[Thomas Leong]

Suggest you google for Video Conferencing and/or Webinars and/or Webcasts.

Good luck,
Thomas Leong

[Thomas Leong]

“(For example, someone talking and walking on the Left projector, and then ‘walking’ to the Main, and then to the right projector)”

Complicated. Are you shooting with one camera or 3 cameras?

With one camera, next question is what is the native resolution of your projectors? If they are HD, you’ll probably need at least a 4K camera (Red?) to maintain some sort of resolution.

With 3 cameras, they will have to be aligned similar to the projection as well. If side-by-side projection with no overlap, each camera’s field of view edge would have to meet the adjacent camera’s field of view edge.

With a one camera shoot at min. 4K, next step is to take the footage into, say, After Effects, and output 3 files aligned to the alignment you intend to project – either side-by-side, or with overlapping areas, taking into account the amount of the overlap in pixels.

Then you need a playback software that can synchronize playback of the 3 files, and have a machine or multiple machines that can handle the 3 files at those resolutions. The more pixels the more demanding from the machine(s) playing back without hiccups. Some options here are –
Apple: RenewedVision Playback Pro2
Windows: Dataton Watchout; AvStumpfl Wings Platinum

Sounds complicated, and expensive? It is.

Thomas Leong

[Thomas Leong]

Try this freeware from –
https://www.zachpoff.com/software/multiscreener/

Thomas Leong

[Thomas Leong]

I think the simpler method is to make 2 comps – one for the 3 walls at 5760×1200, the other for the floor as a standalone. For Comp 1’s output, use the Render Module crop feature to output to the 3 files – left, centre, right. The floor (standalone) is not linked/joined in imagery to Comp 1.

A more involved method may be to comp the lot as a 3D Cube (do a search – the Cow has tutorials on this) or for Dome projection but only output the relevant parts without distorting for a dome, maybe something like this methodology may be helpful reading –
https://www.lochnessproductions.com/reference/2014IMERSA_ae/2014IMERSA_ae.html

good luck!
Thomas Leong

[Thomas Leong]

Apart from being able to find a voice in a particular language, one main problem in foreign language voice-overs is a professional read. Fluency and intonations, etc play a major role in delivering a convincing VO.

The other major problem is the language itself. A short English sentence may require several words in a particular foreign language to convey the same meaning. This creates time/duration problems in re-syncing to your English cuts/footage.

Years ago, I had to do a Japanese VO sync and client had a local ‘friend of a friend’ who had volunteered to do the VO for free. We used a pro studio and recorded the voice. Fine. I then hired someone from the local Japanese Language Institute to sit with me and help me with the re-sync session. Went well. But when finished work was presented to some ‘test subjects’ who knew the language, the main comment was “The voice sounded amateurish”. And this was a presentation meant to sell tourism promotion to Japanese tour agencies. In the end, client heeded the advice of the ‘test subjects’ and I had to contact a Japanese colleague in Japan to re-record the script with a professional voice-over in Japan. Cost was obviously higher, but it did sound much better.

So be aware that just because it sounds, say, Greek, does not mean it sounds convincing for the intended audience in that particular language. Then again, it may be alright for a non-formal presentation like a software tutorial, but if the software costs $thousands, a professional read is what I would recommend.

Thomas Leong

[Thomas Leong]

Part 1
If your problems are similar to this thread –
https://forums.creativecow.net/readpost/158/857738

suggest you try his solution as posted in the ControlBooth forums as indicated.

Part 2
Not being familiar with Screen Pro at all, I have no answer.

Thomas Leong

[Thomas Leong]

Not sure, but maybe Vasco da Gama might do it –

https://www.motionstudios.de/product_info.php?info=p109_Vasco-da-Gama-8-HD-Professional.html

Thomas Leong

[Thomas Leong]

Apart from what the others have advised, I have one question: What is the duration of each comp/resultant video? Reducing the duration and using the playback software to loop the animations will help reduce your render times.

