kye

Great stuff and congratulations!! Let's hope there are many more and that the process gets easier and faster each time!

Great stuff.. and gives me subject matter ideas

Just playing with an idea.... Taking the concepts of resolution and micro-contrast and trying to emulate them in post. This is an attempt to match the look of the Zeiss SS at T1.3 with the Xeen at T4 (one of the sharpest most clinical looking images). Obviously there is no match for the aperture, but in terms of skin softness / naturalness I wonder how well you think I closed the gap? You'll have to zoom to the full images for this one. Zeiss SS at T1.3: My grade: Original Xeen at T4 image: Even if I only got the image half-way that would still mean there's a huge ability to degrade images in post to match between lenses (and different apertures of the same lenses) in any sensibly chosen set.

What people say about the aesthetics of lenses. Starting with the lenses that had more character and less sharpness. Super Speeds: The super speeds are crisp at a 2/2.8 split without being clinical, allowing us to forgo any diffusion in front of the lens. We tested several sets of modern and vintage lenses that all opened up to at least a T1.4. Ultimately, we landed on the Zeiss Super Speeds because they had a cooler, less saturated color rendition and a softer contrast when shooting wide open. I remember when Super Speeds were the cheap lenses, I’d use the Panavision zooms and add a set of Super Speeds because back then they were cheap. Now everybody wants them because they are a little flawed and we want to make this digital cinema look more human. It’s all about the glass today. (talking about Zeiss Masterprimes) BUT: They have no personality, an ugly flare and are ridiculous large. All these things can my Super Speeds much better. Especially the flare is second to none. They are softer wide open (t1.3) but get just as sharp as Ultra Primes at 2.8 and up. They are still decently sharp wide open, just give a smoother appearance rather than crisp. Skin looks many times better on Super Speeds. On a 5K chip, the SSpeeds even wide open deliver an awesome look, almost like all the "defects" are perfectly rendered, thus making them sem alive. The wide-open look that Matt describes is particularly charming, since it combines a slight softening with lower contrast and highlight blooming. Best of all, the effect is somewhat variable between wide open and 2.8, so you have a bit of control if you can light to a certain stop. Cooke S4: We chose Cooke S4s because of their softer attributes and warmer skin tone Well, it's not so much the image being softer, it's the way things go smoothly out with S4 lenses. The play of light and the roundness of its quality, bokeh, the way a lens renders the human face, are all more important from my own personal point of view. The S4s are plenty sharp when they stand on their own, and render a very rich and complex image. I just like the Cooke lenses because they're very sharp, but feel very natural. I like the way they work on faces. they have a softness to them and I think it’s very important with digital cameras not to let the image get too sharp. The other thing is that the Cookes don’t flare very much and we had a lot of practical lights in shot, so they worked very well with our sets chose the Cooke S4s because they are "gentler" than other lenses. I decided on the Cooke S4s because of our multi-racial cast, and the warmth I knew the lenses would bring to their rich skin tones. I couldn't find anyone saying good things about the Xeens. Master Anamorphics: However, once we were on set I could only use the Master Anamorphics because they were so pin-sharp, perfectly straight and square – all the way to the corners – that it was impossible to match them with the other brand. I'm a huge fan of the Master Anamorphics because they don't really look like anamorphics but they have that anamorphic aspect ratio. And they also are one of the fastest anamorphics, you can actually shoot them wide open and still be reasonably sharp. I’m leaving out the Master Anamorphics because those are so clean and crisp it almost defeats the point of shooting anamorphic imo Other comments of note: Yes, in many ways, what's nice about the modern primes is not their sharpness or flare-prevention, it's the fact that they MATCH each other in a set so well, compared to older lens sets which are a lot more quirky in terms of matching in color, contrast, and performance. Hence why I'd be happy to shoot either Primos, Cooke S4's, or Zeiss UltraPrimes.

This leads me to wonder what happens if we look at a Tiffen Black Pro Mist. Observations: No real loss of fine detail (even though I played with it trying to show up the differences), even with the BPM2 It only seems to act on the sharp edges with very high contrast like the edges of the light globe, and edges of the text on the slate. It's also slightly visible on the edges of the colour chart squares. I guess this confirms that it spreads light but doesn't blur, so you'll get less micro-contrast but not zero micro-contrast (which is what blurring is). In theory they're the same, but the similarity is so slight I couldn't really highlight it. I think this, combined with the last post about the lenses, confirms (or at least supports) the idea that the perceived softness of filters like BPM and the famous Zeiss T* are almost independent of resolution, so you can have a lens that looks soft but is also high resolution.

