Home of Professional Cinematography since 1996

HDCAM Time Lapse

Published : 29th November 2003


Separate from the HDW-F900's single frame record capability, HDW-750 and HDW-730 HDCAM camcorders both use an optional cache board, similar to what's offered with the MSW-900 IMX standard def system. The cache board provides up to 7 seconds of RAM recording : when the REC start button is pressed, action that occurred up to 7 seconds earlier is recorded to tape. No more missing those "must capture" special moments.

The second capability is Intermittent Recording. In the menus you select the how long the event occurs in real time, then how long you want that event to run on tape. Time selections range from frames to weeks and all of the recording is done without burning up VTR hours and head time and without operator intervention other than setting the real time and tape time durations. You can review the tape immediately to make sure you've captured the time-lapse motion look you wanted.

These features make the cache board great for many nature and wildlife cinematography projects, working with non-professional talent and kids, even news applications. It's a lot of capability added to the camcorder for all of $3,000 MSRP.

Stop by the Fletcher display if you're in Jackson Hole next week and I can show you the cache system. Or feel free to contact me offline anytime.

Dave Anderson
Fletcher Chicago
www.fletch.com

Fletcher is a Sony National CineAlta Master Representative



I've used the time lapse features on the MSW-900 (in 30P mode) on straightforward things like clouds & sunsets, and I've been very pleased with the results. I haven't done enough time lapse on film to know when you really need to go to a longer shutter speed...so I don't know when that becomes a critical advantage.

I'm also a great fan of the cache board...I leave it set to 2 seconds as a default on all of my documentary shoots, which has already saved me a couple of times...and I've used it at the full 7 seconds on a couple of occasions. I can certainly think of a lot of situations in the past when I would have loved to have had this kind of feature available.

George Hupka
Director/DP, Downstream Pictures
Saskatoon, Canada



We do this regularly with our Sony HDW-750 HDCam with optional HKDW-703 picture cache board. This board stores up to 150 frames before recording to the tape, so the heads aren't constantly spinning. You just set the total record time (up to 100 hours) and the record interval (5 sec. to 40 min.) and start recording.

Renting one of these seems like a much more elegant solution than any of the others mentioned. While the 750 shoots at 1080/60i instead of 1080/24p, it shouldn't matter for time lapse.

Jason White
Audubon Nature Institute



Jason wrote :

>1080/60i instead of 1080/24p, it shouldn't matter for time lapse

Let's not confuse (P)progressive capture/record vs (I)interlaced capture/ record. A 60 i frame is made up of 2 fields CAPTURED a 60th of a second apart where as a progressive frame is captured at the same instant.

Both should look exceptional but not the same.

Regards

Michael Bravin
Chief Technology Officer
Band Pro Film/Video, Inc
www.digiprimes.com



>A 60 i frame is made up of 2 fields CAPTURED a 60th of a second apart >where as a progressive frame is captured at the same instant.

Sorry about that. You are absolutely correct. Don't want to confuse anyone.

>Both should look exceptional but not the same.


I understand the interlace/progressive issue, but correct if I am wrong, a frame of progressive video with no motion will look the same as an interlaced frame with no motion. For almost all of the time lapse that I have done (nature mostly), the motion has been almost imperceptible, so both would look the same (disregarding frame rate which can be altered in post).

Jason White
Audubon Nature Institute



Michael Bravin wrote :

>A 60 i frame is made up of 2 fields CAPTURED a 60th of a second apart >where as a progressive frame is captured at the same instant.

Forgive me as I have found this confusing.

In Interlaced the image is drawn on the screen in fields, line 1 then 3, then 5, etc. all the way down to 525, then the even fields are drawn one line at a time. Is this not so? That was my understanding.

I had thought that the capturing of the image was the same process in interlaced.

Please correct me if I am wrong.

I ask because I thought progressive capture/display would mean that the scan lines go sequentially, 1-2-3-4-5-6. . .to the limit of your resolution. However the way it is sometimes described it seems as if all the pixels capture at the same moment.

Can someone clear this up? I do realize that we are talking about fractions of a second here, but I'm curious.

Thanks

Steven Gladstone
Cinematographer - Gladstone Films
Cinematography Mailing List - East Coast List Administrator
Better off Broadcast (B.O.B.)
New York, U.S.A.



In Interlaced capture, the system records the odd lines for the first 1/60 of a second, then the even lines for the next 1/60 of a second. Each is half resolution because only the odd or even lines are captured during each exposure. That¹s why the ³frame² appears to flicker if there is any camera or subject movement.

This is why trying to create slow-motion from an interlaced capture results in a half-resolution image. The Varicam, for example, can capture progressive at up to 60 fps, so the slow motion is full resolution.

In progressive capture, the system records all the lines for a duration determined by the shutter ³angle² and whatever frame-rate is selected.

