Home of Professional Cinematography since 1996

Miniatures fps

Published : 9th January 2004


I am familiar with the equation for figuring out the camera frame rate for shooting miniature objects. The model we are to shoot is 1/48 scale. The square root of 48 is 6.9 and when multiplied by 24fps to achieve the appropriate frame rate for filming scaled models, I calculate 166 fps.

But I have a few questions...

1) Is 166 fps considered the speed to shoot for action/camera movement to appear as normal when transferred at 24fps ?

2) If I want to slow down the action do I consider the 166 fps to be my
normal shooting base so the 332 fps will slow down the action by half of
"normal" action speed ??

3) How can I best calculate the appropriate speed of camera movement?

4) How best to calculate the speed of moving objects?

5) Does this equation change from 35mm to 16mm?

Many thanks in advance.

Best Regards,

Jim Sofranko
NY/DP



Jim Sofranko wrote:

>...1) Is 166 fps considered the speed to shoot for action/camera >movement to appear as normal when transferred at 24fps ?...

166 fps is the correct frame rate to make 1/48 scale objects affected by gravity appear to fall at a natural rate. However, since the atmosphere can't be scaled, the mass of the objects and thus, how they are affected by buoyancy will enter in. If the motion is not actually gravity driven you may be able to get away with slower frame rates. Test it first, use several frame rates faster and slower.

>2) If I want to slow down the action do I consider the 166 fps to be my >normal shooting base so the 332 fps will slow down the action by half of >"normal" action speed ??

Yes, but test it first, using several frame rates faster and slower (see buoyancy, above.)

>3) How can I best calculate the appropriate speed of camera >movement?

Divide 166 by 24 (6.9) to find the factor by which time will be multiplied. One second of film time will equal 6.9 seconds of screen time. Then decide how many seconds of camera movement will be appropriate for your move.

>4) How best to calculate the speed of moving objects?

See above.

>5) Does this equation change from 35mm to 16mm?

No.

Have fun! A 1/48 scale is pretty small, 1/24 is a much better choice, but you may not have that option. What is the subject?

Wade K. Ramsey, DP
Dept. of Cinema & Video Production
Bob Jones University
Greenville, SC 29614



Wade pretty much answered things, but a few more general conceptual observations:

As Wade pointed out, the sq root formula has to do with gravity...it really depends on what you are doing a lot. If you have a move on a miniature that is not being blown up etc, you might well want to shoot at different speeds. For instance, if the miniature is pretty much static and you are moving on it, you can shoot at any frame rate...what you REALLY need to worry about is depth of field :

You need to be able to hold the appropriate DOF to what you would hold if it were the real world : If you are shooting a 1/12 scale for instance (to make the math easy) and your "real world" dolly shot would be a three second dolly shot of 18 feet, you might be holding focus from , say, 5 ft to infinity on your 18 mm lens.

You need to shoot your miniature on an 18 mm lens again, but now you need to hold from 5" to infinity (or the back of the model) (that means a really fat stop).

Your real move would have been 3 secs for 18 feet at 24fps. That's 72 frames of film to move 18 ft.

You need to shoot 72 frames of film while moving 18 ". If the model is totally static , you could shoot them as 1 sec exposures or at 24 fps or whatever, as long as you got the right number of frames to cover the move in three secs of screen time.

This is where motion control is great - you program your move by frames, not by time - you can adjust your frame rate (or exposure time) based on what stop you need and the rig will move at the appropriate speed to make it all work.

Regarding what speed to shoot to slow down the motion on a miniature…test test test...

I tend to think of a shot in "real world" terms, figure out how many frames it would be in the real world, and then work backwards from there, but with explosions and etc, there are lots of variables...the type of material used to create the explosion determines the speed at which things happen and that determines the frame rate you need to duplicate what would happen in the real world...and from there, how much you would want to slow the shot down for dramatic effect.

When we exploded some bits and pieces of the Coliseum a couple of summers ago for the Core, we were inside shooting maybe a 12' high chunk of coliseum and lighting it with a Dino and lots of maxi's... (Lots is more than 5) much as the producers hate this, miniature does not mean small lighting budget.

I hope this ramble is useful.

Mark Weingartner
LA based
(blowing up small things and making them look big)



I'm just in the middle of a shoot involving miniatures and scaled speeds with water.

