It is currently Thu Aug 06, 2020 8:13 pm




Post new topic Reply to topic  [ 30 posts ] 
 Mantis real time - time-lapse tracking experiment 
Author Message
User avatar

Joined: Tue Nov 24, 2009 7:55 am
Posts: 951
Location: UK
Post Mantis real time - time-lapse tracking experiment
Real time verse Time-lapse mode using Mantis2 - 500 frame test. Check out the bugs at 00:20.. :D

A very quick test using Dolly, Pan, Roll and Focus axes. The movement is deliberately erractic to cram as much movement into a short time. The lens is a Canon 100mm macro lens.
The video move was recorded (500 frames -20seconds at 25FPS) then the Mantis software was switched to Time-lapse mode and the video move was repeated for the 500 frames at 1 frame per 2 seconds.
It was interesting to see how much wildlife comes into view on the time-lapse portion of the clip.
So if you have a wood burner and stack your wood indoors........... think again....... ;)

Some of the bumps in the T-L probably due to me walking on the wooden floorboards that the rig rests on.

No colour grading all straight out of the camera ("flat" jpeg stills)


_________________
Smile, it gives your face something to do!

http://www.vimeo.com/user1214873/videos
http://www.youtube.com/user/MrReggub


Wed Feb 12, 2014 11:59 am
Profile

Joined: Wed Feb 15, 2012 3:36 pm
Posts: 408
Post Re: Mantis real time - time-lapse tracking experiment
Hi Mike,

Glad that you are having fun with Mantis. I've had mine for just over two years now and it's good to see how it has been improving with nice tweaks and additions.

Edward


Wed Feb 12, 2014 1:29 pm
Profile
User avatar

Joined: Mon Nov 22, 2010 1:18 pm
Posts: 1294
Post Re: Mantis real time - time-lapse tracking experiment
Hi Mike good to see timelapse and real video can be shot on the same rig . Did you try overlaying the two shots would be interesting to see what the registry is like.
Very nice i am quite envious of all this live move stuff .
D1


Wed Feb 12, 2014 1:58 pm
Profile
User avatar

Joined: Tue Nov 24, 2009 7:55 am
Posts: 951
Location: UK
Post Re: Mantis real time - time-lapse tracking experiment
edward wrote:
Hi Mike,

Glad that you are having fun with Mantis. I've had mine for just over two years now and it's good to see how it has been improving with nice tweaks and additions.

Edward


Hi, Edward, another useful tool especially being able to create moves and view the results quickly before commiting them to time lapse.

DISPLACEMENT 1 wrote:
Hi Mike good to see timelapse and real video can be shot on the same rig . Did you try overlaying the two shots would be interesting to see what the registry is like.
Very nice i am quite envious of all this live move stuff .
D1


Hi, D1, just updated the video clip to include the overlay for you, it works quite well considering there is no sync between the moves, I guessed the video move start from the audio timeline. The colours do not match and the shooting format was different, 1080p for the video and 5016 x 3744 cropped to 1920 x 1080 to give same output format, so had to rely on the scaling done by Vegas. ;)
Enjoying live move :D but not the snow that is currently falling, flakes the size of ostritch feather ;)

_________________
Smile, it gives your face something to do!

http://www.vimeo.com/user1214873/videos
http://www.youtube.com/user/MrReggub


Thu Feb 13, 2014 3:07 am
Profile

Joined: Mon Nov 14, 2011 11:34 pm
Posts: 401
Post Re: Mantis real time - time-lapse tracking experiment
Hi Mike, just catching up.

That test is pretty impressive. There is no way that I could get repeatability like that. The SMS is also amazing though for a completely reason though as you say...... you would just not notice that sort of degree of movement of critters without the interval. Shows what can be done re natural history stuff. Love it ....

BTW Didn't see snow yesterday but got stuck in London with no trains running ... had to get to Heathrow and drive .... a nightmare. Got home early this morning and slates are off :(

_________________
For further info re keyframing see >> http://keyframemoco.com

"tick-tock ....click-clock ...."


