(Article was originally written for film-making forum DVXuser.com: If you do not see images, you may want to register with their website. In the future, I'll try and get the blog posts independent of the forum.)
In anticipation for an upcoming feature film, another cinematographer and myself had a little fun and did a lens test to see just how well matched the Cooke Panchro/i lenses were to the very popular Cooke S4/i's. The feature is looking to use a set of S4's and Panchro/i's on a two camera shoot.
Cooke S4's have been used on films and television shows such as: Entourage, Lie to Me, V for Vendetta, Erin Brockovich, The DaVinci Code, The Assassination of Jesse James, most of the Harry Potter series, Munich, Mr. and Mrs. Smith, Live Free Die Hard, Jarhead, Kingdom of Heaven, Ray, The Interpreter, The Informant, The Illusionist, Hulk, Frost-Nixon, District 9, Chicago, Cinderella Man, Casino Royale, Capote, Brokeback Mountain, Brothers, A Beautiful Mind, O' Brother Where Art Thou, The Village, House of Sand and Fog, and many many more.
So we were interested in seeing how much the Cooke S4/i's little brothers, the new Cooke Panchro/i's, looked in comparison... especially because Cooke has always said the Cooke Panchro/i, S4/i, and 5/i lenses are color matched to each other. Cooke S4 lenses are also commonly considered one of the best color matched lens sets in the top professional level. This test may come to some as a no-brainer as to the result. I was told by a friend in the ASC, that cinematographer Robert Richardson, ASC who just shot Martin Scorsese's Hugo Cabret in 3D on some Arri Alexas and the new Cooke 5/i lenses, also had sets of Cooke S4/i's and Panchro/i's on the production. I would never doubt Bob Richardson, but I had to do the test myself!
So we set up a cinematographer's still life as a test shoot. We were simply testing the similarities of contrast, color, and similar qualities. If we were to use S4/i's and Panchro/i's on a two-camera feature, we needed to see how they varied juxtaposed to each other. Last thing any cinematographer wants is to have to go through a DI correcting things that ought to have been done right in camera. Large shifts in color reproduction and contrast between lens sets can do this. With that said, we were not doing these tests to compare resolving power. We already know from plenty of experience that all modern Cooke lenses are by-and-far sharp and it wasn't a concern if one was more than the other. If we ever feel like exploring this, we will get some proper charts and go about it in a more scientific way.
As for the variables:
Camera: RedOneMX
Lighting: Arri Tungsten 1k Fresnels, bounced. Controlled environment.
Color Temperature: 3200*k
Iso: 800
Shutter: 180*
Format: 4k, 16:9, RedCode 36
Distance from camera to subject stayed the same for every lens.
Lenses at our disposal:
Cooke S4/i: 50mm, 75mm, 100mm, and 135mm
Cooke Panchro/i: 18mm, 25mm, 32mm, 50mm, 75mm, 100mm
Cooke S4/i CXX zoom: 15mm - 40mm
I'll discuss my findings after I post the images. These images have not been touched in anyway. They are screen captures taken from RedCine-X and received absolutely no color grading. The .tiff files were then imported to Photoshop where I layered the correlating images upon each other and made my 'checkered' comparison still. All steps to preserve and maintain the quality/integrity of these results was taken.
Checkered comparison stills is something I created to give more comparison points of reference. Two images are overlayed, and instead of the image being divided in half, the image is divided into quarters. The Upper-left, and lower-right sections are one image, while the other two areas are the other. This allows two boundaries per image to be juxtaposed with the other. If anyone wishes to download the original RedCine-X .tiffs, please let me know. They will lack the checkered system of comparison, but will be the same quality.
Three Cooke Panchro/i focal lengths are compared to the Cooke S4/i CXX zoom lens, and the other three Cooke Panchro/i focal lengths are compared to the Cooke S4/i primes.
Results:For the reputation Cooke has, I went into the test optimistically. When I took a look at the results, my expectations were surpassed. To put it simply, the Cooke Panchro/i's almost
exactly matched the S4/i's. I didn't think two completely different sets of lenses would match as if they were the same. But that is what Cooke went out to do when designing the smaller Panchro/i's and that is basically what they did.