We are using the square root rule as the basis for our speeds but we're also using a hard disk based video assist that will simulate 3 to 200 FPS and are testing various speeds this way according to the shot.

It's really a great way to get exactly the right speed.

Of course we have to deal with water droplets that are the wrong size but are trying various techniques to reduce surface tension to minimise the effect.

Probe & Periscope with waterproof housing and spinning rain deflector!

Cheers

Geoff Boyle FBKS
Director of Photography
EU Based
www.cinematography.net



Wade K. Ramsey writes :

>166 fps is the correct frame rate to make 1/48 scale objects affected by >gravity appear to fall at a natural rate. However ...snip...you may be able >to get away with slower frame rates.

Wade and Mark,

Thanks for all the helpful info. The subject is a O scale train on a miniature terra firma set which is all 1/48 scale. We are shooting with a 16mm Actionmaster and probably an Innovision probe and a 8-64mm zoom. The train has a crash which we will shoot at the maximum 500 fps but the majority of the action is the drama built up before the crash.

There is a series of shots before the crash of the train travelling on the tracks. We have control over the speed of the train and camera but not the smoke coming from the engine nor the intensity of the practical lights on the train model. It is a night exterior with added fog and/or smoke.

Some of the shots are static of track and signal lights but most of are the train travelling at what in real life should look like about 70 mph. Of course, there is no time or budget for tests so, unfortunately, I need to wing it. I have asked for the testing but it is not possible for several reasons.

So here's my question, aside from the crash shots, what is a good way to determine a camera speed and train speed?

Thanks,

Jim Sofranko
NY/DP



> There is a series of shots before the crash of the train travelling on the >tracks. We have control over the speed of the train and camera but not >the smoke coming from the engine nor the intensity of the practical >lights on the train model.

You are possibly stuck here in an "either or". . .

If you can't beef up the brightness of the practicals you may have to shoot slower and have the smoke look like it is moving way to fast.

By the way, we have often used various kinds of flashlight bulbs - mini mag and also the different kinds of underwater flashlights to get hot scaled-down light sources.

If you need the smoke to be right, the practicals will probably be pretty wimpy...it depends on the shots

IF you start with the formula you probably won't be too far off on the smoke speed.

As far as train speed, depending on the shots, you will need to figure the vehicle speeds the same as you would if it were a full size car.

I think that 70 mph is approx 103 ft/sec

The scaled distance for the train at 1:48 is approx 2 ft/sec = 70 mph in its scaled surroundings

So...if you are thinking 24 fps real world, you need to have the train travel 2ft in the time it takes your camera to expose 24 frames

If you are in a 30 fps world, you need to have the train travel 2 ft in the time it takes your camera to expose 30 frames.

So...If you are shooting at 168 fps, you need the train to be travelling at approx 14 fps.

I doubt if you can get an O gauge train to go that fast, and you may not be able to get the stop to get the DOF you need, so I think you will be up against how much stop you can get, how much light you can get out of the practicals, how fast the train will go, and then back into a frame rate which will probably not be fast enough but will have to do.

1/48 scale is pretty darned small for a train crash

A lot will depend on your client's expectations - DOF may be less of an issue because night stuff is often shallow, and for low angles where the train is coming towards or going away from you, you may not have clues in the frame that show the speed etc...

I would STRONGLY suggest that you rent a frame adjustable video record playback system like a PVR or PAR system so you can pre-vis a frame rate with a train speed to see if it is working. Don Canfield at Gear & Rose may have one...Alex Fernbach's company may have one...I am not sure where else in NY you would find one but it will be a lot cheaper than a re-shoot...even if you have to hire the operator for it.

Unless you use a special videotap and a special integrator box, you will be stuck with trying to figure out what is going on through the motion blur of 29.97 video, but at least you can get the timing right...but if you can't get a video system to do this, you can at least set up the shot... figure out how much screen time it should take for the train to go through the shot... figure out how many frames of film that should be... and do the math to figure out how fast the train has to go at your frame rate to make the right number of frames...or work it the other way to figure out how fast you can run the camera with the train running at whatever speed it can go

Ain't Visual Effects Fun!

Good luck, Jim.

This post will self-destruct in 24 hours

Mark Weingartner



Jim Sofranko wrote :

>...So here's my question, aside from the crash shots, what is a good way >to determine a camera speed and train speed?

