DV Info Net

DV Info Net (https://www.dvinfo.net/forum/)
-   JVC GY-HD Series Camera Systems (https://www.dvinfo.net/forum/jvc-gy-hd-series-camera-systems/)
-   -   A feature film and Documentary Shot with HD100U (https://www.dvinfo.net/forum/jvc-gy-hd-series-camera-systems/62487-feature-film-documentary-shot-hd100u.html)

Geoffrey Pepos March 9th, 2006 05:34 PM

A feature film and Documentary Shot with HD100U
 
Hello All,

In January we finished principal photography for our new feature film, "The Violent Kind." A thriller starring Kirk Harris (Hard Luck, Intoxicating) and John Savage (Deer Hunter, Hair, The Onion Field, Carnivale)

And last night I arrived back to LA from Montana where for two weeks I shot and directed a documentary about a couple who have done about 20,000+ sessions for people with disabilities over the past 20 years...with horse therapy.

Two distinctly different projects...yet they both featured horses...you know? The big animals. : )

The Camera was run over by a horse near the end of the feature shoot. It wasn't actually run over but the operator had to grab the camera and the tripod quick and broke off the handle. The shot is stunning by the way as the horse menaces toward the camera.

So, I modified the HD100U with some leather and glue and like it better without the handle because the focus/aperture/zoom are more accessible...I'm sure my warrenty is void at this point. However, a good cleaning and it still works perfectly.

Overall, combined projects, some 50+ hours of footage shot. And 97% captured at this point...and I watched it all on a Sony HD monitor as it was captured.

SSE? Had this problem only one time with the feature film...I was doing a POV shot in the woods at night with a head mounted LED flashlight and gain up to 9. Seems that SSE is an extreme example of not enough light.

Some focus issues, yet that was mainly because of the run and gun style. The focus assist works well. Yet, I'd love for it to only "exist" when touching the focus ring. Ended up shooting with that B/W with blue edges most of the time.

Some flicker from a 1.2K HMI on occasion when filling in daylight. Yet, that kind of lends to an organic look...plus a little custom plug-in would be easy to create to correct it in post if needed.

Used an NRG battery belt and that gave me about 4 hours of shooting time which we supplemented with half a dozen "regular" cam batteries. I prefer the weight of the batteries on my body rather than on the camera. Without the handle the camera balances better in my hands. But, that might just be me. I tend to shoot from the hip to the chest to keep the eyelines more filmic...rather than on the shoulder where the "actors" seem to be always looked down upon...a decreased presence on camera.

Some CA with wide open lens, sometimes, yet not enough to bother me because I'm looking at the stories and the decreased DOF is welcome. Besides, if you look at Kubrick's "The Shining" as the kid and mom are going through the labrinth in the daylight...um...I believe there is some CA in there on the edges of the hedge and the sky.

Both projects being posted with Adobe's Production Studio with Cineform's Prospect HD on a new Intel Dual Core box with 2TB and 2G of memory. (Also, much of the footage for the feature was captured direct to disk virtually uncompressed via the component outs with the Wafian HR-1 and the Cineform Codec.)

Comparing the DtoD and out of the cam? The DtoD is so clean that I'm going to have to add grain to mask my actor's facial imperfections. Not a bad problem to have. Out of the cam? Looks like a super 16 digital intermediate without the gate shudder. Is that the right word? (Except for some camera operator shudder yet even that looks more like film.) Did some green screen with the Wafian's green screen mode and it keys out very well and without much effort.

Is it all perfect? Nah...yet what is perfection? The lightness of the camera and the reduced needs for grip and crew allows me to focus on the important stuff...the actors or subjects and...mostly, the story. I've shot a lot of DV and even had a film in Sundance competition with DV. Yet, this is a new world, and one that allows, from cam to post, us indies to really compete with the bigger budgets...finally.

That's all for now...gotta cut two projects. At least all of the footage is living live and full res on the PC...my first "real" DigiFilm project was in 97 and I only had 4Gigs of storage.

Take Care,

-g

Daniel Weber March 9th, 2006 06:36 PM

Great post and info. Can you talk about your horse therapy doc any?

My wife has been doing this for around 10 years both in California and now on the East coast. She is now doing a pilot program with vets coming back from Iraq. I am looking for funding to shoot a doc on the program and am considering the HD100 to shoot it with. I have shot horse therapy stuff with everything from XL1, Betacam, DVX100 to Z1. I am not sure which format I want to go with right now. A long lens seems great when shooting stuff in an arena, but the small handheld is great when getting in close.

