The end of film for TV production?

[K Borowski]

No, Karl. What you're referring to is closer to the monkey-ready stuff. Do you even have a Kodak H-24 manual? "Module 3, Analytical Procedures" is the kind of stuff that only real chemists can use. You know, the mad scientist stuff with beakers and smarts.

 

Yes I do.

 

I wouldn't call this real chemistry. And I certainly don't consider myself a chemist. Densitometry and processing control is a highly selective aspect of chemistry, and relatively simple by comparison to the discipline as a whole.

 

I think Dominic put it best when he likened it to 19th century chemistry. Even the measurement of pH and specific gravity in processing control is relatively straight-forward and simple.

 

The equipment: beakers (don't think I have a single one), graduated cylinders, pipettes, pH & pAg strips, 1/5 °F (1/9°C) Mercury thermometers and hydrometers are just tools. Nothing special, nothing exciting. Most go unused unless you have a really out-of-control process. I'm happier NOT to see them come out.

[Paul Bruening]

My father, God rest his soul, was a veritable font of wise (wise cracking) remarks. My favorite was, "If you have to be just one thing in life, be lucky." Another good one was, " If you smell dogs**t, check the bottom of your shoes." But, I think the one that applies best, here, is, "If you hear their voice but their lips aren't moving, they're talking out of another orifice."

[Keith Walters]

My father, God rest his soul, was a veritable font of wise (wise cracking) remarks. My favorite was, "If you have to be just one thing in life, be lucky." Another good one was, " If you smell dogs**t, check the bottom of your shoes." But, I think the one that applies best, here, is, "If you hear their voice but their lips aren't moving, they're talking out of another orifice."

I've heard so many ridiculous arguments here and in other places, all based on people's entirely subjective opinion of image quality. It wouldn't be so bad if there was some semblance of consensus, which might indicate somebody has some sort of point. In reality we seem to get endless blowhard expositions from self-styled Industry authorities, as like as not, diametrically opposed to each other. They can't both be right.

 

The only thing it proves to me is, that by golly, maybe it does make you go blind :rolleyes:

 

In any case, the title of this thread is "The end of film for TV production". How did film prints vs electronic projection of movies get into the act?

[K Borowski]

How did film prints vs electronic projection of movies get into the act?

 

That would probably be when I said something along the lines of the recession being bad for TV shows shot on film, but the end of the recession being even worse for the film manufacturers when all the theatres are able to finance a digital conversion.

 

 

As for "endless blowhard expositions from self-styled Industry authorities," it is ironic that you and Paul should be ranting. . . against "them."

[John Sprung]

Are you sure that modern Projectors only expose each Frame twice and not three times?

 

The vast majority are two blade. Century makes a three blade shutter, you can put either a two or a three in a Century projector. Operating at the right light level (16 ft. Lambert), you really can't tell the difference. Universal put together a little screening room in a hurry a long time ago, grabbing a pair af Centurys. They found that one machine was pulling loads more current than the other to make 16 ft. Lamberts. Turns out that it was a three blade, the other a two. You have to look thru the hot hole while you inch and count to know for sure.

 

Some experiments have been tried with making the interrupt blades narrower than the pulldown blade. It turns out that you can't cheat much that way, the flicker comes back.

 

 

 

 

 

-- J.S.

[John Sprung]

..., this stuff actually gets very easy to control to within 3-4 points (thirtieths of an F/stop).

 

OK, suppose we control things to within 3 points. That's 3/30 of a stop = 1/10 of a stop = 0.1 stop.

 

A change of 0.1 stop is a change by a factor of 2^0.1 = 1.07, or about 7%. The JND (Just Noticeable Difference) for human vision is more like 1%. Pay attention and you'll see a slight color and density shift on a lot of reel changes. Sometimes we get lucky and the "within 7%" lands closer, within 1%.

 

 

 

 

 

-- J.S.

[K Borowski]

OK, suppose we control things to within 3 points. That's 3/30 of a stop = 1/10 of a stop = 0.1 stop.

 

A change of 0.1 stop is a change by a factor of 2^0.1 = 1.07, or about 7%. The JND (Just Noticeable Difference) for human vision is more like 1%. Pay attention and you'll see a slight color and density shift on a lot of reel changes. Sometimes we get lucky and the "within 7%" lands closer, within 1%.