Hopefully, Mark or someone can bail me out if my math is wrong, but I figure that a train travelling at 70 mph is covering 1,232 inches per second. At 1/48 scale, that would be 25.6 inches per second. I think the question is, how fast can your model actually run? You need to shoot it at a higher than normal frame rate because, even though gravity isn't involved, a model train tends to wobble a bit down the track and you need to smooth it out. You need at least 48 fps for that and preferably more, if the model can make enough track speed. That would be 51 inches per second, which I think seems pretty fast for an O gauge train. Will you have a lot of straight track to get it up to speed?

Also, since you can't control the lights inside, you can't afford to run the frame rate very high. What stop do you need for depth of field? Shoot it on 18!

Is this supposed to look like a full scale train or is it a commercial for toys? IOW, how realistic does it need to be?

The wreck at 500 fps seems like it will be a pretty slow ballet! How does it wreck? Off a bridge or into a barrier?

Wade K. Ramsey, DP
Dept. of Cinema & Video Production
Bob Jones University
Greenville, SC 29614



Geoff Boyle said:

>Of course we have to deal with water droplets that are the wrong size but >are trying various techniques to reduce surface tension to minimise the >effect.

Are you floating alcohol on the water? Do people still do that for smaller droplets?

Roderick
Az. D.P. (done some miniature pyro before!)
www.restevens.com
12On / 12Off

====================================================

Would it be possible to shoot these shots stop motion (go-motion) on a motion control rig? If so, shooting at a 1/4 of a second would give you much more control over a greater DOF and, with a couple of tests via a video assist to some animation programme on a PC or Mac you could easily work out travelling speeds by eye. I know this would mean shooting without smoke but that can very effectively be added in post.

Of course a crash would be pretty difficult but hey, it's only my sixpence worth.

Regards

Chris Maris
UKDP



Following on this interesting thread, having considered fps, esp. for falling objects due to gravity, speed of camera movement, object movement and depth of field ... what about a variable a little more difficult to control such as fire.

I mean flames rather than fire-balls. How would you calculate, manipulate the flames in, or on, a miniature burning building to look/ feel correct?

Mike Costelloe
UK Based Cam Op/ Lighting Cam



>Will you have a lot of straight track to get it up to speed?


We will have about 20' of track before it enters the shooting area of the propped miniature. But the train is 14' long so it's not much ramp up time for the train.

>Also, since you can't control the lights inside, you can't afford to run the >frame rate very high... Shoot it on 18!

I will be inquiring this morning if 7218 is available with double perf as it doesn't say on the Kodak website. I do agree with Mark that the DOF may not be as critical because it is night and it will be fogged up into the distance.

>Is this supposed to look like a full scale train or is it a commercial for >toys? IOW, how realistic does it need to be? The wreck at 500 fps >seems like it will be a pretty slow ballet! How does it wreck? Off a bridge >or into a barrier?

It is for a TV documentary and it is a correct 1/48 scale. It runs into a bridge and flies into the water. But the crash is less important. I'm thinking 500 fps will give us enough info to play with in post. Plus that's the maximum speed of the Actionmaster. 500 fps only slows the action by about 3 times if the base speed is 166 fps. I understand the need for the equation which suggests 166 fps but if the smoke from the train is the only factor to consider for that frame rate perhaps I may have some leeway to slow down the fps to around 72 - 96fps in the shots before the crash.

My plan, at this point, is to calculate the practicals, speed of the train, DOF and try to find a camera speed which will accommodate all those factors.

Each shot has it's own applications as some shots are just tracks or signals without the train which won't need to be 166 fps but will need to be slowed somewhat, like 48-60fps, to accommodate the smoke and fog so it matches. I'm thinking that if the train is slowed down by half, my fps can be half. Does that seem to make sense?

Thanks for all the help.

Jim Sofranko
NY/DP



Jim Sofranko wrote:

>...We will have about 20' of track before it enters the shooting area of the >propped miniature. But the train is 14' long so it's not much ramp up >time for the train....

I don't think there is any possibility you can get that train up to the speed needed for a 70 mph look for any frame rate over 72 fps, and you'll be fortunate to make that, IMHO. I sincerely hope I'm wrong, but OTOH I'm not convinced you'll want much higher frame rates, considering the smoke and other factors.

>...I will be inquiring this morning if 7218 is available with double perf as >it doesn't say on the Kodak website...

It was, as of the middle of last month, but we had to request it.