What was your shooting style like?

Dan Weber

Geoffrey Pepos March 9th, 2006 06:56 PM

Horse Doc
 
Dan,

Hey. I'd certainly recommend the HD100.

As things turn and with coincidences I've discovered...My feature is about a vet returning from Iraq to his father who was in Vietnam. I'm sure that if they both had horse therapy in the story...it would no longer be a thriller and we wouldn't have the dramatic conflict needed for a film. This is off topic to this forum, I'm sure.

Please contact me directly. We have much to talk about...and maybe some collaboration potential. I believe I've got my forum settings correct for this. Please, post a reply if I don't.

Yet, will relate the "handheld" experience for the forum.

I've shot a lot of doc work...starting with the Sony VX1000, to the XL1, PD150/70, DV500, HD10U, etc...usually with my short shotgun mike on camera. The cams change, the mike doesn't. This time I hired a sound op for the first week. The "kids" on the horses would stare at him as he pointed that thing at them. They called him the antenna. Since I mainly "look down" into my camera...they weren't so concerned about the camera guy.

The second week...was more natural...mainly because I rarely looked at the subjects...with my eyes...and did my best to seem that I was just hanging out...positioning the cam away from my body...rarely on the shoulder unless they couldn't see me...looking away from where I was shooting...wearing all black or earth tones doesn't hurt either. This didn't work for the blind girl...I think she could smell me. Yet, the therapist did make sure that they all knew I was present. And that it was okay and we were being respectful in shooting them. With her, I played a game of clapping so that she knew where I was...as in quantum mechanics...the act of observation affects the observed knowingly with results resembling a collaboration.

Shooting style? Quiet. Small movements...or large movements swiftly quiet. Usually fairly close to the subjet...full wide to 100 to 135mm lens in 35mm still camera equivalent...to avoid long lens voyeurism. Watch the "subject's" body language...move away when appropriate...come in closer when granted permission...gain trust...disappear. I have horses of my own so I have experience shooting horses up close...and on top of them too. A key thing is to let them smell the camera. Horses like to smell things for information. Horses are flight animals and they're really mainly concerned with three or four things...are you part of their herd and are you alpha? (The answer should always be yes)...is it something to eat? Is it something that's gonna eat them? Those two answers should be no. :}

There are other reasons I tend to dissapear (and appear on occasion) for the subjects...and that is a lengthy topic related to camera operator body language and movement which is key in capturing "documentary" footage...and also relates to my style with actors in a story for a feature film.

The JVC cam is just the right size...especially with my accidental "modifications."

The little cameras are too small for steady shots...the "broadcast" boxes are too big for subtle and intuitive and stealthy quick movement...the HD100? I think this size should be the standard for all future cameras no matter how much resolution they cram into them as the tech improves...it's about the same size as the body of an Alto Sax...if it worked for Charlie Parker...

-g

Mike Marriage March 10th, 2006 04:23 AM

Brilliant posts Geoffrey!

I particularly liked this quote, never applied the theory to documentary before:

Quote:

Originally Posted by Geoffrey Pepos
...as in quantum mechanics...the act of observation affects the observed knowingly with results resembling a collaboration.

It is so important in all filmmaking that you subjects feel comfortable, whether it's docu, drama or wedding videos.

How was the HD100 in low light? I am currently torn between it and spending more on HDCAM (probably 750) for a feature, a lot of which is low light. Obviously the Sony is the superior camera, but like you said, the smaller camera is more flexible and can lead to better content. It would also free up some of our limited budget.

Tom Chaney March 10th, 2006 04:36 AM

Geoffrey,

Thanks for the great post. We are starting a film this summer and have already picked up the camera. We are currently doing tests to try and find the right "technicolor look."

Tom Chaney

www.take2themovie.com

Jim Giberti March 10th, 2006 10:44 AM

Good stuff Geoffrey.
In addition to our creative company we also own a horse farm (where we recently moved our new studios on 200 acres across the road).
We have an autistic boy who comes up and our trainer works with him on one of our older dressage horses.
20,000 is an amazing commitment, also nice to hear yet another (Daniel) member involved with it.
When my life slows down I'd like to devote some real time to it.
I'm looking orward to the coming spring to start getting some green HD100 footage.
Everything is pretty white on the mountain.
We did shoot some great 1/2 pipe footage yesterday with a lot of the Olympic kids training for this weekends last World Cup.
I've got to post some shots..the colors are killer.