I've had a few, I get bored then move on and it either sits there dead or some technical aspect annoys me (like moderating comments) and so I take it down. Here I figure I can do the same thing but it can remain available for the long-term. Plus if Andrew gets a few hits to the site out of it then that's a side-benefit and helps me give a little back for his work in creating this thing in the first place. Anyway, mid-level detail. This is the same thing, only I'm comparing one copy of the image blurred at a fine-level of detail to another copy blurred to a larger level of detail, so the differences should only be medium-sized details. Observations: Lots less difference between lenses when they're wide-open and stopped-down, which makes sense Dog Schidt, CN-E, Zeiss SS, and Master Prime are all a bit less when wide-open, but the Dog Schidt is really the one with the lowest contrast (even when it's T2 wide open and the Master Prime and SS are T1.3). I think this is an interesting thing because we might be measuring the contrast amount impacted by coatings. Apart from the Dog Schidt and Cooke Anamorphic all lenses seem to be pretty equal at T4. I kind of adjusted this to try and see differences between the lenses with less contrast / vintage look and the more modern ones, but it's fairly subtle.

Fine detail. The method to generate these images is as following: I put each image on the timeline I copied the whole track to a second track, so now every image has a second copy of itself on top I changed the blend mode on the top ones to Difference so now it only shows what is different - black is no difference, white is maximum difference I then blurred the top image with a very small radius blur Then over the top of the whole lot I boosted the brightness hugely in order to make the differences visible What this will do is if there is an image with very few fine details then a blur will do very little and will therefore show very little difference, whereas if there is fine detail that gets blurred that will show a large difference and will be highlighted. You will note that the background is black in all shots (with the exception of the edges of the bokeh which is sharp on many lenses) and that's because the background is out of focus. Conveniently, it also shows us the lens and aperture Images: Observations: Many of the lenses show very little fine detail when wide-open. We should be careful comparing them as their maximum apertures are all different, but it's an interesting thing. The Master Anamorphic is interesting because it shows some fine-detail when wide-open and not so much at T4, so it's a more consistent performer, unlike the CP.2 which has a significantly different level of detail when stopped down. The Cooke Anamorphic is also on the more consistent end of the spectrum. The Dog Schidt (Helios with coatings removed) remains at the lower end even when stopped down, although the focal plane appears to be slightly further back on the T4 so take that into account. The Cooke S4 on the other hand is at the sharper end of the spectrum both at T2 and T4.

I grabbed a few more stills from the first cine lens test to analyse. Let's start with contrast, the setup would have been meticulously controlled for lighting and exposure so we should just be able to analyse the file directly. Blue is lowest luma values and red is highest, which conveniently also highlights the lens and aperture Observations: The Dog Schidt has much lower contrast than the others - this is by design as it's a Helios with all coatings removed in order to make it flare radically The Takumar also has low contrast, which I think the Takumars are known for The highlight from the practical just to our right of the model gets dimmer when lenses are wide open, but that's just blur distributing the luminance, not a real contrast difference Contrast doesn't seem to increase when stopped down Beyond that I'm not sure I learned anything here. Next up is micro-detail.

Nice! .........and drive it to the store to buy all the Meike lenses!

