First Attempt at 35mm filmout with the HD-100 at DVinfo.net

Stephen L. Noe · November 10th, 2005, 04:53 PM

Hi,

Anyone else trying a filmout yet?

My workflow: I captured from the HD-100 with CapDVHS and then processed the files through HDTVtoMPeg2, then imported the 720p24 files into a Liquid 6.1 rack. Edited them with the default 720p24 preset and then once the editing was done, I "exported as" Quicktime Tiff uncompressed to a firewire hard drive for delivery to the post house. The camera was set to a filmout curve and the colors were very close to linear. The purpose is for a sales demonstration of the HD-100. We'll see how it all goes.

Luis Otero · November 10th, 2005, 06:09 PM

I do not recall the name, but there was a presentation of someone that made a dacumentary for Nat. Geo. and presented part of the footage transerred into film in NY last month. You may want to do some seach within the DVInfo forum to get a hold on the person.

Luis

Luis Otero · November 10th, 2005, 06:10 PM

Actually his name is Andrew Young and his email is andy@ArchipelagoFilms.com
Just found it...

Good luck,

Luis

Steve Mullen · November 10th, 2005, 08:15 PM

Quote:

Originally Posted by Stephen L. Noe

The camera was set to a filmout curve and the colors were very close to linear.

When shooting, what Matrix and Gamma did you use?

Stephen L. Noe · November 10th, 2005, 08:45 PM

Quote:

Originally Posted by Steve Mullen

When shooting, what Matrix and Gamma did you use?

T. Dashwoods "Filmout" recipe.

Click here

Andrew Young · November 10th, 2005, 11:42 PM

I output about 15 minutes of HD100 footage to 35mm for a JVC presentation in New York last month. 8 minutes of it was ducumentary footage I shot of a scientific expedition in Madagascar - very rough and uncontrollable conditions. For variety, I added 4 minutes of a lit studio set-up I did here in New York and in addition, I shot out Charles Pappert's mini35 test, which is posted on this site. The filmout was done by DuArt in New York. I've been meaning to report on the results here but have been very busy editing the Madagascar material for a broadcast deadline. The presentation and I should be coming to LA in the near future.

In a nutshell, the results are good. My Settings were: factory defaults with cinelike: ON. My advice: turn detail down (but not off). I only received the camera a few days before travel, so I did not have time to evaluate other settings, but I have been told by JVC that the "filmout" gamma setting is a straight line with a slope of 1 - essentially no gamma. I would not recommend using this setting without serious testing and a monitoring device with proper gamma correction - otherwise it will be very hard to know what your getting.

Tim Dashwood · November 11th, 2005, 05:59 AM

Yes. Filmout requires the use of a "Gamma Box" or monitor that can display film gamma, or else the image from the camera will appear about 2-stops too dark.

The idea for the film gamma curve is to retain contrast and latitude (and no digital re-compression) when creating a filmout for traditional lab timing. Of course, this would mean you should be exposing your highlights at around 105% IRE and letting the dark stuff ride. The gamma curve of your selected film stocks will do the rest of the work.
If you have the luxury of performing a D.I., then I would suggest shooting with my wide latitude scene file and crush your blacks in the digital realm.

Stephen L. Noe · November 11th, 2005, 06:50 AM

Tim, we set the highlights to 108 and let the blacks ride. The video is definately not too dark and If I kept it in the video realm I'd definately pull the blacks down and adjust the gamma to where I know DV likes to live. We'll see how it shakes out...

Click here for screengrab.

Andrew Young · November 11th, 2005, 07:46 AM

Stephen,

I'll be very interested to hear how filmout gamma works for you, both in the digital realm and in filmout. I hope to be dooing some test of my own down the pike...

Ben Syverson · November 11th, 2005, 01:24 PM

I haven't played with an HD-100 yet, but if it does indeed offer a truly linear image with no gamma (it'll look extremely dark, with electric highlights), it would be wise to avoid it at all costs.

The reason is that fully linear (gamma 1.0) images are horribly inefficient. Most of the data is taken up by highlights, with the midtones and shadows all crushed together. This is fine if you're working in 16bit or floating point, but at 8bit, the images would be essentially unusable, especially once compressed to HDV. Think banding and noise galore.

This is because in linear space, twice the value equals twice the light -- so the stops you can represent are:
255
127
63
31
15
7
3
1

You'll notice that the the brightest stop takes up half the image data! The next stop takes up 1/4 of the image data, the next 1/8th, etc, on down the line. The bottom few stops are unusable because of noise and processing, so you're left with only two or three usable stops.

For more info, check out my page about sensor values and latitude

The best option for the HD100 is probably to shoot at a gamma around 1.8 - 2.5, underexposing by a stop or so....

Nate Weaver · November 11th, 2005, 01:44 PM

Ben, that page is tremendous.

I like to think I understand a bit about this stuff, but you're obviously way ahead. After carefully parsing the page, things made sense.

