Sebastian Cramer
Well-known member
- Joined
- Mar 15, 2007
- Messages
- 112
- Reaction score
- 0
- Points
- 0
After having constantly issues with the focus and mismatching scales on my two cameras #112 and #1513 I did a test over the past 2 days looking into this topic.
As it has been discussed already in other threads, it looks as non matching back focus issues seem to be more obvious starting in the range somewhere arount ser. # 1000. As a reason for this a change in the design of the OLPF was suspected to be the cause of this problem. I do indeed have the strong feeling optical elements (like OLPF, IR filter, UV filter and the sensor glas surface) are the cause here, as I know those elements can easily add to a total thickness of something like 3mm in other digital cameras.
As the outcome of this test was as surprising as frustrating to me, I decided to put up a new thread as it seem this is indeed something to be addressed and solved immediatelly.
If anybody has the gear and a chance to please verify these results, do it and prove me I'm wrong!!! Please!
To make a long story short, I could not see any shift of focus caused by thermal issues, as all effects could be reproduced exactly the same after a few hours. This doesn't mean there are no thermal issues, but what is addressed here seems to be purely related to the camera itself.
I compared the following lenses on Red #112 and #1513:
Master Prime: 25mm/32mm/75mm,
S4: 25mm/32mm/75mm,
Ultra Prime: 10mm/14mm/24mm/32mm/85mm
Superspeed: 18mm/25mm/35mm/50mm/85mm
MP, S4, UP were supplied by FGV Schmidle, a very respected high end rental house in Munich. Superspeeds came from Ready2Roll, which is my company.
After the test all lenses have been checked on a cine projector using an Angenieux Test Cine 35 slide. Markus Schmidle personally checked all lenses by himself on the projector. The decision to go with the projector was made, because this was the closest to the test environment. Offsets are listed below, as well as the focus readings on the lenses sitting in the projector.
As there is no way to physically measure the flange focal distance like on a 35 camera, backfocus can only be adjusted optically. As I was not sure setting the backfocus on a collimator would exactly project onto the plane on which the light it processed, I decided to set the back focus with a lens I felt to be correct. I took the 75mm Master Prime from FGV in this case.
As I was testing the distance of 1m first, I set up a Putora sharpness indicator at an angle of 45° with it's center at exactly 1.00m distance to the focal plane. I cross checked infinety on #112 and it was dead on too.
The image was monitored in 2K 16:9 in magnifying mode as this gives the most zoom into the image of all settings. Image was checked on a 23" Sony full HD monitor out of the PVW out. An angled Putora with such a big monitor allows for very precise adjustments of the focus. Find a jpg of the test chart taken with the MP75mm. Exact same results could be achieved hours later.
The camera itself was sitting on locked off heavy duty sticks with an OConnor 2060, so consider this as ridged. The position of the bridgeplate was exactly marked for distance to swap the two cameras. And yes, they precicly came back to their position.
On #1513 backfocus was adjusted to the 1,00m reading in the environment described above with the same MP 75mm. This already let to focus position of 3mm beyond infinity on the MP 75mm, which is very very odd...
Anyway, I checked both cameras in exactly the same environment with a constant fixed distance of 1.00m to the center of the Putora. Then I moved the Putora to a distance of 3.00m and did everthing again. No adjustments to the back focus have been made here or at any time in this test.
The results are posted as an image to keep the colums straight.
There is no other explanation to me than the optical path of both cameras inside is VERY (!!!) different. It doesn't have to be the OLPF, it can be anything with a different index of refraction than air.
And yes, quite a few lenses require their own backfocus setting if you want to work properly. There is no such thing as a propper flange focus distance of 52mm anymore. All those film lenses have been designed to only hit the film plane with no other elements in between. Bottom line is, if lenses are properly collimated for film, they are not neccessarily giving a propper reading on the Red, no matter where the back focus is. If you set it up for a different lens, it throws off the others. Looks like this is the digital world.
At a closer look at the readings, they appear not to be random. If the 1m and 3m value is compared for both individual cameras you find in most cases any mismatch is increased in the 3m value, if it showed up at 1m in the first place. So there seems to be some sort of rule to this.
#1513 is completly not usable! Bottom line. No way, not usable.
I'm urgently asking Red to modify this camera to behave like #112.
#112 is much better, but it is still very irritating that there are SOME mismatches in high class lenses, which are MUCH MUCH to heavy to be ignored. "Much better" in this regard means "much better to #1513" but far from being at least what I expect from a professional camera. But at least if mismatches are obvious they can be compensated for, whenever the lens is not stopped to a safer T-stop.
Other than I thought in earlier posts I could not verify a significant change of the focal distance when stopping down. Very few lenses showed a minor shift but in a normal production environment it seem to me this should be compensated by the increase of depth of field. At least with the tested lenses...
Please prove, all of this is wrong! Maybe some of the high end rental houses in the world can set up something similar. I would love to hear this can't be reproduced.
All of this leaves me very frustrated. It is not so much the fact that film and digital is not as close together as it seems in the first place. It's the fact I'm facing those two completely different results on both cameras.