>..It is for a TV documentary and it is a correct 1/48 scale. It runs into a >bridge and flies into the water. But the crash is less important. I'm >thinking 500 fps will give us enough info to play with in post...

Keep us posted on where we can see the final results, if possible. I'm really interested in how this all works out!

>...I'm thinking that if the train is slowed down by half, my fps can be half. >Does that seem to make sense?

Yes.

Wade K. Ramsey, DP
Dept. of Cinema & Video Production
Bob Jones University
Greenville, SC 29614



Mike Costelloe wrote:

>...I mean flames rather than fire-balls. How would you calculate, >manipulate the flames in, or on, a miniature burning building to look/ >feel correct?

Various tricks have been used. You do need to use the scale frame rate to slow down their movement. Sometimes the real flames can be positioned a greater distance behind the model to give them better scale, positioning the camera so they appear to be coming off the top of the model. Matting them in optically has been tried, but laying them in digitally is probably the best way to make it look right. Fortunately, we don't have to worry about the feel, until someone comes up with Feel-O-Vision

It's also important, if the scene is in "sunlight", not to allow the flames to be obviously much brighter than the sunlit surrounding scenery. OTOH, if the scene is overlit the flames won't show up well. Yellow flames are always about the same intensity, large or small. I think lighting the scene to about a stop under real sunlight works about right. If you are doing this miniature outdoors for safety (oh, that!) you might want to scrim the real sun down a bit to make the flames more visible.

And speaking of safety, you keep the flammable fire starter far from the scene (the next county is good) and you have fire control people very near with extinguishers in hand.

Wade K. Ramsey, DP
Dept. of Cinema & Video Production
Bob Jones University
Greenville, SC 29614



>I mean flames rather than fire-balls. How would you calculate, >manipulate the flames in, or on, a miniature burning building to look/ >feel correct?

Test test test of course

Fight like hell for enough budget to make the miniature as big as possible to start with. Build sections of the model for cutaways that are as large as possible - a few of those can really help sell the wide shots

As my mentor Richard Yuricih always says,

"My favourite scale is 1:1."

Mark Weingartner
LA based

Look for some of my flame elements in soon to open films and TV promos



Assuming no motion control...

If you have static shots of the set, can you shoot them clean at a slow frame rate and a fat f/stop to get a bit more DOF (which will help sell scale a bit) and then do a separate smoke/fog pass at a faster frame rate with a thinner stop or better yet, more light? You would want to cover the ground with black velvet or something like that so you could DX the smoke/fog in post as a separate layer - this might involve pulling out trees and that sort of thing or maybe not - I would say that if you have the time (which you probably won't) it would be worthwhile to try this - for shots with the train, you have a train to hold interest, but with shots with no train in them, people will watch the smoke and fog

Another way to do this if you have shots of signal lights, for instance, is to light the miniature for a fat enough stop and high enough frame rate to shoot the fog/smoke and the set in one pass and then shoot a separate pass (obviously this only works in a lock-off) at a much slower frame rate BUT THE SAME STOP AND FOCUS SETTINGS to get a good exposure on the signal bulbs.

Gotta run - good luck!

Mark Weingartner
LA based



> I will be inquiring this morning if 7218 is available with double perf as it >doesn't say on the Kodak website.

I ordered some 7218 about a month ago...The fellow I was speaking to (who was in LA) said that it was a special order, and he didn't have any showing in current inventory in his computer. Don't know what kind of lead time they need, as I didn't need it for my shoot but asked out of curiosity.

George Hupka
Director/DP
Downstream Pictures
Saskatoon, Canada



I noticed - no one has commented on the alcohol for miniature water droplets.

Do people still do this? I read many moons ago that by floating alcohol on top of the water, it created smaller peaks, ripples and droplets? Any truth to this? Geoff have you tried it?

Roderick Stevens
Az. D.P. (learning to speak with an Irish Brogue)
12On / 12Off



Don Canfield wrote:

>Not just alcohol, but detergent, PhotoFlo, or anything that will reduce the >surface tension of the water, and allow it to make smaller "scaled down" >droplets...

What do you use as an anti-foaming agent to keep the PhotoFlo or detergent from sudsing like crazy?

Wade K. Ramsey, DP
Dept. of Cinema & Video Production
Bob Jones University
Greenville, SC 29614



>I noticed - no one has commented on the alcohol for miniature water >droplets.