My own experience:
A few months ago, I had a similar assignment – animate 11 psd layered files to be used as backgrounds for a ballroom function. Each psd file was 20,250 pixels width x 4800 pixels height, sizes ranging from 150 – 350MB. Humongous!! My comp for each was 4096 x 1080 (for a panoramic screen covered by 3 x 20K lumens projectors). I had 1 week to complete the job.

Given the above, I made my comp durations at 4 secs each, and used the playback software loop the animations (Dataton Watchout software), carefully keyframing the animated parts so that jerks were not obvious at the loop points (end and start of each animation). The shortest comp I had was 2 secs…and that took 2 hours to render. Surprisngly, the 4-sec comps rendered in about 1 minute – yep, ONE minute for a 4096×1080 H264 file from 20,250 x 4800 psd files, straight out of After Effects, no intermediate codecs. Initially, I made the mistake of reducing the psd resolutions to 4096 x 1080 BEFORE importing into the AE comp. The resultant rendered videos were horrible – jagged edges, obviously grainy, etc. So I had to use the original psd files at their original sizes in order to get quality output. CTRL-ALT-F was used to fit the psd to the comp, and that maintained the quality in the final renders.

The Trapcode plug-ins I have render fast, so I used mainly Shine and Starglow effects with AE’s text position and scale effects to minimize render times. I’m on After Effects 7 (never upgraded since I hardly do this work any more, and my Trapcode effects were bought during those early release days). My render PC, however, was new – slightly overclocked i7 4770K to 4GHz with closed loop water cooling with a 240mm radiator, 16K RAM (now 32K), Windows 7 Pro 64-bit, OS and applications on an Intel SSD, two disk RAID 0 with 7200rpm hdd for data and AE cache, ASUS HD7970 Matrix Platinum graphics card, etc.

Thomas Leong

[Thomas Leong]

He found the answer himself by reading the manual –
https://www.controlbooth.com/threads/screen-pro-ii-resolutions.35551/

To quote him –
“…the SPII doesn’t support 1:1 pixel sampling when the pixel clock is over 150MHz…”
“On the windows PCs we were able to achieve the 1:1 sampling by altering the timings in the advanced driver settings (many default to a pixel clock around 170MHz for 1920x1080p @60)”

Members

[Thomas Leong]

Tough questions to answer and it would seem that no one here has the experience (self included).

For a distortion free image, it would seem that your objective should be a rectilinear image (or video). Examples are those projected at this multi-sensory Shanghai dining experience, called Ultraviolet –
https://uvbypp.cc/
or
https://www.finedininglovers.com/blog/out-of-the-blue/ultraviolet-paul-pairet-shanghai/

Rough calculations indicate that if each projector for each wall were 1024×768 pixels (XGA resolutions), you would need at least a 4K image or video. Google for ‘4K footage’ and you’ll get some hits for free footage. Generally, one feeds the projector at its native resolution, i.e. use XGA resolution images/video for XGA projectors. If projectors are at HD resolutions (1920×1080 minimum), you would need 6K-8K images/video. Anything less would result in soft/blurred images.

To render out what-we-call ‘pre-split videos’ from one large source, I’d recommend After Effects (Render Queue) which allows you to crop the large source into multiple files.

For playback, you may want to consider Dataton’s Watchout (easy to learn), or AvStumpfl’s Wings Platinum (both are Windows-based programs). There are Mac options but I’m not familiar with them.

To shoot your own, I think the 360Hero.com H3PRO7HD or H3PRO6N (discarding the upper camera) holders will cover the surfaces required with little distortion since the cameras are vertically mounted.

Any distortion can probably be corrected with either Autopano Video or Video-Stitch if they can be set to render a rectilinear video, then split with After Effects for multiple pre-split files.

Or modify the GoPro Hero3 cameras with distortion free lenses:

Modding GoPro for Interchangeable lenses –
https://nofilmschool.com/2013/12/back-bone-ribcage-gopro-hero3-plus-interchangeable-lenses/

Ragecam lens –
https://www.ebay.com/itm/181166584421?ssPageName=STRK%3AMEWAX%3AIT&_trksid=p3984.m1423.l2649

Thomas Leong