement was indeed packaged with bad PNG UUID calls (should be “urn:uuid:#####” and it was just “#####”). It didn’t fail it on ingest, one system failed on playback.
A review of Document 12 - Recommended Practice for Ingest Behavior. There are changes that need to be modified to identify which types of content should be ingested (or not). One suggestion it to not ingest malformed XML documents or hash failures.
Discussion of 4K checkerboard - There was a clever image pattern to determine if a 4K decode was actually 4K. Included in the package was a 4K checkerboard. We didn’t reach a conclusion on what it should look like in 4K it was clearly decoded as grey for a 2K and was a good check for decoder performance.
Some of the packages at the plugfest included invalid content. If you are interested in which ones are invalid, talk directly to Steve L.
Is there a need for content providers to include the projector macro to set the projector into a known (correct) initial state. This could help accelerate the testing procedures.
The next “Roadshow” plugfest should be scheduled in November - maybe 14-17 2016. Same content as this plugfest, but testing real world installations. There was a lively discussion on the value of a traveling plugfests. The goal is to sample a representable set of systems and just find out what is going on in the field.
Feedback from ISDCF’s submission to DCI
(We asked DCI to include SMPTE flavor B1/B2 as part of their CTP test package.)
DCI responded that their CTP has a different purpose and is forward looking and includes requirements beyond current (restricted) needs.
Some in ISDCF questioned some of the forward looking requirements and the need, for instance MCA routing.
One opinion was that the prior test material did not exercise a wide enough set of edge-cases and equipment was designed for the test and doesn’t meet all the (untested) requirements.
Another opinion was that there is a disconnect between studios, vendors and theater owners. There may be a need for a organization to give “rational” response to DCI and have the DCI directives to be more cooperative.
The open standard (SMPTE) is so wide that it is restricted in practice and DCI sets the restrictions. Many vendors asked to contribute to the DCI process.
One asked if the DCI test content could be made available with the same distribution freedom as ISDCF test materials. (There is an issue with StEM material restrictions.)
ISDCF did a negotiations between industries - vendors, studios, exhibitors - to develop a compromise (SMPTE Test Package B). It is a balance between needs.
In the long term future there would be an opportunity to upgrade to use new features.
Looking forward: you could make the shift to SMPTE-DCP by step-wise upgrades. One studio tried to do move to SMPTE-A first and found problems. They changed their approach to just take the leap to Test Package B.
Test Content
SMPTE is selling DPROV content, ISDCF has content, DCI has CTP content. What should be the coordination between test contents? We don’t have agreement on all parts and descriptions of what the test content achieves.
What feature set should be the minimum requirement for the short term future?
SMPTE-DCP update
Pete’s Dragon Disney DCDC August 12 - a letter is going to exhibitors.
2D & 3D (burned in Subs to international locations)
Paramount did issue “Everyone Want’s Some” in SMPTE-DCP
Other studios are working on preparing SMPTE-DCP for the future. No other announcements made.
What About ISDCF? It’s Been 10 Years - what should we be doing?
Should we keep meeting? Is it valuable?
Have transitioned from launch to maintenance phase? No, not until SMPTE-DCP is launched.
Comment: Isn’t it self evident? People still come to the meeting and call in. There must be value.
Comment: This forum is the only one to discuss SMPTE-DCP detail issues.
Comment: ISDCF helps the industry and a place to meet up is VERY valuable. We stay ahead of serious issues. We might transition to TMS’s and software updates. There is pain that needs to be addressed.
Comment: More problems will come up, this is the only forum to address these issues.
Comment: VERY valuable for those that live out of area to meet with industry folks.
Comment: It’s very hard to hear on the phone. (Especially if you can’t link a face to a voice.)
Are we meeting too often? Not enough?
Comments:
Schedule 6-8 weeks between meetings.
May be worth getting an “immersive audio system” to help make it easier for remote folks to follow. (Dean will try and test a system.)
Not having the buzz REALLY helps.
Event Cinema Automation - No follow up from last time. Need to get more requirements.
Alt Audio Tracks - Still need to get more information from studio.
Upcoming Meetings
Thursday July 21, 2016 Main meeting at Universal
Thursday September 22 Universal
* Traveling Plugfest - November 14-17 2016
Part 1: General Reporting
Housekeeping:
Antitrust Disclaimer – The official antitrust guidelines are posted on our website and are linked from the main ISDCF page. A short verbal overview of guidelines was given.
Thanks to Fox for lunch. Upcoming (note that the dates are messed up): Barco June ’16, August ’16 WB/Mike Z, Jim W/Wendy A (Consultants) Sept ‘16
Thank you to Universal for the facilities and parking. Universal will be the primary location for future ISDCF meetings - although Uni may loose the conference room
Thank you to Universal for support of the Chairman.
Thank you to Inter-Society for providing the funding for notes, travel, general expenses and admin support.
Thank you to Universal/Sony for the coffee and treats
Meeting notes from March ‘16 was reviewed and approved.
Legal reminder / press reminder
Chatham House Rule:
When a meeting, or part thereof, is held under the Chatham House Rule, participants are free to use the information received, but neither the identity nor the affiliation of the speaker(s), nor that of any other participant, may be revealed.
