HD Progress???

Mike Panczenko · June 18, 2004

It seems to me that when new cameras are announced, the companies are touting the resolution. To me, that is not the main issue in what gives video a video look. The latitude seems to be the largest difference between video and film, yet I rarely hear companies addressing that. I know that they are working on it, and they tout the resolution because it is a good marketing strategy, but I haven't noticed much progress in the latitude of HD. I do not doubt that the people are working their hardest on it, but why is it taking so long? What makes the latitude more difficult than other issues? How much longer do you think it will take until HD can compete with film on that level. Comparing the latitudes, HD and miniDV and the others do not seem to have that much of a difference. I'd like to get your predictions on when video latitude and gamma will compete with film, and what makes it such a difficult goal to achieve.

John Pytlak RIP · June 19, 2004

The latitude seems to be the largest difference between video and film, yet I rarely hear companies addressing that.

Kodak has addressed the issue, presenting and publishing several SMPTE technical papers analyzing the differences:

Assessing the Quality of Motion Picture Cinema Systems, Christopher L. DuMont, Roger A. Morton, and Kenneth J. Repich, SMPTE Journal, February/March 2002

The "Cinematographers Test" clearly showed the difference in latitude, especially in the highlights (rendition of subtle colors and flesh tones was also quite different):

http://www.kodak.com/country/US/en/motion/...h/35hd24p.shtml

Many cinematographers have commented on it in on-line discussion groups like this.

Phil Rhodes · June 19, 2004

Hi,

I think you'll find that people tout the resolution because barring some fairly major technological advances, there's stuff-all they can do about the dynamic range. It may also be now that because this has been the situation for a while, it's probably starting to feel really difficult for them to advertise based on any other characteristic, but I think it's obvious that they shout about what they can, and overlook what they can't.

That said, CMOS imagers are far from mature. I also wonder what happened to the idea of having two sensors per pixel with differing sensitivities - there have been little consumer pocket cameras with CCDs using the technique, but I think that perhaps it's fallen by the wayside for some reason since I haven't heard much about it. Perhaps Mr. Pytlak can ask the Kodak CCD boffs.

Phil

Justin Barham · June 19, 2004

...barring some fairly major technological advances, there's stuff-all they can do about the dynamic range.

Is that really true? I mean, is there some technical reason why video will never have the dynamic range of film?

John Pytlak RIP · June 19, 2004

Lots coming out of the Kodak Research Labs, so I never say never:

http://www.kodak.com/US/en/corp/researchDe...LRCameras.shtml

http://www.kodak.com/US/en/corp/researchDe...ectTB2003.shtml

http://www.kodak.com/global/en/digital/ccd/sensorsMain.jhtml

http://www.kodak.com/global/en/digital/ccd...sArticles.jhtml

Mike Brennan · June 19, 2004

"It seems to me that when new cameras are announced, the companies are touting the resolution."

I'm not so sure about that, the latest Genisis camera is "touting" use of 35mm lenses as well as improved resolution. They are saying it has a stop more range than f900. (not disputed so far in early tests by Alan Daviau)

The f950 *wasn't* touted as having more resolution than f900 but it has! (when one records 444)

The electronics marketing guys respond to whatever they think will sell cameras, rather than imparting really salient information, where as Kodaks recent advertising is esoteric and philosophical. I wouldn't make any connection between marketing and the real world:)

"To me, that is not the main issue in what gives video a video look. The latitude seems to be the largest difference between video and film, yet I rarely hear companies addressing that."

It is not surprising that the term latitude isn't mentioned much as it isn't a video term.

But terms like dynamic contrast control, knee saturation, 444 all relate to the digital equivelant of latitude and are constantly mentioned if improvements are made.

"I know that they are working on it, and they tout the resolution because it is a good marketing strategy, but I haven't noticed much progress in the latitude of HD."

There is a least 1 1/2 stops betweeen first 10 bit HDCAM and current f950. The early responses to the Alan Daviau test would suggest that the Genesis has greater range than f950. Lets call it 1/2 stop. That makes 2 stops improvement in 6/7 years, perhaps this is as much as we can expect without a revolutionary sensor.

"What makes the latitude more difficult than other issues? How much longer do you think it will take until HD can compete with film on that level."

I'll bet that within a decade a sensor will exist that has seperate control of "exposure" of different levels within a image falling on a sensor, maybe a multi channel facility, for example so you can retain some colour in the bright sun on one channel as well as detail in rim lit clouds on another channel, without effecting the rest of the picture.