Thu Feb 13, 2014 8:00 am
Profile
User avatar

Joined: Tue Nov 24, 2009 7:55 am
Posts: 951
Location: UK
Post Re: Mantis real time - time-lapse tracking experiment
JohnB wrote:
Hi Mike, just catching up.

That test is pretty impressive. There is no way that I could get repeatability like that. The SMS is also amazing though for a completely reason though as you say...... you would just not notice that sort of degree of movement of critters without the interval. Shows what can be done re natural history stuff. Love it ....

BTW Didn't see snow yesterday but got stuck in London with no trains running ... had to get to Heathrow and drive .... a nightmare. Got home early this morning and slates are off :(


Hi, John, thanks for the comments, it is amazing what time lapse can reveal, another path to explore :-) . Snow came and went today, sorry to hear about your delays and hopefully you have sorted the roof without too much damage. More doom and gloom tomorrow unfortunately so keep safe.

_________________
Smile, it gives your face something to do!

http://www.vimeo.com/user1214873/videos
http://www.youtube.com/user/MrReggub


Thu Feb 13, 2014 10:51 am
Profile

Joined: Wed Jan 20, 2010 4:50 pm
Posts: 561
Post Re: Mantis real time - time-lapse tracking experiment
Hi, the DOF is good too..

Re the synching. Looks like a sub-frame difference between the live and stopmotion shutters. This can be out by up to 50% of a frame, but no more, in such unsynced passes.

One way to prevent this is to use camera sync - if the camera allows it. Unfortunately DSLR's won't lock because the video monitor output is not synced internally to the recording... (well not the ones I've tested) otherwise if you shoot a few random live passes, one will get very close.

To achieve 100% live shutter sync, you need a professional cinema camera unfortunately, or conversely an old style video camera... Even then the motion blur in the live pass will create a slight mismatch.. G


Thu Feb 13, 2014 3:22 pm
Profile
User avatar

Joined: Tue Nov 24, 2009 7:55 am
Posts: 951
Location: UK
Post Re: Mantis real time - time-lapse tracking experiment
geraldft wrote:
Hi, the DOF is good too..

Re the synching. Looks like a sub-frame difference between the live and stopmotion shutters. This can be out by up to 50% of a frame, but no more, in such unsynced passes.

One way to prevent this is to use camera sync - if the camera allows it. Unfortunately DSLR's won't lock because the video monitor output is not synced internally to the recording... (well not the ones I've tested) otherwise if you shoot a few random live passes, one will get very close.

To achieve 100% live shutter sync, you need a professional cinema camera unfortunately, or conversely an old style video camera... Even then the motion blur in the live pass will create a slight mismatch.. G


Hi, Gerald, thanks for your comments, The video length is edited Ie. start camera sit down ;) press run, end run.... stop camera, so I have to give the start of the 500frame video my best guess based on the audio time line. I will check the number of frames taken in TL mode, on reflection I think the camera may have fired a frame or two before I hit the play button giving more than 500 shots, careless maybe :(
The canon 5d2 video can be triggered by infra red so it may be possible to access the push button contacts and sync using that via Mantis and an opto coupler.

_________________
Smile, it gives your face something to do!

http://www.vimeo.com/user1214873/videos
http://www.youtube.com/user/MrReggub


Fri Feb 14, 2014 1:04 am
Profile

Joined: Wed Jan 20, 2010 4:50 pm
Posts: 561
Post Re: Mantis real time - time-lapse tracking experiment
Hi.. As previously mentioned DSLR's do not sync. I tested a 5D MkII and it was a bit random. Something weird about the relationship between when they shoot frames in video mode and what is coming out the monitor output... you can use a bloop light for matching L/A passes, but generally it's not hard to find the nearest frame in your favorite editor with a quick composite - just nudge one layer in frame increments... bloop lights are useful on set, when you want to line up really fast to check, but not strictly necessary... G