Not just alcohol, but detergent, PhotoFlo, or anything that will reduce the surface tension of the water, and allow it to make smaller "scaled down" droplets...

Don Canfield
New York



>I noticed - no one has commented on the alcohol for miniature water >droplets.

I did suggest it but got the response that they knew I liked Beychevelle and that I'd have to wait until the wrap before I could have any.

Seriously, the PM said "have you seen the size of that tank!, how much alcohol do you think there is in Copenhagen?"

As for wetting agent, I wanted to go this route but again the size of the tank was a problem, if we used too much it would lose us a day or so getting it out and the tank re-filled.

Cheers

Geoff Boyle FBKS
Director of Photography
EU Based



I did suggest it but got the response that they knew I liked Beychevelle and that I'd have to wait until the wrap before I could have any

I knew, as soon as mixing alcohol with water was mentioned, that Geoff would see this as a horrific violation of protocol!!

Kent Hughes
DoP
SoCal



>What do you use as an anti-foaming agent to keep the PhotoFlo or >detergent from sudsing like crazy?

You only need a very small amount. Drops. It takes a little experimentation. The sudsing is a direct result of the reduction in surface tension. Alcohol works because it is less viscous generally. Water with a reduction in surface tension remains as viscous, but is more slippery, if you know what I mean. It's hard to explain.

Don Canfield



> Assuming no motion control...

Oh no, please never assume that....

Yikes!!!!

Don Canfield
Gear+Rose Motion Control
NY
www.gearandrose.com



>You only need a very small amount. Drops. It takes a little >experimentation. The sudsing is a direct result of the reduction in >surface tension.

So you don't dilute the Photoflo to the same degree that is recommended for use on negatives? But if you used alcohol you'd have to use a lot more, wouldn't you? I'm thinking of a temporary tank we built that was about 30 ft. x 50 ft. x 3 ft deep. A lot of water in that tank. Have you used it in a tank about that size or larger? I'm thinking that if you put in enough Photoflo to actually "thin" the water you'd have suds all over the place. We were doing a storm scene, so there was a lot of disturbance in the water. I've never tried Photoflo in a miniatures tank, so I'm just guessing based on what it does diluted as recommended for film drying.

Wade K. Ramsey, DP
Dept. of Cinema & Video Production
Bob Jones University
Greenville, SC 29614



>So you don't dilute the Photoflo to the same degree that is >recommended for use on negatives?

Right. PhotoFlo is a "wetting agent", it makes the water flow off the negs. It does so by reducing the surface tension of the water, causing it to loose it's ability to make drops and "hang". You don't need to dilute with the flow. You are only trying to reduce the water's ability to "clot" into drops, and thereby make smaller drop lets.

I haven't done this in a long time. The most successful gag I did was to create "Niagara Falls" for a plate using a drop tank of about 12' long, 3' deep, and about 1' wide. This was tipped and dumped into a shallow basin on top of a shaped falls lip. It then poured over the lip and dropped into a catch basin about 6' below. We were only interested in the top of the falls. However, we didn't get a lot of suds, but some. Some suds cascaded over the falls, but very small bubbles. It was pretty convincing. We then pumped the water back to the drop tank after takes.

I really don't remember how much "-Flo" we used, but it was not very much.

Don Canfield



Don Canfield writes :

>PhotoFlo is a "wetting agent", it makes the water flow off the negs.

One of those wonderful contradictions that we live with in photography: we
use a "wetting agent" to _dry_ the film.

A bit like using a "fixer" to _remove_ the silver.

Or "reversal film" which doesn't reverse the image but copies a positive
scene as a positive image.

(They are all probably distantly related to Windows' intuitively-named
"start" button which you use to shut down the computer:)

Dominic Case
Atlab Australia



Hi all,

The horse has bolted on this one, I know, but I've been away for the last week. I have been using the Excellence on food shoots for some time now and love it to bits. As Geoff said, amazing quality for all those aerial images & bits of glass, & true to stop for a change in this type of lens.

There is only one in Australia so far, at Cameraquip, & I have been recommending it to as many as possible to make sure it stays here. I too have asked that Optex make a longer lens than 28mm for it - a pain when you have a close up of a tomato and have to cover 30' of background! & of course, true to Optex, way too expensive - A$1800/day +++ = $2210.

David Wakeley acs, Sydney.