Film-Tech Forum ARCHIVE: Open Caption placement

This is topic Open Caption placement in forum Digital Cinema Forum at Film-Tech Forum ARCHIVE.

To visit this topic, use this URL:
https://ft-forum.com/ft/cgi-bin/ubb/ultimatebb.cgi?ubb=get_topic;f=16;t=002861

Posted by Kevin Fairchild (Member # 4985) on 11-08-2016, 06:30 PM:

Can open caption placement be adjusted within the projector or from Dolby DSS servers? Adjust horizontal/vertical placement or text size?

Thanks

Posted by Harold Hallikainen (Member # 5405) on 11-08-2016, 08:48 PM:

I suspect not. The timed text file defines font size and text position.

Harold

Posted by Carsten Kurz (Member # 5396) on 11-08-2016, 09:20 PM:

Position only, on Dolby 'Doremi' Servers, not the DSS line.

Always wondered why this is not configurable on projectors or servers. Not really complicated to do.

- Carsten

Posted by Harold Hallikainen (Member # 5405) on 11-08-2016, 09:46 PM:

I think that those who produce content want it to be displayed the way they produced it. At ISDCF meetings, there's been talk of turning features on and off at playback time. The studios have said "create another CPL." You're supposed to play what's in the CPL as it's written.

I think that's the intent, anyway.

Harold

Posted by Carsten Kurz (Member # 5396) on 11-09-2016, 01:47 AM:

Clear to me - the trouble is, we have seen more than enough occasions where the placement is wrong. It has to be adjusted to screen/masking. I agree that no other changes than position should be possible, although in certain old-style auditoriums also size adjustments could be useful. It should an installer option in general, though, not easily accessible for the average SMS user.

- Carsten

Posted by Randy Stankey (Member # 64) on 11-09-2016, 03:32 PM:

Ironically, I was reading up about captions, just the other day.

Yes, the position of the text is dictated by the party that creates them and that information is encoded into the stream. However, after that, it gets complicated.

As I understand, there are two main formats for captioning. The old, analog system ("Line 21") and the newer system(s) for HDTV and/or movies. The two are not compatible but, still, companies are obliged to produce captions that are compatible with both systems. In order to do that, most companies just create captions for the old, Line 21 system then transcode them into the newer formats.

To make matters worse, the originating device is the one that is supposed to be responsible for displaying captions on the screen but that device may or may not correctly pass the caption data down the line to other devices/displays/projectors.

In other words, if you have a DVD that contains captions, it is the DVD player that is supposed to be responsible for printing the on-screen but, unless you activate captions on the DVD player, you won't see them, even if your TV has the ability to display captions and you turn that feature on from the TV. The DVD player might or might not pass the captioning data along to the TV.

It all depends on how the captions are encoded, whether the devices are capable of displaying them and whether a given device in the chain actually passes caption data to downstream devices.

That's the reason I looked it up, in the first place. I wondered why captions weren't playing even when I knew that the movie was supposed to be captioned.

I can understand why captions in movie theaters might come out messed up or not play at all.

It sounds, to me, like closed caption formats are just as messed up as video and sound formats are. [Frown]

Posted by Kevin Fairchild (Member # 4985) on 11-09-2016, 03:39 PM:

Thanks everyone. What about viewing closed captions from a web browser in sync with the feature?

Posted by Carsten Kurz (Member # 5396) on 11-09-2016, 04:28 PM:

The caption formats in DCI are actually quite clear and streamlined, even as currently there are two standards for packaging/wrapping: Interop or SMPTE.

The trouble in DCI open captions is not the format, but the equipment rendering them in all sorts of unreliable ways.

It has gotten better recently, though, with server and projector updates.

For a while it was like 'Oh, yes, this projector and that server sure supports it, as we have to, but we don't care wether it's implemented properly, because no one really cares if it doesn't work.'

- Carsten

Posted by Harold Hallikainen (Member # 5405) on 11-09-2016, 08:44 PM:

On viewing captions on a web browser, the USL IRC-28C makes captions available on a web browser for system debug. Because it's on the auditorium network, this web interface is not available to the public.

On caption (and subtitle) file formats, there is CineCanvas used by Interop and SMPTE timed text used in SMPTE DCPs. The main difference is that fractions of a second are based on 4ms ticks in Interop, while in SMPTE, the whole time is a count of editable units. HH:MM:SS:EE where EE is edit units (frames for 2D). IF the edit rate is an integer, HH:MM:SS corresponds to "wall clock time." If the edit rate is not an integer, HH:MM:SS does not correspond to "wall clock time." I wrote a formula that's in the standard that converts HH:MM:EE to edit units from the beginning of the reel.