Under heavy examination, (clicking back and forth between two 4k 16bit .tiff images in rapid succession) I found the S4's had an almost undetectable amount of extra contrast... I really only noticed it in the darkest darks. The largest discrepancy between any of the lenses was the Cooke CXX. The zoom lens was warmer than the matching Panchro/i and S4/i lenses. It wasn't horrible, but enough that if I wanted perfection and wanted to avoid a costly DI session to hunt all the CXX footage down, I would make a very small compensation in camera. My $.02.
Out of the Panchro/i range, the 25, 32, 50, 75, 100 are exactly color matched to each other. However, the 100mm Panchro/i is a nano-bit more warm than the rest, noticeable perhaps only at the most scrutinized examination. Within the Panchro/i set, the 5-set mentioned above is matched, no doubt. However the 18mm was a bit different. Although the 18mm Panchro/i is very close to the others, it was noticeably warmer in comparison. I'm not sure if it would be enough that anyone would notice, but if being juxtaposed multiple times to the other focal lengths, and I was aiming for perfection in-camera, I might consider bumping the CT a couple of points. The 18mm Panchro/i is certainly not as warm as the CXX zoom, but was the greatest inconsistency between any of the Cooke Panchro/i and Cooke S4/i lenses had to each other. With that said, if that was the biggest issue, it's really a non-issue. Take a look at the Panchro/i 18mm compared to the other focal lengths in the Panchro/i set in the
FoV chart. The FoV chart is actually made up from all the Panchro screen grabs. They are simply resized and stacked upon each other. You can see how warm the 18mm is, and how close all the other primes are to each other in the Panchro/i set. The 18mm is the only lens in the set that stands out as not being almost exact to the others. I think knowing this, a bump perhaps of 300*k might clear this up. I didn't test how much it would take, but it's not terribly much.
18mm Panchro/i v. 18mm CXX:
The CXX zoom lens does not have lines accompanying the focal lengths on the barrel, so I got the lens centered on the 18mm focal length engraved number, and that was that. Getting the CXX on 18mm the best I could, it should be mentioned that it was not as wide as the 18mm Panchro/i prime. The 18mm Panchro/i is warmer than it's fellow Panchro/i brothers, but still is not as warm as the CXX. The way the room was rendered as far as perspective is almost exactly the same between the two lenses. However despite FoV being similar between CXX 25mm and 32mm compared to the Panchro/i's, the distortion of how the room is rendered changes. The Panchro/i's will render a more realistic 'flat' field, while the CXX makes the room look a little deeper because the sides continue to have a small warp. It's not aweful, but there is a difference. However, the 18mm lenses match. The CXX is warm compared to the S4/i and Panchro/i's, but since the 18mm Panchro/i is a warm too (but not as much) the warmth is best seen in the 25mm or 35mm panchro/i comparison.
25mm/32mm Panchro/i v. 25mm/32mm CXX:The 25mm/35mm Panchro/i falls inline with the rest of the Panchro/i series and S4/i primes in regards of color rendition. Thus it is easier to see the Cxx warmth in these checkered images. I also noticed a small discrepancy between how they portray depth and the field. The edges of the CXX seem to be 'closer' while the middle seems to recede. The Panchro/i creates a realistic flat field, while the CXX has a slightly more three dimensional feel from some image barreling. The very edges of the screen seem to 'stretch' around the lens. I would probably have never noticed it without direct comparison to primes. Download the 25/32cxx and 25/32panchro and toggle them back and forth in some sort of picture viewer. It's quite interesting.
50mm/75mm Panchro/i v. 50mm/75mm S4/i:It was the 50mm and 75mm Panchro/i vs. S4/i comparisons that really blew my mind. If I hadn't taken so much care labeling the digital frame grabs, I would have easily lost track of which was which. These lenses are completely interchangeable and even under demanding scrutiny, the differences will probably escape the trained eye. The only way I could see a difference was by toggling between two 16bit 4k .tiff images over and over. I eventually began to see the S4/i's possessed a small bit of additional contrast, and the Cooke Panchro's had maybe a nano-bit of warmness? I still am not sure if that is true or if I'm just trying to see things. Regardless, these lenses are so close to each other, their tolerances are probably equal to the tolerances between two lenses of the same set and same focal length. Perfect match.
