Results 1 to 9 of 9

Thread: Time Slice trial

  1. #1
    Join Date
    Aug 2008
    Location
    South Devon
    Posts
    349

    Default Time Slice trial

    Just wanted to share.

    I got a lot of things wrong on this first trial of my time slice rig in it's current form.

    Things I can see are:

    The mount is linear through being straight, not the arc I really want.
    The camera's white balance settings were not balanced as they should be.
    I'm using a couple of E-20's amongst E-10's, they have to be replaced.
    This is only three quarters of the system, still awaiting camera deliveries.
    I have to match focal lengths between the video and the still cameras.
    I didn't close the eyefinders on the still cams, so they are pasty images.
    Focus the damn video cameras properly!!

    All I've done here is import everything, cut the still cam image durations to 2 frames, re-aligned them a bit, and clagged it into the relevant video camera bits, I've not done any retouching or morphing, so it's pretty rough, and short! .


    Last edited by Jerry Hill; 10-20-2008 at 08:39 PM. Reason: Replaced link to wmv with Vimeo version, eight hours to appear since upload!!

  2. #2
    Join Date
    Feb 2006
    Location
    Surrey, UK
    Posts
    10,842
    Blog Entries
    1

    Default

    Impressive stuff, Jerry. It's clearly good enough to make it worthwhile pursuing the project. It just staggers me that anyone has the time and ingenuity to do these things.
    Tim

  3. #3
    Join Date
    Aug 2008
    Location
    South Devon
    Posts
    349

    Default

    Thanks Tim.

    Yes there's some indicators there, the main thing is that doing this one really highlighted areas to focus on to make it as it it should be, or could be at least.

    As for the doing it, as a hobbyist through and through the act of getting this stuff together is part of the fun really. Many film makers have said I should be using the funds to buy decent video camera, but then the same people say that a good film maker will get a result with whatever gear is to hand, which is a bit contrary. I just like to add arrows to my quiver, they just might not be straight ones sometimes .

    Here's the rig as it is, still down on four cams, on the way though.


  4. #4
    Join Date
    Feb 2004
    Location
    Bracknell, Berkshire, UK
    Posts
    5,191

    Default

    The third camera from teh bottom is focused at the same distance as the others! Obvious I thought.

  5. #5
    Join Date
    Aug 2008
    Location
    South Devon
    Posts
    349

    Default

    Alan,

    Sorry, not with you, could you elaborate?

  6. #6

    Default

    Impressive, and nice looking rig, where did the term time slice come from, is bullet time a registered name these days..??

    I did know a Oz guy who made a 25 camera rig but he had lots of problems with the cameras going into sleep mode, and they were 35mm cameras, so who knows what was shot, nice thing was they could use the same film stock as the cine film cameras, made matching it up for the people in post easier.

    These days digital rules, lets drop that 100 year technology, its had a great run, give a thought what technology today will be around in 100 years..

    The Mad Bomber.
    I know nothing about youtube, even less about video
    http://www.youtube.com/user/leokimvideo

  7. #7
    Join Date
    Aug 2008
    Location
    South Devon
    Posts
    349

    Default

    The name for the effect is a good question, and my use of 'Time Slice' is based simply on the guy that invented the concept of multiple cameras shooting simultaneously at a moving subject.

    The Matrix brought the concept into public view, and it is put across as being the team that originated the concept. But from the research I've done, purely over the net, it was actually first conceived by a chap called Tim Macmillan in 1980, an Englishman. It consisted of a camera purpose made which was a 360 degree vertical ring of lenses with hand made photographic emulsions behind them. Subjects were passed through the ring and the shot taken at the appropriate time. The resultant images were then strung together to create a moving image of a time frozen subject.

    He then went on to use cameras based on 16mm movie filmstock, and so on till he basically perfected the system working with many more camera's, in the 125 range.

    Now, my understanding its that for the Matrix they took this method and developed it into what we are all familiar with, and they chose to call it 'Bullet Time', which seems to have been derived from the effect being employed primarily with the heavy use of computers, and CG bullets heading through space at the time the cameras were triggered. But, significantly they added the dimension of triggering the cameras in sequence, so the time was not so much as a 'slice' but more of a segment played back in slomo.