Harold

Posted by Frank Angel (Member # 248) on 11-10-2016, 05:49 AM:

A friend of mine worked for a company that laser cut subtitles into finished 35mm prints for distributors. One of the shift operators was green and basically didn't know what he was doing and worse, didn't care. He cut an entire 5 reel feature with the subtitles at the very top of the image instead of the bottom. He was fired of course and the client wanted the full restitution for the price of striking a new print, claiming that the position of the subtitles rendered the print unrentable; they said they would sue for the print costs and lost rental damages.

Thing is, before they went to court, the titling company found out that this distributor had booked the very same print anyway and it played in two theatres. Not only that, the distrib pitched it as a new captioning process that was easier to read, especially in non-stadium seating cinemas since the captions were never obscured by patron's heads -- talk about turning lemons into marketing lemonade hype. Once that cat was out of the bag, they settled out of court. Seems supertitles could be seen as an added value. Opera houses run supertitles all the time, and yah, that is quite a different animal because the supertitles are above the "image," but sort of the same concept. Do we stick with subtitles at the bottom of the image only because it been that way like forever? What would be the disadvantage of supertitles in film?

I have run lots (and lots) of subtitled foreign language films and I always thought it would be a smashing idea to put subtitles in the hardmat just below the 1.85 image. The art houses are where subtitled film are booked and they almost always have lenses and plates to be able to project 1.66; they could easily use that setup to show the subtitles in black below the picture -- never a problem with light backgrounds making the text hard to read. And even better, theatres in the country of origin could simply project at 1.85 and the subtitles would be gone, image remaining pristine. But hey, no one asked me.

Posted by Carsten Kurz (Member # 5396) on 11-10-2016, 07:44 AM:

I guess it's less an issue with overhead subtitles at the opera because it is REALLY over head. In Movies, you would have the titles appearing on peoples heads, whereas you can assume that they don't block essential image content at the bottom.

That said, every studio is free to place their titles at the top, both Cinecanvas and SMPTE allow this. They just don't do it for common subtitling applications. In special cases, they place subtitles around the screen.

And yes, one of the cases where I would also like to place the subtitles at my own disposal would be below the image. That, however, with current projection technology, is only possible for Scope Features, as there is no imager estate left on projectors configured for flat. You would need a separate projector or display device there. I would always prefer to have subtitles in an extended masking area.

The way it is now, they always have to display above the configured active pixel area of the feature/preset. With some misconfiguration, you can force them to show below, although I am not sure they are rendered there properly on all systems. As the rendering is done by the ICP, I would assume it depends more on the ICP software version than projector type.

- Carsten

Posted by Randy Stankey (Member # 64) on 11-10-2016, 10:49 AM:

quote: Carsten Kurz
I guess it's less an issue with overhead subtitles at the opera because it is REALLY over head. In Movies, you would have the titles appearing on peoples heads, whereas you can assume that they don't block essential image content at the bottom.

Yes, I have done supertitles for operas on stage.
We would hang a Hollywood flat in front of the main valance, painted gray with black gaff tape for a border. We used a video projector and PowerPoint to project the words, white over black. It looked really good when it was all put together.

Point being that the titles were all sectioned off in their own virtual box, separate and above the stage so that people who wanted to read them could easily do so but people who understood the language (or didn't care) could also watch the action on stage without being distracted.

As much as I like subtitles in movies or TV shows, I hate the distraction they cause. If there was a way to present captions as supertitles or, like others say, in the matted off areas, I would be all for it.

quote: Carsten Kurz
In special cases, they place subtitles around the screen.

I have seen some subtitles/captions that pop up around the screen almost like speech balloons in comics. If done well, I would be all for that. Just so long as the screen doesn't become so busy that it looks like those Pop-Up Videos that they had in the '90s. I hated those things!

I think we need to work out a good system for captions that everybody can use in theaters. Captions on the screen are distracting but, as things are, it's the only practical way besides Rear Window systems.

Personally, I like Rear Window captions but, at the same time, expense and complexity are the great stumbling blocks.

Posted by Kevin Fairchild (Member # 4985) on 11-10-2016, 01:06 PM:

quote: Carsten Kurz
And yes, one of the cases where I would also like to place the subtitles at my own disposal would be below the image. That, however, with current projection technology, is only possible for Scope Features, as there is no imager estate left on projectors configured for flat. You would need a separate projector or display device there. I would always prefer to have subtitles in an extended masking area.