Fri Feb 14, 2014 1:40 am
Profile
User avatar

Joined: Tue Nov 24, 2009 7:55 am
Posts: 951
Location: UK
Post Re: Mantis real time - time-lapse tracking experiment
geraldft wrote:
Hi.. As previously mentioned DSLR's do not sync. I tested a 5D MkII and it was a bit random. Something weird about the relationship between when they shoot frames in video mode and what is coming out the monitor output... you can use a bloop light for matching L/A passes, but generally it's not hard to find the nearest frame in your favorite editor with a quick composite - just nudge one layer in frame increments... bloop lights are useful on set, when you want to line up really fast to check, but not strictly necessary... G


Hi gerald, I just edited the above as your latest post came in when I re-read your post ref. DSLR sync. Take your point ref using editor.

_________________
Smile, it gives your face something to do!

http://www.vimeo.com/user1214873/videos
http://www.youtube.com/user/MrReggub


Fri Feb 14, 2014 1:49 am
Profile

Joined: Wed Jan 20, 2010 4:50 pm
Posts: 561
Post Re: Mantis real time - time-lapse tracking experiment
Hi Not sure if what I was saying was all 100% clear - so hope you don't mind this explanation.

Controlling the exact number of frame captured in a move is not always possible. In stop-motion it is, but in real-time you can't control it. Even if you use a cable to run the camera at the exact moment the moco starts, it will not be consistent. Every camera I have tested, tends to have a variable response time to a start trigger. So you should normally just roll the camera first and then run the moco. If you need to find a common frame in several runs, then you can use the bloop light, but as said it's generally easy enough to find the sync frame in an editor.

That said - the other aspect of "sync" that might not be so well understood is "shutter phase". This means that for each run, the exact point in the move where the shutter exposes should be the same. Otherwise we can never quite get a perfect lineup between passes - especially if things are moving fast. Such as the slight lag you are seeing when you comp your two shots.

To achieve phase sync in live action we need a method to start the moco move at a repeatable and precise moment in relation to the actual exposure time in the camera. In the old days of film this was easy. Most cameras had a shutter pulse output. Kuper or Mantis will read this pulse and only start the move exactly when it fires. Red cameras also have a similar pulse and Alexa can produce one via an adapter. Alternately, if the camera outputs a PAL or NTSC monitor signal, which is locked to the camera shutter, then this can also be used - but only at sound speed. But as said, in DSLR's it's still not accurate...

I hope this clarifies what I am saying... with DSLR's between live action passes you will always see some random offset in shutter phase. In practice though, this is not a big problem, depending on how you are using the shots... G


Fri Feb 14, 2014 7:52 pm
Profile
User avatar

Joined: Tue Nov 24, 2009 7:55 am
Posts: 951
Location: UK
Post Re: Mantis real time - time-lapse tracking experiment
geraldft wrote:
Hi Not sure if what I was saying was all 100% clear - so hope you don't mind this explanation.
G


An A.R.P. momement here Gerald. Age Related Problem. Thanks for the explaination and clarification. As you rightly point out it matters little especially when running real time move to quickly get an idea as to how a time lapse move will turn out.

_________________
Smile, it gives your face something to do!

http://www.vimeo.com/user1214873/videos
http://www.youtube.com/user/MrReggub


Sat Feb 15, 2014 2:27 am
Profile
User avatar

Joined: Mon Nov 22, 2010 1:18 pm
Posts: 1294
Post Re: Mantis real time - time-lapse tracking experiment
Hi Mike me again . Would it be possible to see how good the real video and timelapse moves can be composited . For example you walking into shot really time with you somewhere in the background timelapse . Dont worry if you shy its just a example but i hope you get the idea .
I know Gerald says theirs a problem with DSLRs but lets face it that's what we are all using so would be interesting to see the best results you can get .
Cheers D1