Voigtlander 42.5mm f0.95 vs Veydra 50mm T2.2...... The Veydra has excellent performance: However, the right comparison to that is the Voigtlander at T2.2, which is part-way between these two curves: So, what we have is: The Veydra is the clear winner in the edges but the closer you go to the centre the lower its lead gets The Voigtlander can open up two extra stops (with declining performance of course) should you want/need that Assuming the Meike performs the same as the Veydra, it's significantly cheaper, but the Veydras are roughly similar Now, the question is .....does resolution matter? My observations about the Xeen seem to suggest that it doesn't. The Xeen is terrible wide open, but stops down fine: Of course, the Voigtlander at T1.4 kills the Xeen at T1.5 everywhere except the last little bit at the edge of the frame (remember that the Xeen is plotted across a wider scale), and the Xeen looked fine in the test above when stopped down. I don't have an equivalent MTF of the Voigt at T4 so can't compare, but here's one that shows that it keeps improving from T2.8 to T4: In fact it's gone past 'Excellent' and way off the scale! lol. There are other factors to consider as well, of course. Like colour rendering - No idea about the 42.5mm Voigt but my 17.5mm goes quite magenta when opening up the last two stops. Yes, these all lock me into MFT, although FF is just a fad, so I'm not concerned! I am intrigued to hear why it doesn't though - unless it involves using a TC to expand the image circle back to cover FF ??? What are "bi-nature rendering characteristics"??? There's nothing wrong with filming brick walls in preparation for shooting a film that will feature brick walls that for some reason must have them without any distortion, but in preparing for any other type of film, they're of almost no use whatsoever. Of course, for many of those shooting footage of brick walls, they get confused when you get to the part about "preparing to shoot" as typically the only thing they do shoot are lens tests. The Veydras on ebay I saw weren't that well priced, assuming that Duclos have stock (https://www.ducloslenses.com/collections/veydra-primes) and if the Meike really are the same as Veydras then they're really overpriced! I started with the tech as it was the first question in my mind, but I'm already digesting this and working out where to go next. The images in that review were just great! I do wonder about the curvature though, and wether or not it would just simulated in post with a blur or softening filter around the edges of the frame, plus some vignetting.

So, let's summarise. Lenses (*that I found the data for) tend to lose resolution away from the centre and have varying degrees of field curvature. They get better when you stop down, and even the best lenses wide-open aren't as good as cheap-as-chips lenses stopped down by even a single stop. (* although I couldn't find MTF charts for the Zeiss Master Anamorphic I found comments that the measurements are very impressive and they're sharp from edge-to-edge) Some of the classic lenses measure poorly (eg, Xeen and Zeiss SuperSpeeds) but visually they are not radically different to the best lenses. I'm not going to say that resolution doesn't matter, but what I am going to say is this.. I think the absolute levels of resolution probably don't matter much, but the overall character of it may, for example how the lens goes from centre-to-edge and what kind of issues (like astigmatism) it has along the way. Besides, even if it did matter, there are options around that do a very passable job. Many of the MFT ~40mm lenses were amongst the cine lenses in performance and characteristics, and the Veydras (when on MFT) out-resolve the CP.3s (on any sensor) and slaughter the CP.2s.

Let's look at what's on the table for ~40mm then. This article compares a number of ~40mm MFT lenses: https://www.lensrentals.com/blog/2018/03/finally-some-more-m43-mtf-testing-are-the-40s-fabulous/ The Voigtlander performed very well, and despite being the worst performer wide-open (which is essentially an un-fair test), was comparable when stopped down to 1.4 which was wide-open for the rest of the pack, it was close to the top of the pack at F2.8, and drew for first place in how flat the field curvature was. I think this bears out the idea that this lens is around the performance of the other high-performance lenses at any given aperture, but can simply open up further than the other lenses. From the commentary in the article I think this isn't something that can be assumed however, so isn't a general rule. This article looks at the Konica Hexanon: https://erphotoreview.com/wordpress/?p=1902 The MTF50 shows that there is an absolutely huge difference when you stop down a couple of stops. This bears out in my experience.

This is the Zeiss CP2 wide-open at T2.1: This is the Xeen wide-open at T1.5: This is the Zeiss CP2 at T4: This is the Xeen at T4: To my eye, both look soft wide open and both look great at T4. Here's their MTF charts: The Xeen is quite obviously far worse than the CP.2 when wide open, yet the test shots appear similar to me. Or, at least, the difference in focal distance between them could account for the entire difference. Relating this back to myself, I find focus to be a hit-and-miss affair so there's not much point nit-picking at something I will almost never completely nail. Stopped down they both improve vastly, both in the charts as well as in the images. In theory the CP.2 has way worse performance due to much greater field curvature, but there's nothing in frame in the test shots, so who knows. I think the Xeen was the worst performing lens wide-open that I saw in the above articles, so not being able to see it in the test images is interesting. To illustrate a point, here's the $40,000 Zeiss Master Anamorphic wide open at T1.9: Not so high resolution either, although I couldn't find an MTF for this lens, the resolution of the others (at least in the centre where we actually have something in focus) is better at T4 than this lens is at T1.9, so this is simply a fact of life. What it does mean, however, is that wide-open resolution isn't the focus, it's resolution at apertures that you will shoot. Just for fun, here's a modified and re-branded Helios 58mm F2 at T4: It obviously has less contrast (the mod removed all the coatings to deliberately make it flare - which it does spectacularly) but the resolution is also there.