Geoffrey Pepos March 10th, 2006 12:32 PM

Therapy & Low Light
 
Thanks Guys,

It's great to find people on the same wavelength...especially since I currently live near Holloworld.

My "studio" is a cabin on an acre. My quarter horse is my editor.

THERAPY: The second I entered their barn, I knew immediately something amazing was happening on that horse. I know horses, have two of my own...yet, at first it seemed like alchemy. (They have twelve horses for the program) I've got my work cut out for me to tell their story with respect and fullness...they (Bob & Timi) figure they've walked about a mile or two for each session...I was there for about 65 miles out of 20K-40K!!! miles.

We mostly used existing light in the barn...florescent (ug) and two big windows. Snow outside was our reflector. White balanced to daylight...let the windows blow out...yet with the cine setting on the cam, it looks quite good. I think they had ultra cool florescents or something. Once in a while mixed in a tungten fresnel bounced off of a gold reflector to warm it up. Of course, the first time a horse saw that he had to check it out..."Hey, how come the sun is setting in here?"

LOW LIGHT: Low light and the HD100? I pushed it a few times during the feature...a couple of shots just after magic hour and right before dark and not enough time to set up the small HMI we had. Wide open lens (No gain up) and trees against sky...saw a bit of CA on the edges of the branches (I believe the CA is part lens, part low light on CCD)...I'll just drain the color in post to get rid of the magenta/green fringe...that's how our eyes see night anyway, in B/W...the rods in our eyes...our luminance channel?

I'd say, overall...the HD100 is not as sensitive as let's say a PD150/70 in low light. Yet, that could be misleading because the image is much cleaner and the contrast much better with the HD100. And perception is more important than specs. And the progressive HD image holds up better with post manipulation.

We did a lot of night shooting. Didn't have enough instruments to "Hollywood" the whole set...so, took advantage of pools of light and dark. And lighting planes for depth.

One shot was a woman getting flashed with headlights. It's a pretty stunning shot because it's all dark and when the light flashes we see her framed screen right and in the distance a gray Arabian horse. The lighting? Highbeams on a Suburban.

So, I guess it's all relative...if there is any image at all in whatever low light...it could be used to tell a story. I think of that scene in Godfather Two...Michael and his Mother? That is a dark scene.

I did light tree branches in the woods with an LED headmounted flashlight to great effect. Yet, I'm not making a bright sitcom...it's a thriller.

One thing as far as camera setting...I never stretched the blacks in the setup as some people have suggested. So when it was black in the image...there was less chance of noise in the blacks...the codec doesn't have to compress the noise info and has more bits for the important highlights and upper mid tones.

And, sometimes, turned the detail off for the same reason...usually, kept the detail at min...for some of it it was at -3...which I use for my "documentary" look...somewhere in between the edge enhanced "reality" of video and the "mystery" of film. Also at -3 when I know I'm going to add a fog or "net" filter in post.

HDCAM: Oh, and also in reference to the Sony HDCam. Sony was donating one for us to use...yet, I decided against it and bought the JVC instead. I'd be a lot more timid operating a 100K cam than a 6K cam...plus the extra grip and crew...I like to shoot with a tiny crew...not only for lower cost...but, for intimacy on the set.

For my eyes, the HD100U Component out into the Wafian HR-1 rivals the HDCam because of the Cineform Codec compared to the HDCam Codec...especially for critical color needs such as green screen...yet that all might be equalized because the HDCam is at 1080 rather than 720. Now a Panavized Sony950 SDI 1080 24p out into the Wafian would be ideal if you have the budget, crew and grip support and are shooting more traditionally. Great image quality, no decks needed, and the footage is ready to edit/review right after each take...hook up a network and an editor could be in the back cutting the movie as it's being shot....or green screen stage, can check the composites at full res with the exact footage. Or, or...

It all gets so overwhelming sometimes doesn't it? That's why my favorite pieces of "gear" are the story, the script and the actors.

-g

John Trent March 11th, 2006 01:00 PM

Geoffrey,

Best posts I've read in quite awhile. Hope to see your work. I like your approach and philosophy.