Attendance is at end of these notes.
Action Items from May 19, 2016
Start a thread on Forum on Event Cinema Automation needs / workflow solutions. (CJ and Jerry)
Eikon needs to get Jerry the variable Z test package for posting on ISDCF (the one that was used at the plugfest.
Distributors will ask their movie makers on the absolute need for continuous variable Z - is it good enough to just have a set of fixed Z subtitle placement - would that be acceptable for using this feature?
Get results from EDCF for ‘Stair step’ static Z positioning to emulate variable Z.
Subgroup to edit / recommend changes for Document 12 - RP for Ingest Behavior - Steve L to lead, Dean B, John H, Mike R, Jim W, Chris W, Bill E “volunteered” to participate.
Create an updated suite of well described test content on the ISDCF website (Jerry) Including SMPTE DPROVE content.
Try Dolby conference system - Dean will try to get a system for us to try.
From earlier meetings:
Wendy will check with John H and SMPTE to see if there is a list of SMPTE specs that could be posted on the ISDCF website.
Review the use of ContentTitleText in the PKL for situations where the PKL holds multiple CPLs. Language to improve the Digital Cinema Naming Convention site OR may become a new recommended practice document for ISDCF (Chris W)
Test content to show the difference between font rendering color and expectations. i.e. create a font color that should match the background and blend in - this would show the issue. Should be ready by plugfest in May.
Add to the SMPTE CPL metadata document the use of lower case subtitles indicating burned in subtitles and upper case is rendered. Dean [Post on Forum.]
Still looking for DCP that includes Atmos in test content B configuration (Dean) Should be ready by plugfest in May.
Still a need for a simplified description of FLMx for what it does and doesn’t do (to make sure it won’t be rejected due to lack of knowledge of what it does and doesn’t do.
ISDCF should have a recommended practice for zip/tar content to create drives (NOT for distribution to theaters.) Dean? Nick? Steve?
ISDCF should write description of submission aspect ratios to the Academy. (Ernie / Bill H. will take the lead.)
We had discussion on when the next ISDCF meetings will take place. We deleted the June meeting, we set a July 21 meeting date, we deleted the August 18 date and set the September 22 dates. We also set a traveling plugfest for November. We discussed the need to NOT have meetings closer than 4 weeks apart.
It was pointed out the meeting notes from 10 years ago (at end of last meeting notes).
Discussion of iPath Reflector - still needed? Should we kill this reflector or make it more known? Results: keep the iPath list. Send the current membership out once a year for updates. Use this for highly technical communications (the nerds list).
Part 2: Real Work
Plugfest Results! Thanks again for Deluxe for use of room for two full days!
Deluxe Hollywood - GREAT VENUE!
TMS to Server testing on Monday
Participating: Sony, Hollywood Software, Dolby, GDC
(Missing: Arts Alliance, Unique, others)
Plugfest “Light”
Most worked - but mainly latest version of software, not all the equipment in field is up to this level.
Next plugfest should be a “traveling roadshow” to test field systems.
(Field is about 50% success - mainly due to non-updated software)
Discussion of what should be the guidelines for ingest behavior - probably not ingest EVERYTHING. To talk about later.
Results:
Most everything worked!
(One issue with ingest of Barco Escape content).
Everything followed the guideline of “ingest everything” even if you don’t recognize it.
HOWEVER ingesting of “bad” (corrupted) content and/or malformed (not meeting spec) content.
WHAT SHOULD TMS BEHAVIOR BE? (Basic more to discuss about validation and ingest behavior. )
More details:
Some systems could talk to many servers.
Others have a restricted set of servers that are supported.
Some API’s have changed so communication broke down.
Test Content playback on Tuesday (Wednesday cancelled)
Goal was to test deployed SMPTE-DCP compatibility - didn’t happen. Still testing latest-and-greatest software versions.
Tested Servers: Dolby, Barco, GDC, Qube
Test content:
3D Rendered Subtitles
Static Z - fixed per subtitle - is this useful?
Variable Z - changing Z per subtitle (Eikon)
Definition:
Variable Z = each frame is different in Z for a single subtitle
Static Z = each SUBTITLE is at different Z (all frames of that subtitle is the same)
30% of released titles use variable Z, the rest are static Z (all are burned in today)
This might be emulated by shifting Z every 1 second? This is being tested at EDCF (stair step approach). We will wait to hear how it goes…
Distributors would like variable Z - some in the room believe variable Z is REQUIRED (not sure if the stair step Z without flash would be sufficient).
PNG placement (JW)
A test of color and placement - this may become part of the compliance test plan.
HFR in 2D and 3D
Fox Format Test (encrypted, all frame rates from 24 to 120, including captions and subtitles, 250/500Mbps)
Pretty picture demo material (JC)
Bottom Line:
ALL servers played ALL frame rates that are within spec!
Up to 2D60-2K 3D60-2K - including 24, 25, 30, 48, 50, 60fps
Some could play up to 3D60-4K (with extraction of 2K layer) and 2D120-4K (with extraction of 2K layer)
We did learn that the macro to set up projectors needs to be set prior to shifting frame rates. We found visible problems with some at HFR, but we could trace it to a flash set up that would have been solved with a proper macro. This might be very important for instructions to projectionists that are setting up for Ang Lee’s upcoming movie. (It took us 2 hours to figure out what was going on.)