That¹s why you don¹t get the ³NTSC flicker² with progressive.

I¹m no Engineer, but I believe this is correct.

Best regards,
Leo Ticheli
Director/Cinematographer
Birmingham/Atlanta



Yes Leo, but are the lines scanned one at a time, or do all pixels capture at the same instant?

Steven Gladstone
Cinematographer
Gladstone Films
Brooklyn, N.Y. U.S.A.
East Coast List administrator - Cinematography Mailing list



Steven Gladstone writes :

>Yes Leo, but are the lines scanned one at a time, or do all pixels >capture at the same instant?

I believe all of the CCD cells capture the picture simultaneously and that "snapshot" is buffered. The information is then read out of the buffer, converted to a digital word, and stored to tape / hard drive sequentially. Yes?

Brent Reynolds
DP
Tampa, FL



>Yes Leo, but are the lines scanned one at a time, or do all pixels >capture at the same instant?

I'm no engineer, either, but as I understand it, all the pixels capture at the same instant and their charges are stored. Then the system clocks out the charges in order, from one side of the screen to the other, line by line, because the CRT that will ultimately reproduce the image scans that way.

In progressive scan, every line can be clocked out in order for progressive scan environments, but if we are having to reproduce the image in an interlace environment, they are clocked out as alternate fields, 1, 3, 5, 7, 9, then 2, 4, 6, 8, 10, etc., to be compatible with interlace requirements. But both of these fields were actually captured at the same time, so there is no jagged effect on moving objects.

Correct me if I'm in error here.

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



>Yes Leo, but are the lines scanned one at a time, or do all pixels >capture at the same instant?

For interlaced scanning, the imaging sensor is exposed once per field. For progressive scanning, the imaging sensor is exposed once per frame.

Andreas Wittenstein
BitJazz Inc.
http://www.bitjazz.com/sheervideo



Wade Ramsey wrote :

>But both of these fields were actually captured at the same time, so >there is no jagged effect on moving objects.

I'm not an engineer either, but I believe you are correct about everything in your post except the above statement. In the case of interlaced video both fields are not captured at the same time (1/60th of a second apart in NTSC or 1080i) which is why we DO get jagged edges in fast moving objects.

Jason White
Audubon Nature Institute



>I ask because I thought progressive capture/display would mean that >the scan lines go sequentially, 1-2-3-4-5-6. . .to the limit of your >resolution.

This has changed when the video cameras changed from tube to CCD. In the tube, the pixels were captured sequentially, eg. the last pixel bottom right was quite a field later than the first pixel top left in the CCD, there is an integration over 1/60 or 1/30 seconds which is the same for all pixels of a field (interlaced) or frame (progressive), and then the capture is quite immediate. a small difference is only noticeable in very very fast moving objects.

The CRT monitor has the same time characteristics than a tube camera, but the CCD/CMOS cameras have more a time characteristic of a TFT monitor.

Matthias Buercher
http://www.belle-nuit.com



Jason White wrote :

>In the case of interlaced video both fields are not captured at the same >time (1/60th of a second apart in NTSC or 1080i) which is why we DO >get jagged edges in fast moving objects.

Is this still true when capturing 24 fps progressive (which is what I thought Steven was referring to) but converting to interlace?

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



>I ask because I thought progressive capture/display would mean that >the scan lines go sequentially, 1-2-3-4-5-6. . . to the limit of your >resolution.

Steve - keep in mind that if progressive scan on a CCD were line scanned similar to CRT line scanning, then we would see a scan completed from top to bottom in 1/24th of a second. It would be quite obvious on a pan or fast moving object as the bottom would be stretched behind the top. Ergo, the pixels must be exposing more or less simultaneously. (just looking for an excuse to use the word "ergo").

One thing I am curious about, however, is EXACTLY what the exposure time is when shooting 24frames per second on a Varicam or F900. I would imagine it's necessary to snatch the pixel data and download it, then reset the CCD for the next frame. How long does the CCD "reset" process take? Is it next to instantaneous? It's a hunch, but I feel like my Varicam with no shutter at 24fps is exposing for slightly less than 1/24 second.

By the way, isn't "progressive sca" some type of music?

Greg Gillam
Producer / Director
Red Sands Production Co.



>One thing I am curious about, however, is EXACTLY what the exposure >time is when shooting 24frames per second on a Varicam or F900.

On an F900, you can set the exposure time to 1/48, 1/32, 1/24 and also variable rates for scanning.

The CCD snatch and reset is quite immediately, I think less than a msec. We have seen some very little differences from top to bottom on a very fast motorcycle shot, but I would have to measure it.

However, the processing afterwards is quite long, and you have also audio sync problems with 1/24 at 24p, because the audio is not delayed accordingly.

Matthias Buercher