Let's look at resolution. Here's a bunch of resolution tests of the Zeiss CP.2, Zeiss CP.2 SuperSpeeds, Zeiss CP.3, and Zeiss Supreme Prime series. https://www.lensrentals.com/blog/2019/06/just-the-cinema-mtf-charts-zeiss-cine-lenses/ If you don't know how to read and MTF chart then go read some thorough tutorials and come back - warning: hanging with the brick-wall-photographers that think MTF charts are the scoreboard of their egos will rot your brain, consider yourself warned. My notes from the above article on the Zeiss lenses are that: With the exception of a few lenses (which are mostly wide or long) their centre resolution is typically between ~0.93 and ~0.5 when wide open Edge resolution typically drops off quite significantly and also may go through complex curves on the way (likely due to field curvature, but that wasn't covered in the article) The CP.3 series is typically much better than the CP.2 series, and the Super Speeds, which are a classic, are worse, although they're also at a wider aperture, which is a nice segue to..... This article looks at what happens when you stop down, and includes some other cine lenses too (Canon CN-E, Sigma Cine, Rokinon Xeen) https://www.lensrentals.com/blog/2017/11/testing-lenses-stop-downed-mtf-curves/ From this we learn that: Lenses sharpen up when stopped down... I mean, who knew! Edge resolution is still pretty poor and the Sigma looked completely rubbish (once again, it looks like field curvature) Lots of astigmatisms happening off-axis This article talks about best-MTF curves: https://www.lensrentals.com/blog/2017/11/testing-lenses-best-individual-focus-mtf-curves/ This points out that these lenses are capable of very good performance across most of their image circle, so the issues in previous articles are to do with field curvature. So, let's talk about field curvature and taking the average best performance: https://www.lensrentals.com/blog/2017/11/testing-lenses-finding-the-best-average-focus-point/ That basically says that if you focus so that the most of the subject is in the best focus then it'll be better than making the centre point completely in focus and not caring about the rest of the image. For a change of pace, let's look at the Fujinon MK 18-55 T2.9 lens: https://www.lensrentals.com/blog/2017/02/first-impressions-of-the-new-fujinon-mk-18-55mm-t2-9-cine-lens/ It's only a first impression and the MTF charts only show it wide open, but it looks like the other cine lenses with edge performance dropping off. This article shows that the Veydra Cine Mini Prime lenses are very good lenses with great resolution and very little field curvature, and they absolutely kill the CP.2 for resolution. https://www.lensrentals.com/blog/2018/03/veydra-cine-mini-prime-mtf-optical-bench-tests/ Here's one showing the Rokinon / Samyang stills and cine lenses: https://www.lensrentals.com/blog/2019/04/just-mtf-charts-rokinon-samyang-lenses/ ....and finally here's an article about testing cine lenses, why they didn't do it in the past and why they started. It's interesting because it speaks to the concerns of cinematographers vs photographers but also brings in the topic of resolution and how when you've got a bunch of lenses that all looked sharp before and you increase the resolution some of them will still look sharp and others won't. Fun stuff. https://www.lensrentals.com/blog/2017/07/why-were-going-to-start-testing-cinema-lenses-and-why-we-havent-before/ Something that's worth pointing out is that not all lenses do get significantly worse towards the edges of the frame. While rare, lenses like the Nikkor 85mm f1.4 and f1.9 lenses are almost flat between their centre and their edge: https://www.lensrentals.com/blog/2019/04/just-mtf-charts-nikon-prime-lenses/ After all this talk of resolution, I haven't yet talked about if this actually matters. I'll get there, I promise.