I can't remember once ever seeing specs on primary A/D bitdepth on the HD100, or DSP bitdepth. I'm sure it's been covered, but what are the specs? Anybody know?

Something implied by your page is that latitude in current cameras is at least limited as much by A/D bitdepth as the properties of the CCD itself. This would explain why the Reelstream Andromeda rig appears to pull more latitude than DV output off the DVX. Just how much latitude is being thrown away by A/D compromises?

Stephen L. Noe · November 11th, 2005, 03:58 PM

Ben,

I read your page and have been using the same techniques in DV for years. Excellent page, BTW. A film print is a different animal to me and getting video ready for film print is a different animal for me as well. The BMP (above) is an uncompressed frame right out of the shot footage. Can you download it and give an assessment?

As I wrote above, the brightest highlight is 108 and then let the blacks ride using the filmout gamma.

Marty Baggen · November 11th, 2005, 04:16 PM

The only filmout I have done was to 35mm and was a digital edit of material telecined from Super 16mm, so this may be a truly ignorant question:

If you are going out to film, why set up the acquisition unit for any sort of filmic gamma? Shouldn't that be dictated in the actual filmout process? Put another way, why imitate film if you are going to film?

I guess that begs other questions that I hadn't considered, such as... what considerations should be made in film stock and general handling of the video itself during the transfer?

It would seem to me that the job of the videographer would be to provide the least manipulated, most neutral image possible, and allow the colorist and film technician to do the rest.

This could be way off base.... as I say, my only experience in this was with stuff that was already shot on film.

Joseph H. Moore · November 11th, 2005, 04:56 PM

Marty,
I think you are basically right, assuming that DI color work is going to be done. In that case you want to capture as much information as possible (contrast be damned) and then grade the color on a monitor that is reproducing the gamma and color space (i.e. a LUT) of the film stock. That's the best case scenario.

Steve Mullen · November 11th, 2005, 05:39 PM

Quote:

Originally Posted by Marty Baggen

If you are going out to film, why set up the acquisition unit for any sort of filmic gamma? Shouldn't that be dictated in the actual filmout process? Put another way, why imitate film if you are going to film?

FILMOUT is a linear video gamma that is used during recording. It is NOT a film look gamma. The output looks nothing like film or video!

It is, however, correct for a film recorder.

You can only view on monitors (Panasonic) that have a "Linear Mode" capability.

November 10th, 2005, 04:53 PM	#1
Stephen L. Noe Trustee Join Date: Apr 2005 Location: Chicago Posts: 1,214	First Attempt at 35mm filmout with the HD-100 Hi, Anyone else trying a filmout yet? My workflow: I captured from the HD-100 with CapDVHS and then processed the files through HDTVtoMPeg2, then imported the 720p24 files into a Liquid 6.1 rack. Edited them with the default 720p24 preset and then once the editing was done, I "exported as" Quicktime Tiff uncompressed to a firewire hard drive for delivery to the post house. The camera was set to a filmout curve and the colors were very close to linear. The purpose is for a sales demonstration of the HD-100. We'll see how it all goes.

November 10th, 2005, 11:42 PM	#6
Andrew Young Regular Crew Join Date: Aug 2005 Location: New York Posts: 123	HD100 to 35mm filmout I output about 15 minutes of HD100 footage to 35mm for a JVC presentation in New York last month. 8 minutes of it was ducumentary footage I shot of a scientific expedition in Madagascar - very rough and uncontrollable conditions. For variety, I added 4 minutes of a lit studio set-up I did here in New York and in addition, I shot out Charles Pappert's mini35 test, which is posted on this site. The filmout was done by DuArt in New York. I've been meaning to report on the results here but have been very busy editing the Madagascar material for a broadcast deadline. The presentation and I should be coming to LA in the near future. In a nutshell, the results are good. My Settings were: factory defaults with cinelike: ON. My advice: turn detail down (but not off). I only received the camera a few days before travel, so I did not have time to evaluate other settings, but I have been told by JVC that the "filmout" gamma setting is a straight line with a slope of 1 - essentially no gamma. I would not recommend using this setting without serious testing and a monitoring device with proper gamma correction - otherwise it will be very hard to know what your getting.

November 11th, 2005, 01:24 PM	#10
Ben Syverson Major Player Join Date: Mar 2004 Location: chicago Posts: 434	Shooting with Gamma 1.0 I haven't played with an HD-100 yet, but if it does indeed offer a truly linear image with no gamma (it'll look extremely dark, with electric highlights), it would be wise to avoid it at all costs. The reason is that fully linear (gamma 1.0) images are horribly inefficient. Most of the data is taken up by highlights, with the midtones and shadows all crushed together. This is fine if you're working in 16bit or floating point, but at 8bit, the images would be essentially unusable, especially once compressed to HDV. Think banding and noise galore. This is because in linear space, twice the value equals twice the light -- so the stops you can represent are: 255 127 63 31 15 7 3 1 You'll notice that the the brightest stop takes up half the image data! The next stop takes up 1/4 of the image data, the next 1/8th, etc, on down the line. The bottom few stops are unusable because of noise and processing, so you're left with only two or three usable stops. For more info, check out my page about sensor values and latitude The best option for the HD100 is probably to shoot at a gamma around 1.8 - 2.5, underexposing by a stop or so....