I'm almost ready to buy a HD100 but I still have a few fears, maybe you could help. You sound like a man that can give a objective opinion and you have plenty of experience with this camera.

HDV ARTIFACTS: I'm mostly concerned with the image going soft on pans, giving it a sort of artificial motion blur effect, but I'd like to know of any problems you encountered. One fellow said the camera loses half it's resolution during camera movements (web misinformation alarmist, I hope).

NOISE: I've seen some noisy web footage...maybe it was web compresssion and/or stretched blacks. I've read and re-read what you had to say about low light performance and minimizing noise but was it still an isssue?

DVX COLOR & GAMMA: You didn't mention ever working with this camera. If you have, would you say you could match the Panny's color and gamma with the JVC?

I know I'm asking a lot...I'm kind of desperate for information. That's for taking the time to respond. I appreciate it.

John

Geoffrey Pepos March 11th, 2006 07:17 PM

Hey John, I'll try my best here.

Quote:

Originally Posted by John Trent
HDV ARTIFACTS: I'm mostly concerned with the image going soft on pans, giving it a sort of artificial motion blur effect, but I'd like to know of any problems you encountered. One fellow said the camera loses half it's resolution during camera movements (web misinformation alarmist, I hope).

I believe there is a motion sensor in the camera that detects pan rate and adjusts and lowers resolution accordingly....um...only if you don't wear foil on your head. : )

Seriously, the motion blur effect is real and a primary "signature" in what gives 24p a film look. In a blurry pan, the HDV (mpeg) codec (if I'm not mistaken), has to work less because there is less edge detail and more areas of flat tone. I'm suspecting that the "alarmist" is more concerend with "motion" judder and has mainly shot at 60i in the past. Probably isn't used to how 24p works perceptually. 60i is very forgiving for camera movement and objects moving within the image in front of the camera.

Better than me to explain...an article from JVC Europe:

http://www.jvcpro.co.uk/getResource2...ar.pdf?id=6118

Film DP's have perfected pan and motion rates over the last century and they are described in exact detail in the ASC manual. There is a reason pro tripods have highly tuned friction controls...they even have specified car speed limits depending on if you're shooting straight in front or to the side, etc.

I'd recommend getting it if only to be overwhelmed and intimidated by the math. : ) However, the "numbers" can easily be learned by "the body" by practicing with a 24p camera. This is liberating for us indies...tape is cheap.

Also, I've found that experienced film actors intuitively move at the proper rate...John Savage, as one example from my film...his movement is ever so slightly slower than normal movement. This probably comes from working on so many film projects and knowing where the camera is and where it's pointing at and how fast it's moving, at all times...how he does this while also being in the moment as a character never ceases to amaze me.

For hand held, after a while, a camera operator just "knows" with his/her body how to move the camera and follow the action within a sceen. It's just a matter of, lots of, practice.

Quote:

Originally Posted by John Trent
NOISE: I've seen some noisy web footage...maybe it was web compresssion and/or stretched blacks. I've read and re-read what you had to say about low light performance and minimizing noise but was it still an isssue?

Out of the Wafian box? Almost no noise. It went in there from the component out of the camera and is 10bit 4:2:2. Out of the camera after being compressed to (8 bit, 4:2:0) mpeg...yeah...some noise...Depending on the complexity of the picture. Objectionable? Not for me. There is also more grain/noise if there simply isn't enough light for the CCD...whether you boost gain or not. Yet, I like noise (grain/artifacts) to some extent...and the mpeg "noise" out of the camera has a much better quality/look than DV noise. Also, when shooting at 30p rather than 24p? There seems to be, to my eye, more "noise" because there are 6 more frames to compress with the same number of bits.

[SOAP BOX] The web footage? I'd have to ask. "What was it cut on?" If it was cut in "native" form then any amount of manipulation such as color correction will adversly affect the the image quality. (Mpeg was made for delivery and is not best for editing no matter what people say.) Same with DV. "Native" editing is an effective marketing term at best...just like "uncompressed" is now for HD especially in the high-end shops (they get to charge more)...as it was for SD a few years ago. The best codec to edit with is one that is made for editing, has more color depth than the native...and one that can be proven to be as good visually as uncompressed. [/SOAPBOX]

For my feature film, if the light was on the edge of not enough, I "Michael Manned" it by using a 24 shutter speed and adjusted (slowed) my movement rate to mitigate the extra motion blur. For another shot, I boosted the gain to 9 with 48 shutter...there is extra grain/noise...yet, the scene was an edgy scene...so the extra noise makes sense for that part of the movie. Gain up of 3 or 6 are quite clean relatively.