    So we have the common misunderstanding that the bullet is the main factor in 'Bullet Time', when in fact it's the time slice element that's the core of the effect.

    The Matrix team also engaged computers in other ways. Time slice to this day via Tim Macmillan's business, 'Time-Slice Films Ltd' (now a registered trade mark), and similar companies like 'Digital Air' use real environments, whereas 'Bullet Time' uses computers to control significant elements to the effect.

    By using virtual 3D modelling they generate CG backgrounds for the piece, and are able to use the data from that to set up the still cameras, so that the turn rates and POV can be matched. They also use the comps to trigger the cameras in sequence so that the rotation rate of the CG is perfectly matched to the live action which is shot in green screen setups. They also use comps to smooth the circular rotation of the resulting stills to remove any slight tolerance differences between the still cameras. To be honest, poor camera setting aside, what they get to start with is not too different to what I have here. Get the exposures, focussing, white balance etc right and the cameras are remarkably consistent, but the alignment of them remains a potential for problems. I use a 100mm ball as a target for setting up mine, they probably did something similar, but there is still need to re-align in post. I used 8 cams here, 12 is the end target of the first rig, The Matrix team used 120 I think. I'll keep adding them up to 24, then I'm hoping to jump to 50 but only if a friend of mine is able to design a sequencial trigger system. I've done 'better' time slice tests than this, but not on the target subject, this is the first time out with that.

    So that's why I call it Time Slice, I guess it's in honour of the chap who really came up with it, and should I ever bother to get into a method closer to Bullet Time then I shall call it that.

    As far as this rig is concerned one of the biggest challenges is getting that smooth turn, and that will only be done through having loads more cameras as close together as possible, unless I resort to Morphing.

    Morphing gives smooth results, But in this example, just taken by asking my daughter to sit still whilst I reposition a single camera, you can get an idea of the challenges morphing gives:

    Morph trial

    Basically the morphing has been applied to the girl only, the background was left to do it's own thing, I've found that the large movement and the problem with hidden parts of the scene creates a big issue, hence why green screen is actually much easier, but I can't set up a green screen in the middle of a flying site in 15mph winds, so I have to suss this out. Working on that along with simply adding those cameras as time goes by.

    The mistakes I made in the vid at the top are silly and curable, I must also add to that the simple thing of checking and re-checking the cams before each take, as things like the focussing rings on my choice of camera are electronic, not mechanical, and just brushing the lens barrel can effect focus change, even the mechanical zoom is a very light operation, and the reason why I've zoomed in to the effect segement in this times was not for effect, it was becuase one of the cams was set wrong and I had to re-align in post. In addition I thought I'd get away with using later versions of the Olympus E-10's, the E-20. But ironically these later once release a microsecond slower than the earlier ones, so I think it's cam 3 and cam 6 that are a little out of step to the rest, you can see that most easliy on the movement of the wheel in the segment. I'll keep them in there for now though, but will put them at either end, another E-10 will be with me this weekend, and all being well another next week. once I've got to twelve overall then I'll replace the e-20's. I have a third E-20 used to take the photo of the rig, but that's got a fault which stops it being included at the moment (fungus). It's possible that morphing will negate the problem really at my level.

    Btw one of the many lessons I learnt was to keep records of the takes, by taking a shot with the rig of a marker board I will be able to match things up easier, plus I'm hoping to be able to trigger the still function of the video cameras at the same time, so that I can mark the microsecond when the still cams are triggered, though for now I'm working on a simple audio 'click' sound to see how synchronising with that might work.

    Anyway, glad you liked it, it's been great fun so far and quite challenging, the making of the control unit was a biggy for me as soon as some basic electronics was involved, as it's pretty much new ground for me, but I could repeat it now no problem, so getting there.

    More cameras, more care, better procedures, and I'll have something I can take on location with some confidence. Then all I need to do is ensure the couple of seconds of effect it gives me truly enhances a video I might make. I'm under no illusion that this single thing will be anything more than a possible enhancement.