This thread is mostly for me, but I figured that I'd post it instead of just making my own notes as it might be of casual interest to someone. Read at your own peril - this will likely fall down every rabbit-hole available. Still with me? Ok, to catch up to where I am, watch and study these videos.... If you're not lusting after the $40,000 Zeiss Master Anamorphic after watching the first one then go watch them all again before proceeding. I've just sold my car (wasn't using it) so more money for lenses! I use a GH5 and own the Voigtlander 17.5mm f0.95 which I really like, and I've settled on having a super-wide ~8mm and probably ~40mm lengths. The Voigt is pretty rubbish wide-open, but when compared to F2 lenses the Voigt is already stopped-down two stops when they're wide open, so if you're stopping it down a bit then there's real advantages, plus if I ever need the extra exposure (I shoot in available light so that happens sometimes) or want to have the shallow DoF (which also happens) then I can do that. I own the SLR Magic 8mm F4 which is an ok lens, isn't that great optically but isn't terrible either, but it's a drone lens so the ergonomics are completely rubbish. I also own both the Konica Hexanon 40mm f1.8 and Helios 58mm F2 plus an m42 SB. I am not a fan of the swirly bokeh of the Helios (although using it on MFT gets rid of the majority of it on the edges of the frame), and I used the 40mm wide-open on a recent trip and it did not perform well, to put it mildly. Soft as hell as well as all sorts of nasty aberrations. The Voigtlander 42.5mm f0.95 would be stopped down two stops when rendering the same image, so that's something to consider. I've previously identified that I wanted the Laowa 7.5mm F2 lens and was happy with the Konica, but after the Konica had its minor meltdown I'm now wondering if I should replace it with the Voigtlander, or what I should do. Speaking with some pros on other forums about lenses, and especially the guys that know what they're doing in post, the advice was to capture the highest resolution image you can and degrade it in post, or use something like a Black Pro Mist filter. Before I spend any real money on some lenses I figured doing some real research and working out what I really want might be in order, thus this line of thinking and this thread. Next step is to study cine lenses and see what they are actually doing, even just as a reference point.

Looks good @BTM_Pix.... (the stabilisation, not the inclement weather!) I'd also suggest that if it makes that much difference then you can probably get more performance by adjusting it further. I did a series of stabilisation rig tests by having a path that I would walk through the house, so I'd make a change, do a circuit, make another change, do another circuit, etc. Then sit down and review the footage. Every clip began facing a mirror so it was obvious what setup was being tested in that clip.

Ah, well, if we're comparing RAW to compressed anything then it's a different ballgame! In terms of that highlight recovery - I'm not convinced. More likely the camera was shooting the first thing but the ISO settings that the camera recorded in metadata would have blown the image all to hell, then the NLE gets the file, applies the ISO blowing the highlights well and truly, then you adjust them down and voila - the signal that was recorded. This protects against camera operator error, but not if you actually blow out the image. Try shooting in full sun wide open without an ND at base ISO then try and "recover" the file... 12-bit won't save that!!

I tested it in real life, filming my own test footage, including pushing and pulling heavily in-post as well as doing only light grading, but the differences really weren't that much. I'm curious to hear what that difference is actually useful for, not just a subjective impression. Oh, and did you test the Canon 12-bit against Canon 10-bit? Otherwise we're introducing other factors into the question.

I'm curious how much extra you feel that 12-bit gives you over 10-bit? I definitely agree that 10-bit is a decent jump up from 8-bit, but when I was playing with ML I did a bunch of tests comparing 10, 12 and 14-bit RAW and didn't really see any useful differences above 10-bit. I'm not saying there was none, but I can't really think of when those differences would really matter.

Wasn't it meant to somehow integrate with their cinema cameras? Or be able to take their accessories? I recall it had some kind of accessory port and the implication was that you could take the RED accessories and rig up the phone, either so that the phone was attached to a camera or that the phone was the camera but with extra bits attached. So much of the promise was that it wasn't really a phone, but would actually be modular in some way? What was the port actually for?

I saw in a recent YT video by Mark Bone he mentions that he's using his monitor as a light source for the video. What CRI is a monitor likely to have if it's been calibrated and is of reasonable quality? Does the fact that it can produce the entire sRGB spectrum (or even a wider gamut) mean that the panel in it has a high CRI? I'm curious because I do camera tests every now and then and would like a powerful and high CRI light source, and if I already own a really large one then that would be very very handy!!

I'd be interested in reading about it. Partly because with all the hype, cryptic teaser media releases, and my general lack of interest, I still don't actually know what it is or what it does, apart from underwhelm people!

Same reason that they're not all screaming at Canon for lack of un-cropped 4K... Every brand has weaknesses, to expect otherwise is naive. I love it when Canon moves slightly, all the people that believe in pixies, santa, the perfect camera, unicorns, the company not driven by profit, and other mythical creatures all appear and emote all over the internet!!

Nice! Get a GH5, some half-decent lenses, and a smoke machine and you're half-way there already

Personally, I thought The Breakup with Vince Vaughn and Jennifer Aniston was so true to life that it was terrifying, but I somehow get the impression that's not what you're asking for.......