 

 

 

 

 

-- J.S.

 

Come on. Do you really think that current theatrical prints are only showing 1% of a stop difference? I am sure in the past decade, they have gotten worse, not better than 1/100, let alone ~1/14 of a stop.

 

So, who, then, is to blame for this current abortion of a simple densitometric calibration of one print to another: "chemical film" or abysmal quality control?

 

This is a simple matter of matching three colors to another set of three colors, no percentages or calibration factors involved, especially three generations removed.

 

You decide. . .

[Michael Most]

Come on. Do you really think that current theatrical prints are only showing 1% of a stop difference? I am sure in the past decade, they have gotten worse, not better than 1/100, let alone ~1/14 of a stop.

 

So, who, then, is to blame for this current abortion of a simple densitometric calibration of one print to another: "chemical film" or abysmal quality control?

 

This is a simple matter of matching three colors to another set of three colors, no percentages or calibration factors involved, especially three generations removed.

 

You decide. . .

 

Simply saying "it all sucks" is not very constructive. If one really wants to say something that addresses the issue, rather than simply continuing to point out that there ARE issues, one needs to offer some explanations and solutions. So here are mine:

 

1. Why does it all suck when it didn't used to?

The answer is that in the days when "it didn't used to," there were far fewer screens and far smaller resultant print orders, even on the widest releases. When that was the case, most if not all of those prints were made in one lab, with tighter QC, and with better control of things like print light interpretation. Today, the print orders are 4 to 5 times larger, the prints are made in different labs in different cities in different countries simultaneously, with all of the attendant differences that brings, and with far less ability to tightly QC every print that gets made. They also are made under tighter deadlines, thanks to the need for waiting until the last minute for things like visual effect revisions. So you're trying to do more, with less time, in multiple labs at once. That is a recipe for inconsistency, no matter how you look at it.

 

2. What can be done?

What is already being done - which is moving to a digital distribution system that puts the consistency and QC requirements in the hands of the exhibitors rather than multiple labs working under tight deadlines. Because the files distributed are direct copies of the original, variations between distributed elements are eliminated. Digital cinema projectors - in particular, DLP Cinema projectors - are probably the most stable display devices yet built. Once properly installed and set up, they require very, very little further adjustment for months at a time, other than periodic lamp replacement (I speak from direct experience on this, under conditions more stringent than theaters).

 

You can't turn back the clock. The fact is that the more screens there are, the more conditions become ripe for inconsistency with print based distribution. To some degree, it's a victim of its own success. The notion that suddenly everyone is going to start running fewer prints and doing better QC in order to ensure better quality at a time when the labs are barely making money to begin with is just not sensible. One needs to look to the future, and fortunately, there are solutions on the way that are, in the end, better than anything that has come before, regardless of one's own sentimental attachments.

[John Sprung]

Come on. Do you really think that current theatrical prints are only showing 1% of a stop difference?

 

No. That's not what I wrote.

 

A difference of 1% in absolute light level, linear light, is the JND for human vision.

 

You wrote that print density can be controlled to within 3-4 points, and that 30 points make a stop. From that, I just did the math. It turns out that 3 points corresponds to a 7% delta in linear light. We can see a 1% delta, just barely.

 

Therefore, most of the time, if you pay attention, you can see the reel change shift. Sometimes, not often, we get lucky and it does land within 1%. The best that can be achieved with careful work is within 7% -- going by your 3-4 point assertion. And again, all this is in linear light, not stops.

 

 

 

 

-- J.S.

[Terry Mester]

The vast majority are two blade. Century makes a three blade shutter, you can put either a two or a three in a Century projector. Operating at the right light level (16 ft. Lambert), you really can't tell the difference. Universal put together a little screening room in a hurry a long time ago, grabbing a pair af Centurys. They found that one machine was pulling loads more current than the other to make 16 ft. Lamberts. Turns out that it was a three blade, the other a two. You have to look thru the hot hole while you inch and count to know for sure.

 

Some experiments have been tried with making the interrupt blades narrower than the pulldown blade. It turns out that you can't cheat much that way, the flicker comes back.

What is your opinion on Micheal Most's assertion that flicker is a problem? Do you personally find flicker to be a problem? I personally don't.

 

------------------------------------------------------

 

Possible solution to quality problems

I think that there is a simple way to redress these quality control problems that have been noted. The American Society of Cinematographers should set up a Quality Rating system which would rate both Developing / Printing Labs and individual Movies as regards quality. The ratings could be very poor, poor, acceptable, good, very good, and excellent. In the case of rating Movies, they could randomly pick out one of every hundred release prints to view and assess. The ASC is of such high a reputation that nobody would question their integrity.

[John Sprung]

What is your opinion on Micheal Most's assertion that flicker is a problem?

 

I really can't dig back thru 230+ posts to see what Mike said. I can only assume that whatever it was, it didn't move me to the point of posting a reply as I was skimming thru all this .... stuff. ;-)

 

I've known him for 20+ years. He knows his stuff, and tells the truth, even when that's not the comfortable thing to do. We don't disagree very often, and when we do, usually I learn something.

 

 

 

 

 

-- J.S.

[Paul Bruening]

Hello Terry,

 

Sometimes, it is difficult for viewers to distinguish between lamp an blade flicker. Could some of what you are concerned about be lamp flicker? Lamp flicker is easily detectable by the eye. Blade flicker is, usually, considered undetectable by the average viewer. Lamp flicker is when the lamp gets old and the arc rolls around between the contacts.

[Dominic Case]

G'day!

I just dipped into this to see what could take 236 posts to discuss. Turns out it's a free-for-all on a whole barrel of things.

 

But I have to butt in here:

You wrote that print density can be controlled to within 3-4 points, and that 30 points make a stop.

Good maths on bad information isn't any help to anyone. I can't track down who said what but this needs straightening out.

 

0.30 logE is a stop, but that isn't "30 points" anywhere I've heard the term "point" used in labs.

 

A printer "point" is one trim or timing light. It is an exposure change of 0.025 logE, which for those who know what logs are, is equivalent to a linear change of approximately 6 per cent. (Curiously,that is the same as a semi-tone in music).

 

But because of the high gamma of print stock, that results in a density change (also logarithmic) of about 0.07 in the print.

 

Labs vary, but typical quality control from print to print would allow +/-0.05 for show prints, and probably more than that for release prints. However, whatever the threshhold of JND (just noticeable difference) is, that is to do with being able to detect the difference between two adjacent colour patches, not the difference between two image reels screened sequentially.

 

In practice, if you lay two reels that differ by a point (0.07) in colour or density, next to each other on a light box, you will easily see the difference. I'd defy an amateur to pick 0.01 though, or even 0.02.

 

Yes, everything in the production process has a certain variability, that is the way the process works. Typically, the tolerance levels are worked out on the basis of what the lab can achieve. No point in setting a limit that results in rejecting 20% of the work.

 

By the way, Kodak has a quality rating system for labs. It is called Image Care. It's not just based on sampled output, but on a rigorous inspection of the lab's processes, procedures, record-keeping, and training every year. Naturally it's voluntary. Many labs participate, though not many in the USA. Most of them only participate for their negative processing operations. (At least that is the most important part of it!).

 

Some studios or individual filmmakers send in their own QC inspectors to check the prints on a big release coming out of a lab. Basically they screen sample copies in real time, one after the other. Can you imaging watching Star Wars Ep 2 twice a day for several weeks?

 

Over the years, several bodies have attempted to set up a quality rating system for theatres. Kodak, SMPTE for example. Theatres in general won't participate.

 

Not sure what all that proves about the end of film for TV production - except that people will write about what they want to write about, regardless of the topic. How about a bit of rigorous QC here :P

[Terry Mester]

I really can't dig back thru 230+ posts to see what Mike said. I can only assume that whatever it was, it didn't move me to the point of posting a reply as I was skimming thru all this .... stuff. ;-) ...

Well, according to him theatre goers are complaining about flicker. I don't know. I can't see shutter-based flicker as a real problem. However, a failing light as Paul suggests, yes.

 

G'day! ...

Some studios or individual filmmakers send in their own QC inspectors to check the prints on a big release coming out of a lab. Basically they screen sample copies in real time, one after the other. Can you imaging watching Star Wars Ep 2 twice a day for several weeks?

 

Over the years, several bodies have attempted to set up a quality rating system for theatres. Kodak, SMPTE for example. Theatres in general won't participate. ...

G'day Dominic! ;)

I definitely don't have confidence in the Studios as regards quality control for protecting movie goers. The assessment of quality on a given movie needs to come from an independent body that truly cares about this issue. It must be a body that can play bad cop without concerns of backlash.

I can't remember if it was SW 2 or 1, but I was falling asleep and slumping in my seat. George Lucas was way too obsessed with CGI, and not concerned enough with plot line.

 

----------------------------------------

 

Does anyone know definitively that Peter Jackson has "converted" to Digital, or is he just "testing it out"? Many will test it out, and see its flaws first hand. The photographic world went through this novelty of Digital, but after several years about 80% of professionals say they will continue to use Film either in part or exclusively.

[georg lamshöft]

Does anyone know definitively that Peter Jackson has "converted" to Digital, or is he just "testing it out"? Many will test it out, and see its flaws first hand. The photographic world went through this novelty of Digital, but after several years about 80% of professionals say they will continue to use Film either in part or exclusively.

 

His new movie was shot nearly entirely on 35mm ("behind the scenes" showed 35mm-cameras even with bluescreen-work!?) - but we shouldn't mix up technical with artistic quality, some of the greatest storytellers have no clue about technology. James Cameron for example is capable of outstanding work and he cares about technology but strangely, he uses 2/3"-sensor-technology that wasn't even developed for the big screen and at the same time complains about it's technical restrictions...

[Keith Walters]

As for "endless blowhard expositions from self-styled Industry authorities," it is ironic that you and Paul should be ranting. . . against "them."

Hey! You said you weren't going to respond to my posts anymore. You just can't help yourself, can you? Some "professional" you are.

[Keith Walters]

Does anyone know definitively that Peter Jackson has "converted" to Digital, or is he just "testing it out"? Many will test it out, and see its flaws first hand.

Probably one of the premier off-pissing aspects of this whole sad business.

The moment anybody with any relevance to the film industry decides to dip their toe in the Digital waters so to speak, all the fanboys come come out of the woodwork crowing that they're "converted to digital".

I'm sure a lot of them feel like the girl who is terrified of bringing any of her boyfriends home to meet her mother, because Mom's second or third sentence is invariably going to involve discussions of wedding formats or how many children he'd like to have and so on :lol:

There's so much arrant BS spouted about this subject, that I'm sure most of them have already made up their minds in private, and the results speak for themselves.

By the way, has George Lucas made anything since Star Wars Episode 3? (Apart from the last Indiana Jones movie, which was on FILM!! incidentally.

[Thomas James]

With film projection you get 24 frames per second and with a 2 blade shutter you are supposed to get 48 images per second. However with digital projection you get 48 or even 72 frames per second. Now these may not be distinct individual frames but rather repeats of the previous frame but nevertheless its still a lot better than using a shutter.

[Michael Most]

Well, according to him theatre goers are complaining about flicker. I don't know. I can't see shutter-based flicker as a real problem. However, a failing light as Paul suggests, yes.

 

That is NOT what I said. Allow me to quote myself:

 

... given a look at both, theatergoers prefer digital projection to film projection by a very, very wide margin. Reasons often cited are lack of dirt, lack of scratches, lack of weave, less flicker, and the fact that the picture still looks good in its second or third week of release.

 

I didn't say "theater goers are complaining about flicker." I cited it as one of a number of factors as to why they generally prefer digital projection. And what I meant by that is that digital projection - at least as commonly exhibited today by DLP Cinema projectors - has no discernible flicker. Film projection has flicker caused by multiple issues - 48Hz refresh rate (with a 2 bladed shutter, common in theatrical projection) with part of that period blanked, inconsistency in prints and an exaggeration of that inconsistency over time, and lamp issues. A DLP Cinema projector has none of that because there is essentially no blanking period, no physical media, and lamp failure is usually catastrophic and so the lamp is replaced.