In theory this can be achieved today on a bench with a prism system. A mirror and a second sensor in the Genesis could create a second channel that could be recorded on one of the two SR HDSDI tracks for instance. But more processing more weight ect. Ikegami had a 4 sensor prism in the nineties. The forth sensor could be NDed to recieve 2 or 3 stops less light than the other sensors.

A 14 bit A to D would make a slight improvement over the current 12bit A-D in f900 and 10bit A-D in 750, Varicam and most other DV cameras.

"Comparing the latitudes, HD and miniDV and the others do not seem to have that much of a difference.

This is a subjective apprasal of HD vs DV:) The likely hood is that you were watching a crt? where well shot film, HD or DV may be difficult to tell apart.

In respect to distribution of images to cinemas, HD (transferred to film) *is* competing with contrast range of film. By this I mean that the quality of cinema prints falls to within the range of HD.

However it is a different story at the HD camera end, as one must expose correctly and control the subject lighting to fit the range of the sensor.

If there were a way to project all the quality and range that is in the neg it would blow current digital formats out of the water.

"I'd like to get your predictions on when video latitude and gamma will compete with film, and what makes it such a difficult goal to achieve."

Sony's 12 megapixel ccd sensor was a surprise, the Fuji dual pixel per photosite, is so far a dissapointment.

It took Kodak 100 years to get to this point with far more revenue over these years much from consumer use of film than broadcast TV camera manufacturers have earnt. The digital sensor revolution is as likely to come from product developments or spin offs from satisfying the mass markets, than from in house developers. Nikons new factory is making 60,000 pro digital cameras a month, perhaps the next big thing will evolve from the stills market?

Mike Brennan

Mike Brennan · June 19, 2004

"The "Cinematographers Test" clearly showed the difference in latitude, especially in the highlights (rendition of subtle colors and flesh tones was also quite different):

http://www.kodak.com/country/US/en/motion/...h/35hd24p.shtml

Many cinematographers have commented on it in on-line discussion groups like this"

Yes they certainly have!!

It has been reported that one of the cinematographers stated that he used the test to push HD to the limit of exposure. Although there may be some merit in the test, in that it potrays how HD can "fall over", the results should not necessarly be regarded as being indicative of how a DP, shooting a HD feature (and not a 'lets see how fast she'll go' test) would impliment the medium!

This film was withdrawn from IBC as apparently Kodak would not let it be shown without a Kodak representative in attendance. Originally it was titled "There is more to the story" a title which perhaps was too strong a pointer to its heritage? so it was later changed to "The cinematographers Test" possibly to make it appear more impartial?

"The Cinematographers Test" was made by Brava Films Production and was executive produced and financed by....

Mike Brennan

Mike Brennan · June 19, 2004

"The "Cinematographers Test" clearly showed the difference in latitude, especially in the highlights (rendition of subtle colors and flesh tones was also quite different):

http://www.kodak.com/country/US/en/motion/...h/35hd24p.shtml

Many cinematographers have commented on it in on-line discussion groups like this"

Yes they certainly have!!

It has been reported that one of the cinematographers stated that he used the test to push HD to the limit of exposure. Although there may be some merit in the test, in that it potrays how HD can "fall over", the results should not necessarly be regarded as being indicative of how a DP, shooting a HD feature (and not a 'lets see how fast she'll go' test) would impliment the medium!

This film was withdrawn from IBC as apparently Kodak would not let it be shown without a Kodak representative in attendance. Originally it was titled "There is more to the story" a title which perhaps was too strong a pointer to its heritage? so it was later changed to "The cinematographers Test" possibly to make it appear more impartial?

"The Cinematographers Test" was made by Brava Films Production and was executive produced and financed by....

Mike Brennan

Mike Brennan · June 19, 2004

"Is that really true? I mean, is there some technical reason why video will never have the dynamic range of film?"

Yes there are technical reasons why digital cameras do not have the dynamic range of neg film at the moment.

But there are no laws of phsyics, religous beliefs or acts of parliment to outlaw development of a sensor that is superior to film:)

Not "if" but "when".

That is not to say that such a sensor will replace film overnight.

Mike Brennan

John Pytlak RIP · June 20, 2004

Mike Brennan wrote:

It has been reported that one of the cinematographers stated that he used the test to push HD to the limit of exposure. Although there may be some merit in the test, in that it potrays how HD can "fall over", the results should not necessarly be regarded as being indicative of how a DP, shooting a HD feature (and not a 'lets see how fast she'll go' test) would impliment the medium!

The superior dynamic range of film puts fewer restrictions on what a cinematographer can do. You don't have to "work around" the issue of limited latitude and make as many compromises in your lighting.

Yes there are technical reasons why digital cameras do not have the dynamic range of neg film at the moment.

That's the bottom line. You are choosing the best way to make your production NOW, not ten years in the future.

It took Kodak 100 years to get to this point with far more revenue over these years much from consumer use of film than broadcast TV camera manufacturers have earnt.

Yes, Kodak has been constantly improving film for over a century. When you open each can, you are getting the latest technology. You don't have to buy a new camera to take advantage of the latest improvements like VISION2 technology.

The digital sensor revolution is as likely to come from product developments or spin offs from satisfying the mass markets, than from in house developers. Nikons new factory is making 60,000 pro digital cameras a month, perhaps the next big thing will evolve from the stills market?

Kodak makes its own sensors for its professional digital cameras, and is an OEM supplier to other camera manufacturers:

http://www.digitaloutput.net/back%20edit/aug03/feature3.html

Sizes of sensors vary greatly. Some pro medium and large format sensors go as high as 20 megapixels (MP). Kodak has built a line of 16MP sensors that are included in a number of leading medium format digital backs.

The highest resolution digital SLR, the Kodak 14n, has a 14MP sensor, with the Canon 1Ds being close behind at 11MP. The next level for professional SLRs is around 6MP, with a number of models from different manufacturers in that category.

Nikon is also going with a type of interpolation to increase the effective resolution of its D1X digital SLRs. Trying to compete with the 14MP Kodak 14n, which is equipped with a Nikon lens mount, and Canon, which has the 11MP EOS 1Ds, Nikon has developed enhanced in-camera firmware and related software to intelligently interpolate the 6MP images that the D1X captures up to 10MP. With both Fuji and Nikon, the resulting image quality is very good, but they?re still forms of interpolation.

Jason Rodriguez · June 20, 2004

The Genesis only has a stop over the F900?

That's not much then, and wouldn't coincide with the comments about it looking like 35mm, or even what somebody in the Genesis thread described when he aimed the camera at a crowd of people and it captured both the shadows and the hot-lit sunshine falling on the crowd.

One stop over the F900 is only around 9-10 stops of lattitude. Not much of an improvement IMHO. The Dalsa has 12-stops of dynamic range according to their engineers.

Mike Brennan · June 20, 2004

"That's not much then, and wouldn't coincide with the comments about it looking like 35mm, or even what somebody in the Genesis thread described when he aimed the camera at a crowd of people and it captured both the shadows and the hot-lit sunshine falling on the crowd."

Hard to comment on the shadows sunlight observation apart from saying current HD doesn't exaclty fall over in such situations however a stop in the highlights on video can be the difference between no highlight detail at all and a block of detail popping into the zone, especially true with back Caucasian lit skin tone.

Properly/sympathetically handled HD that is intercut with film is very difficult to spot. Direct split screen is more obvious.

I'd guess that the Genesis test was produced to get the best out of the camera, not as a demonstration of the difference between film and HD. I'm sure we could conjure a test where film is more obvious. A High Key cosmetics commercial would do it!

The Genisis is a ccd camera, I wouldn't expect miricles, but do expect more positive comments from the community as DPs and others warm to a camera that has 35mm lenses.

From what I've heard from the technical side, (I haven't seen the test) it is likely that intercutting between Viper/ f950 with 444 recording and Genesis will be possible as there is likely to be not more than a stop in difference in the latitude to tape. A slight difference in sharpness should be evident, but be difficult to spot except perhaps in cityscapes and landscapes or other scenes that are rich in fine detail. Will be interesting to see, but don't expect Panavision to allow a side by side comparison unless the Genesis is actually much better than 2/3 inch cameras:)

Sony produced a promo shot f950 444 which was projected at the Digital Cinema Summit in Vegas on the 2k projector but was largely rubbished by the self proclaimed experts.

A month later the very same experts claim the Genesis 1920x1080 SR recording is a close match to film! I understand that digital noise reduction was applied to the film shot version of the Genesis test so it probably looked less like grainy film than purists would be happy with.

Earlier this month I was shooting a wedding scene for a commercial on HD in Africa, Nigerian bride wearing an off white dress. There was also a day exterior of a white car, Nigerian occupants. The difference between lighting this for film and HD was that it simply required more fill for HD and I added a polariser.

If you examine the film vs HD test conducted over the years a common key aspect has been the exploration of the exposure latitude of HD vs film. So typically, DPs would light a scene with the contrast ratio that they are used to and get a gorgoues result on film. The HD camera is placed next to the film camera and no changes are made in lighting.

One such test had a 5k? backlight on a ballerina performing on stage.(it wasn't the early Panavision test) Looked great on film crap on HD as the DP shot the HD at the same time without adjusting the lighting! His comment then was that HD didn't have the exposure latitude of film. I was asked to critique the test and commented that the HD version was poorly lit and not representative of how a DP would work with it on location.

One should be aware that the Genesis test like all tests before it should be viewed with suspicion.

Borrow a HD camera, good monitor and have a play. Transfer to film at a company that has a vested interest in dealing with digital rushes. This is particularly important in Soho!

Even a (minor) part of the highly regarded DCI StEM test shown in Vegas has been critisised as being described as X when it was actually Y.

Digital has some way to go to match the extreme highlight handling charecteristics of film, test to see what it can do for you today.

Mike Brennan

Mike Brennan · June 20, 2004

The superior dynamic range of film puts fewer restrictions on what a cinematographer can do. You don't have to "work around" the issue of limited latitude and make as many compromises in your lighting.

This is not a one way street, there are comprimises in lighting that come with using film that are beyond this present discussion of highlight latitude.

Depth of field of 2/3 inch HD vs 35mm is one where in low ambient light levels it is film that is compromised. Shadow detail monitoring, spot metering of highlights, are all easier reliable and faster on HD.

As far as I can tell most DPs note that HD is faster to light than film, even though more has to be done to control some highlight areas of the image. In a real world situation being able to light fast means more time to fine tune.

If a DP is trying to achieve a high key look then he will have problems. I'd love to hear of any other specific lighting styles that are difficult to achieve on HD and also of lighting styles that are easy on HD, difficult on film.

Mike Brennan

Max Jacoby · June 20, 2004

As far as I can tell most DPs note that HD is faster to light than film, even though more has to be done to control some highlight areas of the image. In a real world situation being able to light fast means more time to fine tune.

In what situations do you consider HD being faster to light than film?

I for one wouldn't want to be stuck shooting sunlight exteriors with an HD camera.

David Mullen ASC · June 21, 2004

Overall I wouldn't say that lighting for HD is any faster or slower than with film. For the most part, lighting is lighting -- you want to give someone a soft side key by putting a light through a frame of diffusion, it's not going to go any faster or slower with film or video. You want to add a backlight, the time spent is the same.

I don't even think that HD is all that bad in full sunlight -- it handles most things EXCEPT white objects and overexposed areas well, and just because you're shooting in sunlight doesn't follow that the subject is white or that you're overexposing part of the frame.

I have more problems with HD when dealing with sunlit backgrounds but the scene is indoors or in the shade -- THAT'S a challenge. But being in sunlight itself is just as hard in HD or film if you need to reduce the hardness or contrast of the light.

Of course there are subtle differences. For example, I save some minor time in HD simply because it keeps me from OVER-lighting the scene, whereas in film I may worry about something not having enough light on it and spend the time to add something as a "just in case". But that's not a technical advantage to HD; it's just a sign that I'm not always confident with lighting and take advantage of the instant feedback with HD in terms of seeing the results. In other words, it makes me a little lazier...

Yes, you may have to spend more time, on the other hand, controlling bright highlights with HD. That's why I tend to think the net time spent is the same whether you are shooting film or HD, with some minor time savings from reloading HD less often and the fact that with HD you tend to use zooms and with 35mm, you tend to use primes because the 35mm zooms are slower and heavier than the HD zooms.

The only lighting style that works less well with HD is overexposure tricks like the bright slash on some furniture that bounces up and lights the subject in the shadows. I don't think any lighting style works BETTER in HD so much as some lighting problems can be fixed better when you digitally color-correct, so it works for film as well if it is posted digitally. For example, taking the hazier flatness of shots made in overcast and being able to punch up the colors and contrast to match sunlit shots better.

Nate Downes · June 21, 2004

I'll stick to my film cameras for a long while, despite being one of the "new generation" of filmmakers. While HD is interesting, I "understand" film better, I know its limits better, I can work with it better. When I had lighting problems on a shoot, I knew it before the first frame was shot. In time, possibly, I could learn HD, but it is too expensive for me at this time.

And I'd note, I've seen the Fuji's "Interpolation" technique first hand. It's a joke and a half, great for marketing hype and wasting xD space, and nothing else. Absolutely *ZERO* image improvement. Take a normal and an interpolated print side-by-side, and you will not be able to tell which one is which, they're identical. I'd sooner have real-world pixels than some marketing trick.

Mike Brennan · June 23, 2004

"In what situations do you consider HD being faster to light than film?"

This is a fast response excuse typos..

This is down to technique and communication. Nothing mysterious.

There are two aspects to this. One is the use of HD monitor the other is taking advantage of greater depth of field of 2/3 inch imagers (compared to 35mm) for given amount of light

HD monitor

Communication is faster, gaffer sees the shot sees what you are up to faster with a decent monitor. During blocking he watches. He knows what is in shot and what isn't.Point to the monitor, a slash here a flag there, you are not gesturing in space you are pointing at the picture.

The monitor empowers the gaffer. Lots of finger prints on the monitor:)

In frame practicles, reflections, are faster to adjust.

Dropping light sources into a scene, observing where the light falls from the cameras POV is really usefull. This helps gaffers and electricains drop in cutters fingers ect.

Light measurement across the image is instantaneuos. No need to use a spot meter measuring 10 different parts of the picture, as a good monitor has 1920x1080 spot meters that you see at a glance. Their are also zebras for the real enthusiast.

Communication with a director is fast. It is very easy and quick to nail a look under the black cloth. (this is particullarly usefull when working with a new director)

Night scenes can be shot with absolute confidence. Worried about a colour temperature of a neon or a dodgy looking flourescent?

Worried about subject failure? strange casts?

Bounce light in particular is great fun and fast to play with on HD. The brain doesn't need to number crunch. No estimating what colour shifts are happening, you have a ringside seat on the image plane.

Lighting continuity is a breeze if you record a few seconds of every scene on a seperate tape. recalling the scene file and switch between playback and live image to swiftly compare moods and tones from one location to the next. Particularly usefull if something unplanned is occuring.

I'm sure being able to strike a set and derig lights through the benifit of instant replay, without having to wait to the following day has an impact on the speed of the production as a whole. This may be off topic.

Depth of field

Greater depth of field on 2/3 inch imagers. Less light required usually equals faster setup/derig reposition.

Bluescreen, less light for same depth of field. Do you want to light a bluescreen studio and subject to T5.6 or T2? All the subject must be sharp for blue screen

No question packshot lighting is faster on HD. Positioning products in respect to reflections, mini bounce cards ect is a breeze with a big HD monitor a few feet from the table. Assistants look at the monitor. Greater depth of field plays a roll here too. No need to hang a 1.2 with a Chimera from a truss or goal post, use a 800 watt bug light with chimera on a right angle arm. Easy to adjust. Extreme closeups are much easy to light/meter. Fiberoptics are great for closeup work but moving them half an inch equates to a few stops if they are close to a subject. Instead of setting the light then metering your are metering while you set the light.

Sometimes we can use ambient light levels, say in a night exterior as a base, rather than calling in Muscos.

These are lighting techniques that have lent themselves to video and have been practised over many years on video.

DPs and gaffers without video experience may not be tuned in...

My first feature I worked with a very established Italian gaffer who had worked with the greats. He had just come off his first HD feature and was looking stressed! After a week with me he was over the moon. In his words he felt after 25 years in the buisness he was actually crafting the light exactly (to the 1/4 stop) the way *we* wanted it and seeing it live he felt closer to the image than he had ever been on film. I invited him into my decision making process and gave him confidence in the HD monitor.

Film making/shooting is a team effort the HD monitor is a brilliant communication tool. Even a small one is usefull in this respect.

"I for one wouldn't want to be stuck shooting sunlight exteriors with an HD camera."

It just a matter of knowing how to do it, the right approach, ideally from all departments.

If you have a generator or butterfly frame it is simply a question of more fill or go up a level of diffusion on the silk, pop in a pola, or .6 grad, ensure makeup do their job, craft the image to look good. You will only be "stuck" if you and the director have an unrealistic expectation of what the combination of set, crew equipment format grading can achieve. In the whole scheme of a production, films superior dynamic range may be more critical for some projects than others. But one of the first HD movies was shot in a snowfield. Are there special video cameras for the winter olympics?

Now the above comments are more relevant to a tightly scheduled tightly crewed production.

Bear in mind whenever I talk HD it is usually across features docs commercials, not only high budget features where the established working practises of a production crew numbering 200+ remains pretty much unaltered regardless of format.(perhaps bluescreen movies is an exception)

Mike Brennan

Max Jacoby · June 24, 2004

Bear in mind whenever I talk HD it is usually across features docs commercials, not only high budget features where the established working practises of a production crew numbering 200+ remains pretty much unaltered regardless of format.(perhaps bluescreen movies is an exception)

I think the biggest factor is the lighting style of each Dop. 2 features that I worked on over the last year shot on the same set. Obvioulsy the set was dressed differently and some architectural changes were made as well, but the space was the same and the light came from the same windows. On top of that the mod of the scene was the same (soft side-lighting).

Well one Dop put up 2 18Ks through light gridcloth with 1 Kino-Flo for fill, while another one bounced lots of HMIs into bedsheets and used tons of flags. The first one took 30 minutes to light the scene, the second one 4 hours...