Sat Feb 15, 2014 3:49 am
Profile
User avatar

Joined: Tue Nov 24, 2009 7:55 am
Posts: 951
Location: UK
Post Re: Mantis real time - time-lapse tracking experiment
DISPLACEMENT 1 wrote:
Hi Mike me again . Would it be possible to see how good the real video and timelapse moves can be composited . For example you walking into shot really time with you somewhere in the background timelapse . Dont worry if you shy its just a example but i hope you get the idea .
I know Gerald says theirs a problem with DSLRs but lets face it that's what we are all using so would be interesting to see the best results you can get .
Cheers D1


Hi, D1,
My first couple of attemps have not worked out, the real / TL shots are working but the content composition does not composite well enough, mask is too compex, will try again during next week.

_________________
Smile, it gives your face something to do!

http://www.vimeo.com/user1214873/videos
http://www.youtube.com/user/MrReggub


Sun Feb 16, 2014 6:57 am
Profile
User avatar

Joined: Mon Nov 22, 2010 1:18 pm
Posts: 1294
Post Re: Mantis real time - time-lapse tracking experiment
Cheers for trying lets us know how it goes . Thought it would be handy for putting atmosphere on timelapse and for pulling off a few magic tricks :D
Dont worry if it was easy everybody would be doing it.
D1


Sun Feb 16, 2014 7:54 am
Profile
User avatar

Joined: Tue Nov 24, 2009 7:55 am
Posts: 951
Location: UK
Post Re: Mantis real time - time-lapse tracking experiment
DISPLACEMENT 1 wrote:
Cheers for trying lets us know how it goes . Thought it would be handy for putting atmosphere on timelapse and for pulling off a few magic tricks :D
Dont worry if it was easy everybody would be doing it.
D1


I don't think my "advanced" planning was up to scratch ;) ;)

_________________
Smile, it gives your face something to do!

http://www.vimeo.com/user1214873/videos
http://www.youtube.com/user/MrReggub


Sun Feb 16, 2014 10:35 am
Profile

Joined: Wed Jan 20, 2010 4:50 pm
Posts: 561
Post Re: Mantis real time - time-lapse tracking experiment
Hey Mike. Good on you for trying. Compositing can quickly become non-trivial, unless you just make a quick and dirty mix. Personally I thought the difference matt showed the potential quite well already, the rest is hard work... :) G


Sun Feb 16, 2014 3:32 pm
Profile

Joined: Sun Feb 16, 2014 9:55 am
Posts: 4
Post Re: Mantis real time - time-lapse tracking experiment
DISPLACEMENT 1 wrote:
Would it be possible to see how good the real video and timelapse moves can be composited. For example you walking into shot really time with you somewhere in the background timelapse.


Not sure if this is quite what you're talking about here, but here's a test I did with no motion control, just a Canon 5Dmkii and 17-40 f/4 shooting timelapse then a video clip and then roughly composited together in After Effects. Clearly there's some color work to be done, and the timelapse loops, but you get the idea.



Sun Feb 16, 2014 11:17 pm
Profile
User avatar

Joined: Mon Nov 22, 2010 1:18 pm
Posts: 1294
Post Re: Mantis real time - time-lapse tracking experiment
Yes that's the sort of thing . Think the motion control makes it a bit more problematic though . Sure it will be sorted sometime :D


Mon Feb 17, 2014 11:35 am
Profile
User avatar

Joined: Tue Nov 24, 2009 7:55 am
Posts: 951
Location: UK
Post Re: Mantis real time - time-lapse tracking experiment
Hi, NattB, similar to what I am after but with motion. My main problem was I recorded the video sequence with a Pan of my workshop whilst I "fiddled" around on the left of the shot, did same using TL while standing to the right, unfortunately the pan was such that I was either in the left or the right of the composite and not in both :cry: I have since re-done the shot but school Half Term is upon us and my 10 yr old nephew has taken over as DOP ;) on his "Green Lizard racing circuit" project for the rest of the week. So far 2 seconds of Stop Motion shot and eight laps to go, although that might get cut to one at the current rate :D