    By the way, these cameras are powered by a seperate power supply, this is buffered by the on board batteries, so I can leave the cams fully powered up for about three hours before they might shut down, sleep mode can be turned off, but when an external power supply is used that's bypassed anyway, so no problem with that on this rig. The camera type was chosen very carefully, there's a lot of criterion that needs to be satisfied.
    Last edited by Jerry Hill; 10-22-2008 at 12:35 AM.

  8. #8

    Default

    Wow, Jerry thats quite a reply, and I will give you some more of what I have seen over the years, especially some of the tests done on the first Matrix film which were never fully developed.

    Of course these days these types of effects would be done all CG, as that gives you total image control, I would say the roots of this type of effect go right back to a time before motion film cameras, where multiple still shooting cameras were used to try to capture movement, and a very early form of time compression.

    So using multiple cameras can be used for frozen and moving moments, the Matrix were thinking of a super slow motion camera on a very fast moving track, (remember the Matrix camera moves through a roller coaster move, not a straight arc), the issue with this are speed, you could not achieve the acceleration / deceleration required to achieve the shot over the required distance, it was a attempt to keep the effect in (one) camera, history tells multiple still cameras were used on the film which was in production now 10 years ago 1998.

    The next two film did not use this method at all to achieve the same type of effect.

    Yes focus, and framing is difficult, and I have seen a few different ways of getting a line up on the point of interest, I have seen lazer pointers attached to rings which clip onto lenses, its not bad for getting a general line up, the final is done via the viewfinder.

    The rig I saw built up (based on 35mm film) went out on three jobs, around 8 years ago, and that was it, some other companies built up digital camera rigs but the work here in Oz is very limited, and dominated by mainly one person.

    The main guy down my way who does this work is link below.

    Time Splice - Home

    The system was designed by Mark Ruff, a leading Melbourne‐based commercial photographer who has produced 3D digital content for local and international television commercials for over 10 years.

    He has been doing it for some time now, has had lots of media exposure, interesting also called Time Splice, (funny he was not known as Time Splice a few years back) I notice he has dropped his links to Matrix as I understand he had nothing to do with those camera rigs, always nice to drop that film name as its so well known for this type of effect.

    Good luck with building up your rig, and thanks for the morph video, very interesting.

    As for me I just blow things up, thats why I am the Mad Bomber.
    I know nothing about youtube, even less about video
    http://www.youtube.com/user/leokimvideo

  9. #9
    Join Date
    Aug 2008
    Location
    South Devon
    Posts
    349

    Default

    Quote Originally Posted by rollingstock View Post

    Yes focus, and framing is difficult, and I have seen a few different ways of getting a line up on the point of interest, I have seen lazer pointers attached
    I have a solution to that, which I think is what the Bullet Time team used. Effectly micro adjusters on the camera mounts.

    The problem is that on a 'beam' type arrangement like this, it's very easy for one adjustment to have an impact on another camera, it's just not ridgid enough, and there's a detectable variation in what's seen through one viewfinder to another, even though they are effectively on the same plane. It's not huge, but enough. So I'll be using a simple mount under each cam that allows for keeping the horizon level, and panning and tilting via simple threaded adjusters, not too dissimilar to devices used for aligning still cams for macro photography. Regular pan/tilt heads are too crude for this unless big buck geared ones are used.

    First though far more groundwork to do, one step at a time.

Similar Threads

  1. Time Slice rig trigger challenge
    By Jerry Hill in forum Cinematography
    Replies: 9
    Last Post: 09-30-2008, 08:58 PM
  2. Image manipulation for time slice (Bullet Time0 - Premiere CS3
    By Jerry Hill in forum Adobe Premiere, Premiere Elements, and After Effects
    Replies: 2
    Last Post: 09-03-2008, 10:40 PM
  3. prem cs3 trial - hd?
    By innocent_bystander in forum Adobe Premiere, Premiere Elements, and After Effects
    Replies: 1
    Last Post: 11-01-2007, 01:55 AM
  4. How to slice in detail the time line is too small
    By pielcanela in forum Sony Vegas video editing apps
    Replies: 4
    Last Post: 10-04-2007, 10:55 PM
  5. where can i get a trial
    By PMAC in forum General video editing software help and advice
    Replies: 2
    Last Post: 04-15-2004, 03:30 PM

Bookmarks

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •