When to intentionally overexpose negative?

[Dimitri Zaunders]

I've heard from a number of sources, such as in the great book 'Film Lighting' by Kris Malkiewicz, that it's generally best to overexpose negative for a denser, 'healthier' negative.

 

But is this always the case? When I asked him today about it, my cinematography professor said that this only applied when you intended to make prints, and that today when digitizing film there is no benefit to overexposing.

 

So is there any truth to this? I'm shooting a short next weekend and I had intended to overexpose 2/3 stop, now I'm not so sure.

[Saul Rodgar]

This subject has been covered here extensively, and that may be an understatement. You may want to check the archives. But yes, generally speaking most of us overexpose our negative stock regardless of digital of print finish. Overexposing cleans up grain and prevents the shadows from going completely black, among other things. As to how much, well, most people here say 2/3s of a stop is what they feel comfortable with, sort of conservative overexposure, which is a good thing. I usually go 1 stop, depending on the situation. And sometimes I don't overexpose at all. Like I said, it just depends what is needed case by case.

Edited April 13, 2010 by Saul Rodgar

[Dimitri Zaunders]

Thanks for the reply Saul. I've seen this question come up a couple of times since I came here, and I've come across it in the archives a few times too. But unless I'm mistaken I don't think a consensus has ever been reached about whether the standard practice of overexposing is still a good idea when you plan to telecine rather than make prints. I've seen a few people recommend it anyway, but is this just a case of old habits dying hard or do the benefits such as tightening grain structure and retaining detail in the shadows still apply?

[Saul Rodgar]

I have done overexposure for both prints and telecine and have always had excellent results. My lab techs have never complained. My printer lights are therefore higher than if I hadn't overexposed, but well within tolerances.

 

I guess this is a practice that really is based on experience and taste rather than recommendation. Someone can say whatever they like or have had good results with in the past, but that doesn't mean that it will translate well to what you are trying to convey with cinematography. That is why shooting tests is always good, and nothing replaces hands-on experience, of course. So I don't really think you will find definitive consensus about overexposure. Some people may argue that the only way to expose the negative is by the Kodak / Fuji charts, and some others will disagree. Just like anything else, I guess . . . ;)

 

Good luck!

Edited April 13, 2010 by Saul Rodgar

[Dimitri Zaunders]

Thanks again Saul, you're completely right about the need to shoot tests to know for sure. As a student I wish I could convince others that shooting tests is necessary, but sadly it's already hard enough to even convince people to shoot on film in the first place!

[Chris Burke]

Thanks again Saul, you're completely right about the need to shoot tests to know for sure. As a student I wish I could convince others that shooting tests is necessary, but sadly it's already hard enough to even convince people to shoot on film in the first place!

 

 

film or digital, tests are always a good idea. The consensus is overexpose color negative.

[Adrian Sierkowski]

+1 vote for over exposing by 2/3 of a stop to 1 stop. Personally; I always over expose any stock 200T or above just to tighten up the grain. For stuff like 100T, 50D, I don't really see the need personally; but people do it.

[Tom Jensen]

Kodak is in the business of selling film, so the more appealing the speed, the better it sells. Film really only has a one stop latitude. Latitude is the ability to under expose film and still get a decent print. Kodak knows that you can shoot their film at there designated ISO because the print will be fine. The best thing to do is test your film. Testing has been done again and again and the results are generally consistent in that you can rate it at 2/3rds of a stop over and get the best quality print. Most film shot goes to telecine and film has wide Brightness Range which a lot of people erroneously call latitude. In telecine over and under exposure is not as critical and you can generally save a lot of improperly expose film that you could otherwise not save if you were going to print. Over exposing a full stop works but you are adding more exposure to the shoulder region of the characteristic curve the you don't really need.

[Saul Rodgar]

Thanks again Saul, you're completely right about the need to shoot tests to know for sure. As a student I wish I could convince others that shooting tests is necessary, but sadly it's already hard enough to even convince people to shoot on film in the first place!

 

I hear you. One can only hope people realize that shooting on film sets them apart ,look-wise, from the rest of the pack --who shoots on video now. Sad days indeed.

Edited April 13, 2010 by Saul Rodgar

[Tom Jensen]

I hear you. One can only hope people realize that shooting on film sets them apart ,look-wise, from the rest of the pack --who shoots on video now. Sad days indeed.

 

Yes and no. Exposing film properly is no easy task but once you have it figure out it becomes a little easier. Video is instant and you can see it right there on the monitor. One day there will be no more film and the reason I say that is the progress video has made over the last 20 years. What kind of progress do you think the next 20 years will bring. It will be mind boggling.

[Alex Malm]

Film really only has a one stop latitude.

 

I think we've been taught different definitions for film latitude.

[Alex Malm]

Bear in mind that this is from Kodak's "The Essential Reference Guide For Filmmakers."

 

"Latitude in exposure is the permissible change in camera exposure that can be made without a significant effect on image quality."

 

So far this is the definition you gave so kudos for that, but my reaction was mainly to your claim that film has only 1 stop of latitude because I have personally experienced the saving grace of film latitude of over one stop.

 

"A typical characteristic curve covers a log E [log exposure] range of 3.0. A range of 1.8 can fit inside that range easily with some latitude to spare."

 

Note: 1.8 divided by .3 is 6 stops, which you might notice will account for 3.5 stops underexposure being empty black and 2.5 stops overexposure being empty white.

 

"A normal exposure would be placed at the speed point [ASA rating]. Moving in steps of .3 log E units (one stop), we see that we can move the brightness range left two times before running off the curve."

 

In other words there are two stops of underexposre latitude before D-min is reached at the low end of the curve.

 

"Similarly, we can move right two times before running off the curve."

 

Of course, this is just an example, but here we have four overall stops of latitude. Two up and two down.

[Alex Malm]

As for the poster's original question here's what I know so far:

 

When to overexpose negative...

 

-When you want more saturated colors overall

-When you want to minimize the appearance of grain

 

When to underexpose negative...

 

-When you want desaturated colors overall

-When you want to emphasize the appearance of grain

[Edgar Dubrovskiy]

For my grad (www.MagiciansWifeFilm.com) I've just wrapped, I did some tests: exposed correctly and with 2/3 overexp, then asked the lab (Technicolor) to pull it back in telecine to the 'normal' level. The test was shot on s16, 7217 Kodak 200T and telecined to HDCAM.

I did expect the grain to be tighter - but frankly the test showed no apparent difference in the grain amount and size when viewed on HD monitor.

Could be that the size of the monitor was hiding the difference, not sure, but on the monitor two shots looked exactly the same, with similar minor grain in the grey card.

 

So I may agree with Adrian, that '2/3 over' trick works for stocks over 200T. I personally did not see the difference, and loosing 2/3 of 200T is a big sacrifice for a student lighting budget, unfortunately.

Edited May 10, 2010 by Edgar Dubrovskiy

[Edgar Dubrovskiy]

For my grad...

 

To add to my previous post - the test was done to see if overexposing 2/3 will reduce the grain.

Nothing else mattered (like more shadows detail etc.) in this test.

So we were looking for the effects of overexposure on the grain ALONE.

Edited May 10, 2010 by Edgar Dubrovskiy

[Chris Burke]

For my grad (www.MagiciansWifeFilm.com) I've just wrapped, I did some tests: exposed correctly and with 2/3 overexp, then asked the lab (Technicolor) to pull it back in telecine to the 'normal' level. The test was shot on s16, 7217 Kodak 200T and telecined to HDCAM.

I did expect the grain to be tighter - but frankly the test showed no apparent difference in the grain amount and size when viewed on HD monitor.

Could be that the size of the monitor was hiding the difference, not sure, but on the monitor two shots looked exactly the same, with similar minor grain in the grey card.

 

So I may agree with Adrian, that '2/3 over' trick works for stocks over 200T. I personally did not see the difference, and loosing 2/3 of 200T is a big sacrifice for a student lighting budget, unfortunately.

 

 

 

could be the monitor. I had about 10,000 feet of Fuji S16, both transferred on a Shadow to HD and a 2k scanner to dpx. From either source, it all depended on the monitor how much grain was visible. Don't be surprised if further down the road you see grain where you didn't before, or vice versa.

[Edgar Dubrovskiy]

In the end - shot it with 1/3 'for safety' overexposure :)

Edited May 10, 2010 by Edgar Dubrovskiy

[Tom Jensen]

Bear in mind that this is from Kodak's "The Essential Reference Guide For Filmmakers."

 

"Latitude in exposure is the permissible change in camera exposure that can be made without a significant effect on image quality."

 

So far this is the definition you gave so kudos for that, but my reaction was mainly to your claim that film has only 1 stop of latitude because I have personally experienced the saving grace of film latitude of over one stop.

 

"A typical characteristic curve covers a log E [log exposure] range of 3.0. A range of 1.8 can fit inside that range easily with some latitude to spare."

 

Note: 1.8 divided by .3 is 6 stops, which you might notice will account for 3.5 stops underexposure being empty black and 2.5 stops overexposure being empty white.

 

"A normal exposure would be placed at the speed point [ASA rating]. Moving in steps of .3 log E units (one stop), we see that we can move the brightness range left two times before running off the curve."

 

In other words there are two stops of underexposure latitude before D-min is reached at the low end of the curve.

 

"Similarly, we can move right two times before running off the curve."

 

Of course, this is just an example, but here we have four overall stops of latitude. Two up and two down.

You're doing math. I was shooting film. It was one stop in either direction. Go shoot some tests over and under, make prints, watch it on the big screen then come back and give me your opinion. If you think a two stop underexposed negative printed up will give you the same image quality as a perfectly exposed negative, you would be mistaken. Sure film has gotten better but a four stop difference is huge. There is really only one definition for film latitude. Everything else is just a misuse of the word.

[Mark Dunn]

I think we've been taught different definitions for film latitude.

Latitude isn't the same as dynamic range, which is many stops.

[Alex Malm]

You're doing math. I was shooting film. It was one stop in either direction. Go shoot some tests over and under, make prints, watch it on the big screen then come back and give me your opinion. If you think a two stop underexposed negative printed up will give you the same image quality as a perfectly exposed negative, you would be mistaken. Sure film has gotten better but a four stop difference is huge. There is really only one definition for film latitude. Everything else is just a misuse of the word.

 

Did you use a densitometer to help determine the actual speed of the film? Or did you do your test based on the EI? Because of batch to batch variances the asa of film won't necessarily match what's printed on the box.

 

Also, I participated in a test of kodak's 5213 this past summer. In the case of that particular film stock we were able to underexpose by more than 2 stops before appreciable differences in quality became apparent. I think it may have even been 3. However, this was scanned at 2k so maybe the differences would be more evident if we viewed a print, as you mentioned earlier. That was a good point. Although that limit is also influenced by how the printer lights are set up since if we're assuming 25 across as our starting point and .025 density per light then we can only compensate by just over 2 stops anyway on a print.

 

We didn't use a densitometer on our test so there could be a margin a error there.

 

You're right, there is one definition of film latitude. That's why I quoted it out of Kodak's reference book just to make sure I wouldn't fudge up the details. What more do you need?

[Alex Malm]

Latitude isn't the same as dynamic range, which is many stops.

 

I'm not talking about dynamic range.

[Bill DiPietra]

Overexposing negative is pretty common practice. How much you overexpose varies.

 

One of my film teachers used to tell me to overexpose negative by one full stop. But over time I've found that a 1/2 stop of overexposure is usually just right...for me. It can also depend on what stocks you are using as well as what kind of look you are comfortable with before any post work. This is why tests are so critical.