 

If you're going to paraphrase me, please do it accurately.

[John Sprung]

Some studios or individual filmmakers send in their own QC inspectors to check the prints on a big release coming out of a lab. Basically they screen sample copies in real time, one after the other. Can you imaging watching Star Wars Ep 2 twice a day for several weeks?

 

Actually it's worse than that -- or at least it was back when I worked for TriStar. We had to check every reel of every print on 70mm releases, mainly because of the magnetic sound. What made it worse is that we had to watch prints of the same reel over and over, not the whole show.

 

As for printer points, perhaps things have become more standardized than they were 30+ years ago. Back when I was dealing with film printing, different labs each had their own number of points into which they divided a stop. I'm sure some had twice as many points as others, but the exact numbers I can't swear to. I seem to recall something like 25 and 50, but that's a very shaky long ago memory.

 

Of course you're right that the JND thing would only apply if the changeover happened within a shot. Because changeovers always happen on cuts, that would greatly mask any difference. So I had the theory wrong on that. Thanks for the correction.

 

 

 

-- J.S.

[K Borowski]

Good maths on bad information isn't any help to anyone. I can't track down who said what but this needs straightening out.

 

0.30 logE is a stop, but that isn't "30 points" anywhere I've heard the term "point" used in labs.

 

A printer "point" is one trim or timing light. It is an exposure change of 0.025 logE, which for those who know what logs are, is equivalent to a linear change of approximately 6 per cent. (Curiously,that is the same as a semi-tone in music).

 

But because of the high gamma of print stock, that results in a density change (also logarithmic) of about 0.07 in the print.

 

Labs vary, but typical quality control from print to print would allow +/-0.05 for show prints, and probably more than that for release prints. However, whatever the threshhold of JND (just noticeable difference) is, that is to do with being able to detect the difference between two adjacent colour patches, not the difference between two image reels screened sequentially.

 

In practice, if you lay two reels that differ by a point (0.07) in colour or density, next to each other on a light box, you will easily see the difference. I'd defy an amateur to pick 0.01 though, or even 0.02.

 

 

Not sure what all that proves about the end of film for TV production - except that people will write about what they want to write about, regardless of the topic. How about a bit of rigorous QC here :P

 

It was I who was referring to .01 ln as a point.

 

Perhaps I was confusing others into thinking I was talking about additive printing points. But densitometry, on a calibrated machine, is more or less standardized so that 0.30 equals a stop of density difference.

 

 

John is right in pointing out that the system is confusing, and there isn't always standardization in printing points, so I figured I would confine the discussion to densitometry, as "chemicals" has been continually thrown around here as a dirty word.

 

I know of some machines that can adjust the value of a "point", and there used to be the added confusion of subtractive printing numbers too. Those numbers were even more non-standardized, as the filters would fade more quickly.

 

 

 

I really don't have any knowledge of release printing machines, just your standard contact printers, but I assumed that you would have very fine, even finer control than with standard printing lights to allow results to be consistent from reel to reel.

 

What I am really trying to get at is that there is little subjectivity and artistic decision at the release printing step. You're trying to match a grey "bullseye," that you are reading with a densitometer on the rlease print to a number.

 

And, while I guess there are instances where color will drift from the start of a reel to the end of a reel, I think that getting the grey dot to a lot closer than 0.25 log off will go a long way to improving color shifts. Maybe things have gotten a lot better or a lot worse since I was up in a booth with a densitometer, bored one day, but the type of drifts you are talking about, Dominic, are no where near what I was seeing.

[K Borowski]

Actually it's worse than that -- or at least it was back when I worked for TriStar. We had to check every reel of every print on 70mm releases, mainly because of the magnetic sound. What made it worse is that we had to watch prints of the same reel over and over, not the whole show.

 

That would really suck. I assume you had to watch in real time because faster speeds was result in the potential for damage? I always thought it would be done at 2-3 times projection speed with a squawk box, editor screens, and several reels running simultaneously.

 

 

As for printer points, perhaps things have become more standardized than they were 30+ years ago. Back when I was dealing with film printing, different labs each had their own number of points into which they divided a stop. I'm sure some had twice as many points as others, but the exact numbers I can't swear to. I seem to recall something like 25 and 50, but that's a very shaky long ago memory.