_________________
Smile, it gives your face something to do!

http://www.vimeo.com/user1214873/videos
http://www.youtube.com/user/MrReggub


Tue Feb 18, 2014 1:01 am
Profile
User avatar

Joined: Mon Nov 22, 2010 1:18 pm
Posts: 1294
Post Re: Mantis real time - time-lapse tracking experiment
Hi Mike Nate you prob seen this one but do like the mixing of different time frames . Wouldn't it be great with timelapse to

Happy animating Mike bet your nephew loves visiting


Tue Feb 18, 2014 11:59 am
Profile
User avatar

Joined: Tue Nov 24, 2009 7:55 am
Posts: 951
Location: UK
Post Re: Mantis real time - time-lapse tracking experiment
DISPLACEMENT 1 wrote:
Hi Mike Nate you prob seen this one but do like the mixing of different time frames . Wouldn't it be great with timelapse to

Hi, D1, thanks for the link, this is similar but uses stacked GIF images http://blog.spoongraphics.co.uk/tutoria ... -photoshop
DISPLACEMENT 1 wrote:
Happy animating Mike bet your nephew loves visiting

He sure does, photography, bug hunting and deer stalking are a few half term activities :D

_________________
Smile, it gives your face something to do!

http://www.vimeo.com/user1214873/videos
http://www.youtube.com/user/MrReggub


Wed Feb 19, 2014 7:37 am
Profile

Joined: Wed Jan 20, 2010 4:50 pm
Posts: 561
Post Re: Mantis real time - time-lapse tracking experiment
Hi - cinemagraphs are cool...

Combining bits of natural realtime movement with either frozen time, or timelapse are similar ideas. But soon as the camera is moving it all gets more complicated to composite. If you just use a simple soft matte then it has to travel, which is either done with manual key framing (rotoscoping), or by motion tracking. Either way it can draw attention to itself if not done carefully. It's also time consuming. An alternative is to use chroma screen to separate elements, but this is not always practical in real life.

To achieve true frozen time is not really possible when moving the camera, though it is often achieved by cheating. In controlled situations actors are often rigged to hold them as though in mid movement, then a slow-motion pass is made to reduce any remaining movement. Otherwise even 3D or "time-slicing" is used...

The Kinetic example has some subtle movement, but that is done with pan and scan, cropped from a larger plate.

There are other techniques too, but basically to combine passes at different speeds takes some hard work in post...


Wed Feb 19, 2014 3:43 pm
Profile
User avatar

Joined: Tue Nov 24, 2009 7:55 am
Posts: 951
Location: UK
Post Re: Mantis real time - time-lapse tracking experiment
Two real time video moves.
This is a "watching paint dry" video, a two run composite using difference blending.
Dolly move is 100cms total moving forward 50cms and then back to start whilst the Pan, Tilt and Roll axes each move throught 360 degrees. The lens moves across the power and signal wires that themselves are moving slightly in a couple of points producing banding of the background. I an quite pleased with the result that shows quite good repeatablility of the rig movements.
Servocity Pan, Tilt and Roll rig, modified to run with stepper motors, Dolly home brew using 2 x 100mm square x 2000mm carbon fibre tubes. Stepper drives are Leadshine DM442 driven by Mantis 2 hardware and software.

D1, will do that real/tl move in the next day or so hopefully.


