- Banned
- #1
Pawel Achtel
Banned
- Joined
- Oct 12, 2007
- Messages
- 3,648
- Reaction score
- 0
- Points
- 0
- Location
- Tasmania
- Website
- www.achtel.com
Today I had a great evening at Arri Australia projecting and evaluating various lenses that I intend to put on my Red One and later on Epic.
My Evaluation Criteria
The assumption was that, in order to resolve for Mysterium sensor, a clear and contrasty image of ~70lp/mm should be obtained across the image plane. Accordingly, my main criteria for glass selection was to achieve ~70% MTF at 70lp/mm. To aid my estimation, I assumed that the 70% modulation corresponds to roughly 5-10%MTF (barely visible) at ~200lp/mm. Other than that, the testing was very "unscientific" and mainly aimed at giving me a ball-park guidance.
Wide Angle
We projected 16mm MP and 16mm UP. The difference in the 60% centre section was very small, but noticable. Both lenses resolved 70lp/mm nicely in the centre. However, at the edges the 16mm UP is soft - very soft. Essentially, if you shooting 3k widndowed, the UP will be OK, but at full sensor coverage expect significant bluring and CA towards the edges of the frame and corners are even worse strugling with resolving even 30lp/mm. Master Prime had some softness in the very corners too, but the edges were as crisp as it gets.
My opinion is that in wide angle department Master Prime is an absolute minimum glass to have on a 4k (or 5k) chip, no ifs or buts. In the corners, the limitation is still in the glass and not in a 4k sensor, not even in a 3k sensor! But anywhere else, including edges, the 4k sensor would limit the final image resolution.
Medium Angles
We projected 35mm MP and 32mm UP. There was less difference between UP and MP here, however I still decided to put 35mm Master Prime in the shopping basket in favour of an equivalent UP. Not as prominent as at 16mm wide angle but, after careful consideration, I concluded that the MP is worth the extra cost as the difference in corner resolution was still significant. What I mean significant is that it is clearly a limiting factor over the CMOS resolving power. At 3k crop area both lenses out-resolved the sensor and produced 70lp/mm with high contrast and I could not see any significant difference in their performance in this region. For budget concious, going UP in those angles (~35mm focal length) is tempting, specially if shooting 1:1.85 or 1:2.35 aspect ratios. I decided to "future proof" for Epic and selected a MP again.
Long Lenses
We projected 75mm MP and 80mm UP. I was hard pressed to tell them apart except for the very corners and even that was marginal. Both performed superbly resolving far beyond what Mysterium is capable of.
We also put 180mm UP on the projector. Again, beautiful lens: crisp corner-to-corner and out-resolving Mysterium sensor comfortably.
Conclusions
I am taking delivery of 14mm MP (wider than 16mm tested) and this is a must-have lens at those angles. If you think any lesser glass would resolve 4k at super wide angles, you are kidding yourself! Master Prime barely achieves this and it is miles apart from Ultra Prime.
If you are aiming at Red One's or Epic's full raster, medium angles of about 35mm focal length are still in Master Prime territory. The difference is on the edges and corners only, but the difference is clearly visible. Border line, but enough to swing me into selecting Master Prime.
At longer focal lengths of ~75mm, you will save heaps of money by going Ultra Primes. They are razor sharp corner-to-corner, edge-to-edge and, in my opinion, there is absolutely no justification to go Master Prime. Mysterium chip resolving power will be a limiting factor here, certainly not the glass.
It was quick and rough "sanity check", but I hope some of you guys will find the above observations useful.
It was certainly good fun playing with all this glass and I would like to thank Rey and Stefan of Arri Australia for their kind invitation and assistance.
My Evaluation Criteria
The assumption was that, in order to resolve for Mysterium sensor, a clear and contrasty image of ~70lp/mm should be obtained across the image plane. Accordingly, my main criteria for glass selection was to achieve ~70% MTF at 70lp/mm. To aid my estimation, I assumed that the 70% modulation corresponds to roughly 5-10%MTF (barely visible) at ~200lp/mm. Other than that, the testing was very "unscientific" and mainly aimed at giving me a ball-park guidance.
Wide Angle
We projected 16mm MP and 16mm UP. The difference in the 60% centre section was very small, but noticable. Both lenses resolved 70lp/mm nicely in the centre. However, at the edges the 16mm UP is soft - very soft. Essentially, if you shooting 3k widndowed, the UP will be OK, but at full sensor coverage expect significant bluring and CA towards the edges of the frame and corners are even worse strugling with resolving even 30lp/mm. Master Prime had some softness in the very corners too, but the edges were as crisp as it gets.
My opinion is that in wide angle department Master Prime is an absolute minimum glass to have on a 4k (or 5k) chip, no ifs or buts. In the corners, the limitation is still in the glass and not in a 4k sensor, not even in a 3k sensor! But anywhere else, including edges, the 4k sensor would limit the final image resolution.
Medium Angles
We projected 35mm MP and 32mm UP. There was less difference between UP and MP here, however I still decided to put 35mm Master Prime in the shopping basket in favour of an equivalent UP. Not as prominent as at 16mm wide angle but, after careful consideration, I concluded that the MP is worth the extra cost as the difference in corner resolution was still significant. What I mean significant is that it is clearly a limiting factor over the CMOS resolving power. At 3k crop area both lenses out-resolved the sensor and produced 70lp/mm with high contrast and I could not see any significant difference in their performance in this region. For budget concious, going UP in those angles (~35mm focal length) is tempting, specially if shooting 1:1.85 or 1:2.35 aspect ratios. I decided to "future proof" for Epic and selected a MP again.
Long Lenses
We projected 75mm MP and 80mm UP. I was hard pressed to tell them apart except for the very corners and even that was marginal. Both performed superbly resolving far beyond what Mysterium is capable of.
We also put 180mm UP on the projector. Again, beautiful lens: crisp corner-to-corner and out-resolving Mysterium sensor comfortably.
Conclusions
I am taking delivery of 14mm MP (wider than 16mm tested) and this is a must-have lens at those angles. If you think any lesser glass would resolve 4k at super wide angles, you are kidding yourself! Master Prime barely achieves this and it is miles apart from Ultra Prime.
If you are aiming at Red One's or Epic's full raster, medium angles of about 35mm focal length are still in Master Prime territory. The difference is on the edges and corners only, but the difference is clearly visible. Border line, but enough to swing me into selecting Master Prime.
At longer focal lengths of ~75mm, you will save heaps of money by going Ultra Primes. They are razor sharp corner-to-corner, edge-to-edge and, in my opinion, there is absolutely no justification to go Master Prime. Mysterium chip resolving power will be a limiting factor here, certainly not the glass.
It was quick and rough "sanity check", but I hope some of you guys will find the above observations useful.
It was certainly good fun playing with all this glass and I would like to thank Rey and Stefan of Arri Australia for their kind invitation and assistance.