 

This has been my experience too. I've seen machines where the number of points in a stop could be adjusted.

 

I thought things had been standardized so there was a correlation between ln points and printer points, but I guess I am wrong.

 

Still, if there are 0.025 nat. logarithmic units (notice I didn't say "points" :-p) in a printer point, I've seen results that are off by eight, ten, twelve printer points. That seems a little excessive if there really is competent quality control going on

 

Of course you're right that the JND thing would only apply if the changeover happened within a shot. Because changeovers always happen on cuts, that would greatly mask any difference. So I had the theory wrong on that. Thanks for the correction.

 

 

You make a good point though. And when you are cutting from a closeup of a group of people to a long shot of the same, 1% difference (are you talking about printer points here too?) might not be noticeable, but 10% would be :(

 

 

 

 

As for how this whole thing got derailed to the topic of motion picture release prints, that was a sidebar comment I made, so I guess I am at fault. Discussions that go off topic are so much fun though! :P

 

BTW, Dominic, I guess I had been going on the assumption that most color variances were chemical in nature, not due to light discrepancies.

 

Would I be correct in assuming that both sets of discrepancies are additive, not compounding in nature, but can be corrected out at each step in the process?

[John Sprung]

That would really suck. I assume you had to watch in real time because faster speeds was result in the potential for damage? I always thought it would be done at 2-3 times projection speed with a squawk box, editor screens, and several reels running simultaneously.

 

We watched in real time, primarily because we were checking sound and picture, and wanted to be sure that we were catching anything that the audience would catch. Magnetic sound is recorded onto each print individually, and there can be dropouts if the oxide stripes don't adhere. Some of the AA-2 Norelco's had 30 fps motors which we could have used with no risk to the film, but the company was willing to pay for real time QC. These were the 70mm road show prints. The real fun is when you find a problem with all the prints of a reel, so you have to check the replacements 24/7 to make a deadline. Happened with reel two of a horror picture called "Life Force".

 

High speed QC is done on 35mm prints in some cases. But they're going on the assumption that if it's not grossly effed up, it's most likely OK. Most often, that's true.

 

 

 

-- J.S.

[Michael Most]

High speed QC is done on 35mm prints in some cases. But they're going on the assumption that if it's not grossly effed up, it's most likely OK. Most often, that's true.

 

You and I also go back to the days when the lab (well, MGM Lab, at least) would high speed dailies every night. I got more than one call at 5AM from Bill Bickford back in the day....

[Terry Mester]

...

And, while I guess there are instances where color will drift from the start of a reel to the end of a reel, I think that getting the grey dot to a lot closer than 0.25 log off will go a long way to improving color shifts. Maybe things have gotten a lot better or a lot worse since I was up in a booth with a densitometer, ...

I think that a root cause of these colour inconsistencies in a print is the disparate Speeds of Film used to shoot the movie. Directors can't expect to get even closely comparable colour quality from 500 ISO that they would get from 50 ISO. Those old Technicolor movies we all love were shot on 5 ASA (yes, 5) Film.

 

However, Digital isn't lily white. Here is an impartial Wiki quote about the colour accuracy problems with Digital Projectors (which have been praised here as if the best thing since sliced bread):

--------------------

"According to DLP.com, the three-chip projectors used in movie theaters can produce 35 trillion colors, which many suggest is more than the human eye can detect. The human eye is suggested to be able to detect around 16 million colors, which is theoretically possible with the single chip solution. However, this high color precision does not mean that three-chip DLP projectors are capable of displaying the entire gamut of colors we can distinguish (this is fundamentally impossible with any system composing colors by adding three constant base colors). In contrast, it is the one-chip DLP projectors that have the advantage of allowing any number of primary colors in a sufficiently fast color filter wheel, and so the possibility of improved color gamuts is available."

--------------------

 

There are then the separate problems of expense of purchase and maintenance, and the fleeting lifespan of electronic technology. I will gladly go out on a limb right now, and predict that a few years from now many Digital Theatres will be bringing their 35mm Projectors out of retirement rather than go heavily into debt to buy new Digital Projectors.