_________________
Smile, it gives your face something to do!

http://www.vimeo.com/user1214873/videos
http://www.youtube.com/user/MrReggub


Sun Mar 02, 2014 2:28 am
Profile
User avatar

Joined: Mon Nov 22, 2010 1:18 pm
Posts: 1294
Post Re: Mantis real time - time-lapse tracking experiment
Hi Mike very impressive for a watching paint dry video :D . Just posted a live move test on Johns thread just Df arduino again but interesting to see how it copes with data .
D1


Sun Mar 02, 2014 10:21 am
Profile
User avatar

Joined: Tue Nov 24, 2009 7:55 am
Posts: 951
Location: UK
Post Re: Mantis real time - time-lapse tracking experiment
DISPLACEMENT 1 wrote:
Hi Mike good to see timelapse and real video can be shot on the same rig . Did you try overlaying the two shots would be interesting to see what the registry is like.
Very nice i am quite envious of all this live move stuff .
D1


Hi, D1, better late than never. Sorry about the model but my dolly bird was on a fashion shoot ;)

A 20 second video with panning left to right was composited with a 500 frame time lapse sequence with the same movement, video and time lapse sequences controlled by Mantis 2 software/hardware.
Sync by eye, main area of concern is the light bulb "alignment" caused by different flare regions otherwise I am quite pleased with the test result result on the verticals.
After alignment there are a few less frames in the final clip.


_________________
Smile, it gives your face something to do!

http://www.vimeo.com/user1214873/videos
http://www.youtube.com/user/MrReggub


Thu Mar 27, 2014 3:46 am
Profile
User avatar

Joined: Mon Nov 22, 2010 1:18 pm
Posts: 1294
Post Re: Mantis real time - time-lapse tracking experiment
Hi Mike that's not bad at all . Very usable look forward to some interesting time bending stuff going on did i read you off to Scotland again .
Didn't manage it last year and not sure if gonna make it this year either :(


Thu Mar 27, 2014 11:38 am
Profile
User avatar

Joined: Tue Nov 24, 2009 7:55 am
Posts: 951
Location: UK
Post Re: Mantis real time - time-lapse tracking experiment
DISPLACEMENT 1 wrote:
Hi Mike that's not bad at all . Very usable look forward to some interesting time bending stuff going on did i read you off to Scotland again .
Didn't manage it last year and not sure if gonna make it this year either :(


Hi, thanks for the comments, the colour boundry draws the eye so quite a harsh test, opens up some interesting possibilities.
Off the Scotland hopefully May 1st, same day as new car may arrive, timing is not good but will sort that out when the time comes :cry: Will be in the Tyndrum area, so good for Ranoch Moor,Glen Orchy etc and some gold panning not. Will not try taking the Servocity rig though ;)

_________________
Smile, it gives your face something to do!

http://www.vimeo.com/user1214873/videos
http://www.youtube.com/user/MrReggub


Thu Mar 27, 2014 11:58 am
Profile
User avatar

Joined: Mon Nov 22, 2010 1:18 pm
Posts: 1294
Post Re: Mantis real time - time-lapse tracking experiment
Lots and lots of possibility's i am still very envious of that real time thing .
Know Tyndrum well. The campsite their i use a lot on my way up to the top bits . Funny place as i wild west thing going on and a lots of bikers


Thu Mar 27, 2014 12:10 pm
Profile
User avatar

Joined: Tue Nov 24, 2009 7:55 am
Posts: 951
Location: UK
Post Re: Mantis real time - time-lapse tracking experiment
DISPLACEMENT 1 wrote:
Lots and lots of possibility's i am still very envious of that real time thing .
Know Tyndrum well. The campsite their i use a lot on my way up to the top bits . Funny place as i wild west thing going on and a lots of bikers


Don't forget the chippy and the other GoldMine in the form of The Green Welly :D

_________________
Smile, it gives your face something to do!

http://www.vimeo.com/user1214873/videos
http://www.youtube.com/user/MrReggub


Fri Mar 28, 2014 11:04 am
Profile
Display posts from previous:  Sort by  
Post new topic Reply to topic  [ 30 posts ] 


Who is online

Users browsing this forum: Majestic-12 [Bot] and 9 guests


You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot post attachments in this forum

Search for:
Jump to:  
cron
Powered by phpBB © phpBB Group.
Designed by Vjacheslav Trushkin for Free Forums/DivisionCore. pozycjonowanie