Looking foreward for an explanation to all of this from Red. Thanks.
Best,
Sebastian
As it has been discussed already in other threads, it looks as non matching back focus issues seem to be more obvious starting in the range somewhere arount ser. # 1000. As a reason for this a change in the design of the OLPF was suspected to be the cause of this problem. I do indeed have the strong feeling optical elements (like OLPF, IR filter, UV filter and the sensor glas surface) are the cause here, as I know those elements can easily add to a total thickness of something like 3mm in other digital cameras.
As the outcome of this test was as surprising as frustrating to me, I decided to put up a new thread as it seem this is indeed something to be addressed and solved immediatelly.
If anybody has the gear and a chance to please verify these results, do it and prove me I'm wrong!!! Please!
To make a long story short, I could not see any shift of focus caused by thermal issues, as all effects could be reproduced exactly the same after a few hours. This doesn't mean there are no thermal issues, but what is addressed here seems to be purely related to the camera itself.
I compared the following lenses on Red #112 and #1513:
Master Prime: 25mm/32mm/75mm,
S4: 25mm/32mm/75mm,
Ultra Prime: 10mm/14mm/24mm/32mm/85mm
Superspeed: 18mm/25mm/35mm/50mm/85mm
MP, S4, UP were supplied by FGV Schmidle, a very respected high end rental house in Munich. Superspeeds came from Ready2Roll, which is my company.
After the test all lenses have been checked on a cine projector using an Angenieux Test Cine 35 slide. Markus Schmidle personally checked all lenses by himself on the projector. The decision to go with the projector was made, because this was the closest to the test environment. Offsets are listed below, as well as the focus readings on the lenses sitting in the projector.
As there is no way to physically measure the flange focal distance like on a 35 camera, backfocus can only be adjusted optically. As I was not sure setting the backfocus on a collimator would exactly project onto the plane on which the light it processed, I decided to set the back focus with a lens I felt to be correct. I took the 75mm Master Prime from FGV in this case.
As I was testing the distance of 1m first, I set up a Putora sharpness indicator at an angle of 45° with it's center at exactly 1.00m distance to the focal plane. I cross checked infinety on #112 and it was dead on too.
The image was monitored in 2K 16:9 in magnifying mode as this gives the most zoom into the image of all settings. Image was checked on a 23" Sony full HD monitor out of the PVW out. An angled Putora with such a big monitor allows for very precise adjustments of the focus. Find a jpg of the test chart taken with the MP75mm. Exact same results could be achieved hours later.
The camera itself was sitting on locked off heavy duty sticks with an OConnor 2060, so consider this as ridged. The position of the bridgeplate was exactly marked for distance to swap the two cameras. And yes, they precicly came back to their position.
On #1513 backfocus was adjusted to the 1,00m reading in the environment described above with the same MP 75mm. This already let to focus position of 3mm beyond infinity on the MP 75mm, which is very very odd...
Anyway, I checked both cameras in exactly the same environment with a constant fixed distance of 1.00m to the center of the Putora. Then I moved the Putora to a distance of 3.00m and did everthing again. No adjustments to the back focus have been made here or at any time in this test.
The results are posted as an image to keep the colums straight.
There is no other explanation to me than the optical path of both cameras inside is VERY (!!!) different. It doesn't have to be the OLPF, it can be anything with a different index of refraction than air.
And yes, quite a few lenses require their own backfocus setting if you want to work properly. There is no such thing as a propper flange focus distance of 52mm anymore. All those film lenses have been designed to only hit the film plane with no other elements in between. Bottom line is, if lenses are properly collimated for film, they are not neccessarily giving a propper reading on the Red, no matter where the back focus is. If you set it up for a different lens, it throws off the others. Looks like this is the digital world.
At a closer look at the readings, they appear not to be random. If the 1m and 3m value is compared for both individual cameras you find in most cases any mismatch is increased in the 3m value, if it showed up at 1m in the first place. So there seems to be some sort of rule to this.
#1513 is completly not usable! Bottom line. No way, not usable.
I'm urgently asking Red to modify this camera to behave like #112.
#112 is much better, but it is still very irritating that there are SOME mismatches in high class lenses, which are MUCH MUCH to heavy to be ignored. "Much better" in this regard means "much better to #1513" but far from being at least what I expect from a professional camera. But at least if mismatches are obvious they can be compensated for, whenever the lens is not stopped to a safer T-stop.
Other than I thought in earlier posts I could not verify a significant change of the focal distance when stopping down. Very few lenses showed a minor shift but in a normal production environment it seem to me this should be compensated by the increase of depth of field. At least with the tested lenses...
Please prove, all of this is wrong! Maybe some of the high end rental houses in the world can set up something similar. I would love to hear this can't be reproduced.
All of this leaves me very frustrated. It is not so much the fact that film and digital is not as close together as it seems in the first place. It's the fact I'm facing those two completely different results on both cameras.
Looking foreward for an explanation to all of this from Red. Thanks.
Best,
Sebastian