100mm Panchro/i v. 100mm S4/i: The Panchro/i 100mm and S4/i 100mm are
very closely matched. Not exactly matched, like the 50mm and 75mm demonstrate, but because the Panchro/i 100mm is a very small bit warmer, I must mention it. However, this is nowhere near as warm as the CXX and I don't believe it's far enough to correct for it, even on the smallest level. In fact, I think it's slight enough that it's actually nice. Considering I rarely use the 100mm lens for regular shooting and when I do, it is often for a very long-lens beauty shot, I welcome the little nuance... even if it doesn't comes across on screen.
Conclusion:It's simply amazing how much the Panchro/i's did compare to the S4/i's. It was much more that I required. The CXX zoom is a fantastic lens, and it's a difficult lens to make being a 15mm-40mm zoom and all. However, it does need to be cooled down in order to match the S4/i and Panchro/i's. Just by a small amount. It also has some barrel effects mostly seen when it's longer focal lengths are compared to the primes. Regarding the Panchro/i and S4/i's, I would not think twice about using these lenses in a two camera shoot, but not just for different units, but even for mother-daughter or cross coverage shooting. The CXX I would probably tell my AC's to make sure they bumped the CT a couple hundred k's when it goes on. Although not to the same degree, the Panchro/i 18mm could use a small bump in CT too, maybe as little as 100 or 200 kelvin from my tests so far. Otherwise, it's all gravy.
***Disclaimer***By now you have probably noticed how carefully I've been trying to choose my words when describing these findings. The differences I describe are almost all nuances and differences almost unnoticeable by the naked eye. I highly recommend downloading the images which I have taken so much time to compile before interpretations of my words become fact. I tried to be very thorough and detailed, and I don't want people thinking the differences I've mentioned are larger than they are, or smaller than they are. Please take a look for yourselves!
After writing the first review, I received a concern from one individual who had deeply examined the images. He was concerned about aberration in the lenses, so I posted his main concerns here, and then investigated myself. The following is his post.
=================================================
=================================================
Hmmm.... is it just me or do they fringe (chromatic aberration) like mad in the out-of-focus areas?
Look at the book letters. Looks like transverse chromatic aberration to me.
And what is going on in that pistol? Random colored dots, compression artifacts, moire?
No offense here, but I´m surprised, I thought that a RED with Cooke glass would do way better.
- Frank
That still camera has a strange blue halo, the letters are fringing and something funky is going on in that pistol grip.
Looks like chromatic aberration to me, but who knows, maybe it´s the chip, Bayern pattern and codec.
Click on the pictures to enlarge and see the effects in full glory.
=================================================
=================================================
My response to comment regarding aberration concerns:
I would be happy to examine these things you are discussing, but I must say it would be very very helpful if you would give a reference to which images you are examining. (Being so cropped, it could be many of the images, if not specified)
With that said, I never went through these images thoroughly for aberration, as this test was for color matching and contrast. However, I have looked into it, and you have found some very interesting things worth mentioning.
Out of the examples you gave, and I mean no offense, I only think the image of the Yashica camera is worth examining. I would be happy to take a look at the other examples should you give me an image you are looking at. So far those other images look like very very zoomed in images portraying what looks like random noise in the image. It's kinda funny that it appears the arrows you added to the image has spawned more of the noise around them. Perhaps a culprit of compression, the camera sensor, who knows. I don't think this has anything to do with the lens, quite frankly, but I would be happy to revisit it if I find this is more than just super enlarged imaging.
Ok, so back to the Yashica camera. From the image you posted, I must say you have found what I consider an usual suspect of chromatic aberration. I don't know what you were talking about, but I assume it's the same. Anyhow, despite not specifying which image you pulled the still from, luckily the Yashica camera is one of the outer-most props on the table, isolating your image from either the 18mm or 25mm FoV's. It should be noted that
chromatic aberration is often found in focal lengths with wide field of views and near the peripheral of that FoV. Thus it seems our culprit fits the bill. Considering the above.
So let's examine two areas of the images, first with the Yashica camera. I took the 8bit .jpg images available in this thread, enlarged greatly, and used mac's Shift+Apple+4 to screen capture a .png file. I assume this is not a loss-less process, but chromatic aberration cannot be caused by this process, so please, for now, disregard what may have been an increase in pixel compression. If we want to explore that at another time, I'll do it correctly.