Pretty picture playback was fun to watch.
It was impressive the similarity between 3D120 and 3D60. We don’t think anyone could tell the difference.
• Variable3D was in B2.1 format with the MCA metadata (no true labeling) in WTF format, no issues with that at all anywhere.
• Still having some issues with PNG color.
• We have 3 formats that are the highest common FPS, which is playable on ALL of the systems tested. Anything past this is either downscaled or completely unsupported on some systems, even if one or two are ok with it.
• 2D 2K 60fps (tested @ 250 mbs)
• 2D 4K 30fps (tested @ 250 mbs)
• 3D 2K 60fps (tested @ 500 mbs)
Future Plugfests, we NEED to get ALL content a full week prior to make sure all content conforms to specs. [Except for those pieces purposely put in to trip up and test how systems handle bad content.]
Notes: PNG Placement was indeed packaged with bad PNG UUID calls (should be “urn:uuid:#####” and it was just “#####”). It didn’t fail it on ingest, one system failed on playback.
A review of Document 12 - Recommended Practice for Ingest Behavior. There are changes that need to be modified to identify which types of content should be ingested (or not). One suggestion it to not ingest malformed XML documents or hash failures.
Discussion of 4K checkerboard - There was a clever image pattern to determine if a 4K decode was actually 4K. Included in the package was a 4K checkerboard. We didn’t reach a conclusion on what it should look like in 4K it was clearly decoded as grey for a 2K and was a good check for decoder performance.
Some of the packages at the plugfest included invalid content. If you are interested in which ones are invalid, talk directly to Steve L.
Is there a need for content providers to include the projector macro to set the projector into a known (correct) initial state. This could help accelerate the testing procedures.
The next “Roadshow” plugfest should be scheduled in November - maybe 14-17 2016. Same content as this plugfest, but testing real world installations. There was a lively discussion on the value of a traveling plugfests. The goal is to sample a representable set of systems and just find out what is going on in the field.
Feedback from ISDCF’s submission to DCI
(We asked DCI to include SMPTE flavor B1/B2 as part of their CTP test package.)
DCI responded that their CTP has a different purpose and is forward looking and includes requirements beyond current (restricted) needs.
Some in ISDCF questioned some of the forward looking requirements and the need, for instance MCA routing.
One opinion was that the prior test material did not exercise a wide enough set of edge-cases and equipment was designed for the test and doesn’t meet all the (untested) requirements.
Another opinion was that there is a disconnect between studios, vendors and theater owners. There may be a need for a organization to give “rational” response to DCI and have the DCI directives to be more cooperative.
The open standard (SMPTE) is so wide that it is restricted in practice and DCI sets the restrictions. Many vendors asked to contribute to the DCI process.
One asked if the DCI test content could be made available with the same distribution freedom as ISDCF test materials. (There is an issue with StEM material restrictions.)
ISDCF did a negotiations between industries - vendors, studios, exhibitors - to develop a compromise (SMPTE Test Package B). It is a balance between needs.
In the long term future there would be an opportunity to upgrade to use new features.
Looking forward: you could make the shift to SMPTE-DCP by step-wise upgrades. One studio tried to do move to SMPTE-A first and found problems. They changed their approach to just take the leap to Test Package B.
Test Content
SMPTE is selling DPROV content, ISDCF has content, DCI has CTP content. What should be the coordination between test contents? We don’t have agreement on all parts and descriptions of what the test content achieves.
What feature set should be the minimum requirement for the short term future?
SMPTE-DCP update
Pete’s Dragon Disney DCDC August 12 - a letter is going to exhibitors.
2D & 3D (burned in Subs to international locations)
Paramount did issue “Everyone Want’s Some” in SMPTE-DCP
Other studios are working on preparing SMPTE-DCP for the future. No other announcements made.
What About ISDCF? It’s Been 10 Years - what should we be doing?
Should we keep meeting? Is it valuable?
Have transitioned from launch to maintenance phase? No, not until SMPTE-DCP is launched.
Comment: Isn’t it self evident? People still come to the meeting and call in. There must be value.
Comment: This forum is the only one to discuss SMPTE-DCP detail issues.
Comment: ISDCF helps the industry and a place to meet up is VERY valuable. We stay ahead of serious issues. We might transition to TMS’s and software updates. There is pain that needs to be addressed.
Comment: More problems will come up, this is the only forum to address these issues.
Comment: VERY valuable for those that live out of area to meet with industry folks.
Comment: It’s very hard to hear on the phone. (Especially if you can’t link a face to a voice.)
Are we meeting too often? Not enough?
Comments:
Schedule 6-8 weeks between meetings.
May be worth getting an “immersive audio system” to help make it easier for remote folks to follow. (Dean will try and test a system.)
Not having the buzz REALLY helps.
Event Cinema Automation - No follow up from last time. Need to get more requirements.
Alt Audio Tracks - Still need to get more information from studio.