The way it is now, they always have to display above the configured active pixel area of the feature/preset. With some misconfiguration, you can force them to show below, although I am not sure they are rendered there properly on all systems. As the rendering is done by the ICP, I would assume it depends more on the ICP software version than projector type.

That is exactly what I'm trying to do, place captions on a separate screen in the lower masking area for scope. If you have any ideas on how to accomplish, I'd love to hear them. Using NEC3240 with DSS220/CAT745

Posted by Carsten Kurz (Member # 5396) on 11-12-2016, 08:17 AM:

There were some discussions here about alternative titling options. I remember one guy being successful setting up his own webserver, but that was to supply his own titles to a DLP projector I think. Actually I don't know if the subtitle transfer between server and projector is encrypted now. If not, there would a chance to tap the connection at play time and create your own renderer on a PC connected to your own display. As long as Interop or unencrypted SMPTE subtitles are used, you could also grab the files from the DCP and use them on a separate machine for playout. You would need a sync mechanism to the main feature though to play them with the proper timing. So, API calls, LTC, or some sort of GPIO/serial/TCP triggers, depending on your servers.

Some Barco projector manuals still contain some documentation about the Cinecanvas subtitle mechanism and how to set it up externally for test or debugging.

It occurs misplaced in these manuals, as that passage doesn't deal with anything common for DCI operations. It gives you some idea, however, what goes on technically and how you may tap into it for special applications.

- Carsten

Posted by Gordon McLeod (Member # 33) on 11-12-2016, 08:30 AM:

cine digitals tms software can send subtitles to your tablet or cell phone using there app so one could i guess send the video out of the tablet to a seperate data projecotr and place it where one wants

Posted by Mark Gulbrandsen (Member # 72) on 11-12-2016, 07:48 PM:

Does it make the data available for all screens in a given complex or does it just do one screen at a time?

Posted by Gordon McLeod (Member # 33) on 11-13-2016, 11:02 AM:

sents the captioning to the ip address of a wifi emitor in the auditorium you want so if you put a seperate one in each room it would do them all

Posted by Carsten Kurz (Member # 5396) on 11-14-2016, 06:25 AM:

Would be interesting to know where they capture that data? For mainstream encrypted DCPs, the subtitle MXF would also be encrypted and they couldn't read it of the DCP files. So they would need to tap it from the projector-server communication at playtime? Then how do they do this for all the systems that do not allow this, e.g. IMS systems?

- Carsten

Posted by Harold Hallikainen (Member # 5405) on 11-14-2016, 08:20 AM:

If the system connects to the server on port 4170, the server uses the Auxiliary Content Synchronization Protocol to send information about the captions. It send the URL of the Auxiliary Resource Presentation List. This contains (among other things), the URL for each reel of captions. The captions are transported to the theater encrypted, but are decrypted before the external device fetches the captions (either on ingest or show selection). The content synchronization protocol then tells the external system where the server is in the timeline and whether the content is playing or not. This is typically done by a device in each auditorium, but if a central device has access to all the servers, it could also do it. Putting Wi Fi access points on the auditorium network would seem to raise security concerns. If an external server (like a TMS) could drive a separate network just for captions, that seems a bit safer.

Harold

Posted by Carsten Kurz (Member # 5396) on 11-14-2016, 10:21 AM:

Hi Harold - that would apply for closed caption devices. But I think the OP meant open captions, traditional subtitles. There is a connection between the renderer (projector) and server for classic HD-SDI/IMB systems with an external server providing the open captions to the projector at play time.

But not for IMS systems? Only closed caption system would need to have that external connection to closed caption display devices?

- Carsten

Posted by Mark Gulbrandsen (Member # 72) on 11-14-2016, 12:15 PM:

quote: Gordon McLeod
sents the captioning to the ip address of a wifi emitor in the auditorium you want so if you put a seperate one in each room it would do them all

Gord,

Well, I was thinking about your idea of a video projector in each auditorium projecting below the screen on a smaller one. I just don't see how that could work in a multiplex unless one does ethernet to video translation. At that point though with aux projector and lamp costs one might as well just use the emitter panel.

Mark

Posted by Carsten Kurz (Member # 5396) on 11-14-2016, 01:14 PM:

The idea is to have individual tablets to connect to a specific stream or Wifi to receive it. Of course, if the smartphone or tablet supports HDMI/MHL, you could as well connect it to a projector.

For the given reasons, I just don't think it is a universal solution. It would probably work more universal (and that is CDMs main intention) for closed captions.

- Carsten