Please remember that the two enlarged images following the one below, are just portions of this 4k... please note the scale of the object as to understand how much chromatic aberration is taking place. The Yashica camera has been colored Red to show you how much we are zooming into the still frame from the video. If my math is right, the full image is about 32x larger than the portion we are about to examine!
18mm FoV checkered image with Yashica camera size noted:
Enlarged from Cooke CXX T/2 15-40mm zoom lens at 18mm T/2.8:
Enlarged from Cooke Panchro/i 18mm T/2.8 wide open:
25mm
**Next is the 25mm focal lengths. The following image is the FoV of the 25mm, with the Yashica camera highlighted in red, so you can understand the scale of how zoomed in we are looking and the offset of the chroma aberration.
**
Enlarged from Cooke CXX T/2 15-40mm zoom lens at 25mm T/2.8:
Enlarged from Cooke Panchro/i 25mm T/2.8 wide open:
Conclusion from Yashica camera analysis:
(as it happens, if you only looked at the 'checkered' comparison files I offered in this thread, the Yashica camera was always in the lower right quadrant, which was always designated as the CXX. So I suppose it would have been impossible to see how the Panchro/i's handled aberration on the Yashica, unless you explored their .jpg stand-alone image.)
Postmaster has brought to light some images which have spurred me to dig deep into these images. It appears that the CXX-zoom does have
considerably more aberration than the Panchro/i's when you really explore the image. This was a little shocking, as far as how much the CXX produced but upon further contemplation, it wasn't as surprising. As mentioned above, chromatic aberration is the offset of wavelengths of light (in most cases the blue-violet wavelengths (the shortest in our visible spectrum)) and lenses succumb to aberration increasingly so when bending light from very off-parallel angles, such as the peripheral of wide FoV's. Also, I believe that the more optical elements within a lens, the more difficult it becomes to properly bend the wavelengths of light correctly, meaning it's no surprise to me that an ultra wide 2.67:1 zoom (starting at 15mm and T/2) has more aberration than a prime lens which has considerably less optical elements. Looking at the CXX aberration now, it's definitely more than most DP's would like to see, but given the fact it's a 15mm-40mm zoom lens, the complexity of the lens and how it exists in the realm of chromatic aberration optimal conditions, I'd be interested to know if anyone could find a better ultra wide-angle zoom lens with no aberration. I guess that's being fair to a $38,500 zoom (or so says ebay).
Now examining the 25mm, we still see the CXX has much more aberration than the Panchro/i. The Panchro/i 25mm does pick up a little, because now, with the tighter focal length, the Yashica camera is on the very very edge of the FoV. This means the light coming from the Yashica camera is entering the lens at the most extreme angle opposed to the parallel light entering from straight ahead. Thus the light rays are bent to near perfection. It's there on the very fringe, but look at the variables. We are unforgiving zoomed into a 4k file by over 32x magnification on the edges of a wide lens. At this level, I'm certainly not concerned nor do I think they are performing outside the tolerances of what we expect from top pro-lenses.
Sadly, we did not have any S4/i primes in the wide variety, so I cannot test those results. Just the CXX v. Panchro/i.
It's not perfect, but I don't think they are fringing like mad, not the Panchro/i's certainly, and not quite the CXX. I do agree, now that I have examined the CXX, that it does have considerably more aberration than it's Cooke panchro/i brothers, but as mentioned, the lens is practically built to be prone to aberration, I consider the amount there is as a valiant battle of suppression. Show me a 15mm
zoom lens that does better, and perhaps I'll start to shame Cooke and their CXX. There's definitely some there mate. I give you that.
The original review pointed out the CXX had a slightly less 'flat-field' perspective. Probably also why it has more aberration than the primes? I bet so.
What do you think? Take a look at the 'Arri' hat, and you'll see the same results.
Take a look at the Arri hat, which at 32mm is VERY close to the edge, which should theoretically be most prone to aberration, yes?
32mm FoV with approx. designated zoom area identified for reference of magnification:
Cooke CXX 15-40mm T/2 zoom @ 32mm T/2.8:
Cooke Panchro/i 32mm T/2.8 Wide Open:
