Axel

I often admitted that I can't grade. That's true, though I technically understand everything Huczek did in his tut and think I could have done even better. I agree that the key is not as good as it could have been. BUT: The difficulties start when you don't grade a single, isolated shot so that it looks much better than before, but a whole sequence, in which almost every shot presents different problems. Yeah, Resolve has this very easy-to-use wipe-splitscreen for saved grades. Now the way to make the sequence look "consistent" (term by Stu Maschwitz) is to grade your favorite shot from the sequence first, for look. Then you compare the rest of the shots one by one with the splitscreen. Not just with the naked eye, you watch the RGB parade, which also shows the split and indicates if the balance matches. So far the theory, and by following this advice, a half-blind will get 90% of the task accomplished. The remaining 10% separate the boys from the men. Because then you balance without net, and you will just need "an eye" for it, much like a good cook needs "taste" not to spoil everything but instead make it special, sensational ... Two more aspects to consider: the whole philosophy of skin tones which need to be preserved because they are "memory colors" (Maschwitz again) is arguably wrong, the vectorscopes' skintone-line just a crook. If the skin color is perfect "porange" (Maschwitz) doesn't mean that it's looking alive and beautiful. Because that - second objection - has more to do with the richness of nuances, with the number of different shades and colors the skin is composed of while it was stored in the recording file. Who says that colors are more believable if they are "pure"? There is a reason why the classic painters didn't just smear some porange on their portrait sketches. They applied different shades in layers, they let shadows wash away the saturation, they let the skin reflect the light or nearby objects. You could say they did a lot of grading. To get beautiful and not just acceptable skin, you need the right light/exposure, the right WB, the right color science of your camera's profile. The best options offer 10-bit or raw. And you need that extra bit of delicate taste for colors that will never be taught in a tut. EDIT: On the Larry Jordan tut. This is not grading, it's saving a shot with a terrible cast through color correction. We don't see the girl full screen, but we will all agree ("everybody knows the color of human skin") that the color balance looks natural. It is an SD-approach actually. In SD it was enough when the outlines were filled with the right color. For true HD, let alone UHD, this doesn't suffice anymore. We are able to see so many details that the chroma values can't just represent skin. It's like a jump from a children's coloring book to a Vermeer.

Don't expect 4k to be a revelation, but it is noticably cleaner (particularly for HD as target) than 1080. This and the external monitor: a no-brainer. If you feel your setup needs an external monitor ...

I actually had the Ursa Mini in my hands at IBC, and it feels gooood ... As far as my experiences with BM are concerned, I only had the Pocket (and once borrowed the BMPC. Now I have sold the Pocket, because I gave up operating camera altogether and rely on a friends' skills. Invested in A7s, Shogun, lenses, Ronin. Tests are running since two months now, but right now the results are at best, er, mixed in comparison to the Pocket. Too many parameters to check, almost no chance to grade this stuff (the way I was used to), and very poor skintones. Would buy the UM, but unfortunately my friend just hates BM. I think there is no need to preorder. First periods of shipping they will be delivered with terrible delays anyway, in a few months you'll get them within days.

Former owner of 7D, used 5Dii and 5Diii a lot, former owner of GH2 and BMPCC, now using A7s with or without Shogun, having seen a lot of reviews on GH4 and the competition, I feel fit to draw some very general conclusions. First of all: colors. You may not know now, but with a Canon DSLR you already have the best color palette for 8-bit currently available. Very recently, there was this thread about an Ursa Mini clip, and everybody raved about the colors. And for a reason. The colors are rich. But then watch this ancient clip, and you will notice that despite the fact that the skin colors seldom touch the ominous "skinline", they are all believable and beautiful. In fact, to get close with any other non-Canon camera is hard, unless you shoot 10-bit or raw. I am very sceptic about using flat picture styles like Slog or V-log. If you are limited to 8-bit: forget that, is my advice. Reverie shows poor resolution and poor DR, for today's expectations. Without knowing your specific camera, I guess that's true for the 600D as well. Now, without Slog, the A7s can't deliver the theoretic 14 stops. If the GH4 with V-log and external 10-bit can show it's full 12 stops (and if glorious colors can be graded from that) still remains to be seen, imo. The current tests/demos don't impress me much. Resolution-wise, many cameras top your 600D. 4k of course is a better HD, but right now I feel the A7s' HD resolution to be sufficient to decide against the Shogun (if that meant the setup became to cumbersome for run&gun). If there ever will be a hack for this camera, and they had to decide between 4k and 10-bit internally, I hope they choose the latter.

Funny, I am trying to get the answer for the exact same question. First of all, resolution. The A7s' resolution is fine. I can further improve it with the Shogun, but as I wrote, the resolution should suffice. Then the colors and gradability. The colors are few (8-bit). In theory, the A7s can depict almost 15 stops, more than even the 4,6 Ursa Mini. But only at the native ISO of 3200 (that's why this is the default ISO for S-Log). S-Log fits to landscapes, but imo not for skintones (too few to make them look alive). So until further experiences, I stick to autumn leaves then, shoot only at ISO 3200, use only WB 3200, 5600 and 6500. I need to buy fixed ND filters, because with such high ISO, ND faders all destroy the colors. I'm not finished with my evaluation, these are my initial findings. I hope I can get close to this look ...

You are right. I wrote 'tasteful teal/cyan' (it should have read orange/teal of course) because it isn't Transformers. I think most of the enthusiastic reactions to the clip come from the two things a subtle o/t grading does: 1. the image looks balanced, harmonic. 2. the image looks clear because it's the highest hue contrast combined with luma contrast possible without looking outright wrong. But even if a lut can do wonders, my own experience with ProRes of the Pocket tells me that you always need to apply corrections, if only for gamma. Our Mr Wedding may be a laughing stock and make embarrassingly cheesy videos, but he must be brashly clever, if he did nothing but produced such nice images. So maybe everybody else makes too big a fuss about grading, log, finding the right profiles, ask for the right workflows. All you need is the right lut ...

I did a lot of weddings. You may detest them, but you hardly find better occasions to test your equipment and your skills - as well as videographer as as storyteller. A wedding is as good a proof of a camera's strengths and shortcomings as any image video or fictional short, imo.

You can see clearly that it's graded. It's a very tasteful teal/cyan grading. I'm curious if he used Resolve.

After I overcame my first enthusiasm, I took a look at the former work of this wedding videographer, which was shot with BMCC and BMPCC. What we have here is a guy who is a skilled colorist. The resolution is good, the DR is good (but not sensational, he always crushes the blacks!), but what really shines is the guys' taste in finding just the right colors. Though personally I'm convinced that the camera is worth every cent of it's price (and feels good in the hands, held it on IBC on monday), I don't think this particular clip proves anything. Meanwhile, there is a UHD clip by the same man:

4k-internal and 5-axis stabilization? A no-brainer compared to the predecessor, except for the higher price.

Ever since the first "gamechanger" threads on handheld gimbals emerged, I remained sceptic. Particularly the effect of travelling to, away from or around inanimate objects (cars, trees) is something that wears off quickly, more so if done with ultra wide lenses. This is true for Glidecam/Flycam/Steadicam shots as well, as long as they don't serve a well-defined purpose. Similar with extreme sDoF that was en vogue for a few years. But wouldn't you agree that it's nice to have these opportunities in our tool box? We also made some shots with a 70mm, and they were smooth too, something you'd hardly try with a Glidecam. Didn't try longer lenses so far. Very few nowadays explore what tele lenses have to offer. Of course these things shouldn't be misused to just show off. Watch this clip on 'Bayhem' from everyframeapainting: https://www.youtube.com/watch?v=2THVvshvq0Q This is what we missed on the airfield yesterday: long lenses. As you can see, the best dynamic effect comes from cutting between wide shots and tele shots. And I don't rue to have bought the Ronin, because it offers some things you can't do with a Glidecam. But doubtlessly, two or three years from now, the construction of these devices will have improved to such a degree, that you need to change then. Maybe they become lighter, maybe 5 axis stabilization will be standard, who knows now?

It has been tested very often. In raw, you can use ISO 200 or 400 to see something in the display when using ETTR. The ISO setting has no permanent effect then. In ProRes, as said above, every other ISO than 800 will decrease dynamic range.

Well, you mount the camera, balance it (quite easy, no tools needed), then you establish a bluetooth connection with your phone to tune it with the "DJI Assistant App"(doesn't work all the time, we had to restart the app / the Ronin a few times, no idea why). With that app, you initially just need one button called "Auto Tune Stability". Once hit, the Ronin makes loud servo noises like an ED209 or a T800 , the calibration stand shivers, and the camera turns upright. After a few seconds, it's useable. But there are other parameters to tune. By default, the camera follows your movements, but way too late and too slowly. This is what these guys are for: Probably you don't want any roll behavior with a gimbal, but you'd want the camera to follow your deliberate pans and tilts, you just want these movements smoother. So you have to play with those settings, until you are satisfied ("Deadband" = the higher the value, the more the Ronin ignores your movement). Took round about an hour, until we were satisfied. This morning, we went to the glider airfield to test the thing in the field. The sun was so bright, we couldn't see shit on the tiny A7s' display. So we mounted the Shogun on top. We still couldn't see enough to really frame the images. What is more, the additional weight of the Shogun made holding the Ronin for more than, say, two minutes - impossible. And we both go to the gym on a regular basis and do quite a lot of shoulder and neck workout. We also changed the lenses. One was considerably heavier than the first, but balancing it (of course you have to carry the calbration stand with you) was a matter of two minutes. It took another ten minutes for the phone to find the Ronin (this is kind of a PITA) to hit Auto Stability, but afterwards everything worked fine. For the time being, having not much experience with the device, I dare to draw some first conclusions: 1. To believe that such a gimbal replaces tripod and shoulder mount would be naive fallacy. 2. It replaces a steadicam. 3. Even though this is a lightweight model and we had a small camera, you really don't carry this around all day. 4. I am musing about video glasses. Any suggestions? Comments? Has anyone tried this?

Only yesterday we received the Ronin M. Works from the start. Only concern is that it feels heavier (with camera) than the demo videos suggest. Maybe one day we change ...

The title says it all. Can't comment on the app so far, must read the 341 page manual. I'm looking forward ...

Yes, of course. But if you spread the 256 shades too far to the shadows and highlights, few are left for the midtones, where the action is, usually. Log profiles work with 10-bit (1024 shades) or raw. Even then, grading for 8-bit (i.e. Vimeo) means crushing blacks and sacrificing highlight detail for a better image, and in most cases the results will look like autumn leaves in the end ... And: does 'grading' really imply excessive tone-remapping? HDR? Squeezing 15 stops into an 8-bit image at any costs? Many succeeded in color correcting threadbare consumer video before. Of footage that was much worse than that of the A7s. Very few, on the other hand, can present beautiful material from SLog-recordings. See what Mr. Sudhakaran, seller of the Wolfcrow guide, has to show. Does this convince you?

Yes, the Wolfcrow E-Book is $80, and though it may gather all information the author knew about at the time he first published it, he has very headstrong opinions, which imo don't help to get the best out of your camera. Me and my buddy are only testing the A7s since a month now or so. We have a lot of Canon glass, so we bought the Metabones adapter. Can't recommend it, because it turned out to be unreliable (aperture). We have a Sony Zeiss 16-35 and a Sigma 24-105 (imo the better one, 24mm is also wide enough), both E-mount of course. For faster primes, we are planning to buy some Rokinons, perhaps 50 + 85. As I wrote before, I'm no friend of superflat PPs with an 8-bit codec. Try autumn leaves, and you will get better skin tones than with any SLog+LUT-treatment. But that's my strong opinion.

Since the master format (see imdb) for MI5 is 2k, what would you expect? 2k on a 4k projector is still 2k. In an IMAX theatre, where the seats are closer to the screen, you'd see the limits of resolution. A cinema floor plan is a compromise always, and the only suitable format for an IMAX theatre is IMAX - in 4:3 aspect ratio. 4k and above is best appreciated on an UHD TV, and with a documentary. You can position yourself as close to the display as you like (at some point, a curved display is an advantage). The images of such a film should move slowly, if at all, because otherwise motion blur (like in an action movie) blurs away the detail.

I didn't watch Coreys tut, but I know one thing: You never cut precisely to the beat if you want the greatest emotional impact. You cut to the rhythm, and you do it intuitively. Don't believe me? Watch any good music video and tell me how many edits are precisely on the beat!

You are right. But Kubricks approach was deconstructivistic. He used the fisheye and ultra wide lenses in 2001 to make us perfectly aware that he distorted our view at things intentionally. He stretched the preconditions for total immersion radically away from any naturalistic style and forced the audience to believe everything nonetheless, be it weightlessness, absence of orientation in space, relativity of time (and therefore to some extend logic) and finally with an "answer" (the ending) to that we (as humans) were just the question. With the fight scenes in the Revenant trailer, I am unfortunately reminded of reality soaps. Perhaps this was Inarritus intention, we'll see. Everything is possible as long as it is appropriate to carry the emotions of the scene. I think Cameron will prove HFR to work in Avatar II. If your TV has one of these smooth motion modes, see Avatar on BD on it. You'll be astonished. It looks better.

Yes, but your eyes do not sharpen outlines. For me, the term 'organic' means a soft image, no matter if I can see 'film grain' (analog films' smallest 'pixels') or not (better resolution). I saw this trailer too. And I didn't like the look at all. Because extreme wide angle distortion looks like video (although it has actually nothing to do with analog or digital). It looks like GoPro, it looks cheap, unpleasant. I agree with you. But I suggest we nail down what is really meant when people distinguish video and film by the overall look. Video stands for an image that transports no emotions, for what reason ever. Be it, like in most cases, that the maker didn't invest any care, not for the actors, their clothes, their make up, the productions design, the lighting, the framing, what have you. Be it that the maker chose the wrong frame rate, like in The Hobbit. Be it that, like in many amateurish shorts, he tries the vintage look you describe. You recognize a transvestite by his over-the-top appearance, and very few will mistake him for a woman. On the Arri 65, Arri CEO Franz Kraus made some interesting statements on this years' NAB. He said DoPs chose the Arri 65 for the aesthetic of the larger sensor which suited epic themes. Epic films demand a broader scope and really big images, therefore 4k (greater resolution allows a softer, more organic look, this has little to do with 'detail' and nothing at all with 'sharpness'). I find this an intelligent remark. Because in the reverse conclusion this means that there are narrative films where it doesn't fit. More or less the same reason why filmmakers decide for widescreen or scope.

It's called the Render Cache, go to page 140 of the manual: But usually, you won't need to render in the Edit page, since pretty all funky things happen in the Color page (those few old-fashioned effects can't be a big deal).

Very rarely you will have all channels clip evenly. And Turbogards highlights are particularly badly balanced, as his waveform confirms.

You lose a lot of highlight detail. With raw, use ETTR with 100% zebra. Allow the sun (the perfect circle) and artificial light sources (zebra should appear within the sharp form of i.e, the light bulb) to clip, but never reflections. There is a lot of lost detail on the walls. Of course this could also be a matter of insufficient highlight recovery in Resolve. Use the luma waveform to see how your values are mapped. You can of course deliberately lose detail. If, for example, you have faces in the sun at noon, it can be better to let the sky clip. The sky is always several stops brighter than anything else, and imo there is no point in trying to get an HDR image then. In extreme cases, you can even cause noise in sun shadows (even on faces) if you don't sacrifice the sky. We all know how a sky looks anyway. A professional DoP would bounce the faces but sometimes let the sky clip nevertheless. In Hollywood, as everyone knows, there is always blue sky, and not many things can be more boring. I think so. You can do everything in Resolve. The tools there are much better. Add saturation (and everything) with more nodes. Turn on soft clip. And if a Vimeo upload file is all you want, you can as well deliver in ProRes422 ...

How about NO Picture Profile? Every more or less flat profile robs values from the skin tones and results in worse images, because in the 8-bit codec, little can be done in post to add vibrance to those. This Michael Jackson video shows this, imo. The skin tones are poor. Even if you had the best luts to nail down the right *colors*. Rungunshoot (Brandon Li) just uses autumn leaves, and then apparently you have some options for CC (though not for heavy grading, but this, again, is hampered by the 8-bit). I don't say I'm right, I'm just starting to test with the A7s. I'd like you to show me that I'm wrong.