I just shot a band in Hollywood last night...and there was almost no light on the stage and it was all red light! I pushed the gain up to 18...and there is significant noise as a result. Yet it works for the vibe of the band. I did post process it to reduce the noise in the shadows and the low mid tones. It looks like a fast (grainy) film stock.

There are many ways in post to minimize grain/noise as an alternative to the processor intensive "remove" grain plugin in the AE pro package. I use After Effects with stock filters in various configurations on layer stacks. To much to get into here.

Quote:

Originally Posted by John Trent
DVX COLOR & GAMMA: You didn't mention ever working with this camera. If you have, would you say you could match the Panny's color and gamma with the JVC?

I haven't myself. I know there are others on this forum who have created DVX settings files for just that thing.

I hope this helps a little.

Eventually, I'll post some footage...better to illustrate for your visual perception devices...um...eyes.

-g

Stephen L. Noe March 11th, 2006 08:31 PM

Quote:

Originally Posted by Geoffrey Pepos
[SOAP BOX] The web footage? I'd have to ask. "What was it cut on?" If it was cut in "native" form then any amount of manipulation such as color correction will adversly affect the the image quality. (Mpeg was made for delivery and is not best for editing no matter what people say.) Same with DV. "Native" editing is an effective marketing term at best...just like "uncompressed" is now for HD especially in the high-end shops (they get to charge more)...as it was for SD a few years ago. The best codec to edit with is one that is made for editing, has more color depth than the native...and one that can be proven to be as good visually as uncompressed. [/SOAPBOX]

Being on the "native" side of the fence and doing CC for years I'm not so sure about your comments. The comment may be more relevant to what render/fuse codec was used to render the end CC results.

I offer the Chicago ProHD user group meeting (3rd Monday each month, Downtown Chicago). I can demonstrate both the "native" ProHD film workflow as well as the FCP Aja Kona capture workflow side by side and we can take a look at the results.

Geoffrey Pepos March 11th, 2006 09:28 PM

Quote:

Originally Posted by Stephen L. Noe
Being on the "native" side of the fence and doing CC for years I'm not so sure about your comments. The comment may be more relevant to what render/fuse codec was used to render the end CC results.

I offer the Chicago ProHD user group meeting (3rd Monday each month, Downtown Chicago). I can demonstrate both the "native" ProHD film workflow as well as the FCP Aja Kona capture workflow side by side and we can take a look at the results.

Yes, I agree Stephen. The end render is a key and must be of the highest quality. Yet, we do have to think about time to decompress the "native" for our daily editing/compositing work...or if we're bringing in uncompressed...one processor intensive...the other data size intensive. Both of these things take time.

What if there was a codec that was more efficient in decompression time (and compression time for the end result) and was also data thrifty AND had as good or better visual quality?

We could then edit/compose quicker and also operate on more streams at once...and idealy, in real time.

My co-composer lives in Chicago and I've been meaning to visit him. I'll keep a third monday open.

-g

Stephen L. Noe March 11th, 2006 09:53 PM

Quote:

Originally Posted by Geoffrey Pepos
Yes, I agree Stephen. The end render is key. Yet we do have to think about time to decompress the "native" for our daily editing/compositing work...or if we're bringing in uncompressed via AJA...one processor intensive...the other data size intensive. Both of these things take time.

What if there were a codec that was more efficient in decompression time and also data thrifty AND had as good or better visual quality? We could then operate on more streams at once in real time?

My co-composer lives in Chicago and I've been meaning to visit him. I'll keep a third monday open.

-g

Sounds good on getting together. The uncompressed is the least processor intensive because there is no calculation on the frames but it takes up whopping space. The "native" workflow is much better in that the untouched clips remain untouched all the way to delivery format and the CC'd clips are rendered uncompressed 2VUY and subpixel. Once "everybody" is happy with the timeline then it's sent out as uncompressed TGA sequence with audio channels delivered as elementary streams. Efficiency and quality are very good.

We'll be happy to see you when you get here...


All times are GMT -6. The time now is 12:54 AM.

DV Info Net -- Real Names, Real People, Real Info!
1998-2024 The Digital Video Information Network