November 11th, 2005, 01:44 PM	#11
Nate Weaver Wrangler Join Date: Oct 2002 Location: Los Angeles, CA Posts: 2,100	Ben, that page is tremendous. I like to think I understand a bit about this stuff, but you're obviously way ahead. After carefully parsing the page, things made sense. I can't remember once ever seeing specs on primary A/D bitdepth on the HD100, or DSP bitdepth. I'm sure it's been covered, but what are the specs? Anybody know? Something implied by your page is that latitude in current cameras is at least limited as much by A/D bitdepth as the properties of the CCD itself. This would explain why the Reelstream Andromeda rig appears to pull more latitude than DV output off the DVX. Just how much latitude is being thrown away by A/D compromises? __________________ My Work: nateweaver.net

November 11th, 2005, 03:58 PM	#12
Stephen L. Noe Trustee Join Date: Apr 2005 Location: Chicago Posts: 1,214	Ben, I read your page and have been using the same techniques in DV for years. Excellent page, BTW. A film print is a different animal to me and getting video ready for film print is a different animal for me as well. The BMP (above) is an uncompressed frame right out of the shot footage. Can you download it and give an assessment? As I wrote above, the brightest highlight is 108 and then let the blacks ride using the filmout gamma. Last edited by Stephen L. Noe; November 11th, 2005 at 04:51 PM.

November 10th, 2005, 06:09 PM	#2
Luis Otero Major Player Join Date: Apr 2005 Location: Fort Dodge, IA Posts: 221	I do not recall the name, but there was a presentation of someone that made a dacumentary for Nat. Geo. and presented part of the footage transerred into film in NY last month. You may want to do some seach within the DVInfo forum to get a hold on the person. Luis

November 10th, 2005, 06:10 PM	#3
Luis Otero Major Player Join Date: Apr 2005 Location: Fort Dodge, IA Posts: 221	Actually his name is Andrew Young and his email is andy@ArchipelagoFilms.com Just found it... Good luck, Luis

November 11th, 2005, 05:59 AM	#7
Tim Dashwood Wrangler Join Date: Aug 2005 Location: Toronto, ON, Canada Posts: 3,637	Yes. Filmout requires the use of a "Gamma Box" or monitor that can display film gamma, or else the image from the camera will appear about 2-stops too dark. The idea for the film gamma curve is to retain contrast and latitude (and no digital re-compression) when creating a filmout for traditional lab timing. Of course, this would mean you should be exposing your highlights at around 105% IRE and letting the dark stuff ride. The gamma curve of your selected film stocks will do the rest of the work. If you have the luxury of performing a D.I., then I would suggest shooting with my wide latitude scene file and crush your blacks in the digital realm.

November 11th, 2005, 06:50 AM	#8
Stephen L. Noe Trustee Join Date: Apr 2005 Location: Chicago Posts: 1,214	Tim, we set the highlights to 108 and let the blacks ride. The video is definately not too dark and If I kept it in the video realm I'd definately pull the blacks down and adjust the gamma to where I know DV likes to live. We'll see how it shakes out... Click here for screengrab.

November 11th, 2005, 07:46 AM	#9
Andrew Young Regular Crew Join Date: Aug 2005 Location: New York Posts: 123	Stephen, I'll be very interested to hear how filmout gamma works for you, both in the digital realm and in filmout. I hope to be dooing some test of my own down the pike...

November 11th, 2005, 04:16 PM	#13
Marty Baggen Major Player Join Date: May 2005 Location: monroe, or Posts: 572	The only filmout I have done was to 35mm and was a digital edit of material telecined from Super 16mm, so this may be a truly ignorant question: If you are going out to film, why set up the acquisition unit for any sort of filmic gamma? Shouldn't that be dictated in the actual filmout process? Put another way, why imitate film if you are going to film? I guess that begs other questions that I hadn't considered, such as... what considerations should be made in film stock and general handling of the video itself during the transfer? It would seem to me that the job of the videographer would be to provide the least manipulated, most neutral image possible, and allow the colorist and film technician to do the rest. This could be way off base.... as I say, my only experience in this was with stuff that was already shot on film.

November 11th, 2005, 04:56 PM	#14
Joseph H. Moore Major Player Join Date: Oct 2005 Posts: 362	Marty, I think you are basically right, assuming that DI color work is going to be done. In that case you want to capture as much information as possible (contrast be damned) and then grade the color on a monitor that is reproducing the gamma and color space (i.e. a LUT) of the film stock. That's the best case scenario.