squig

400%!

Not oversaturating like a muppet helps too.

Try lighting the shot without the computer screen, or balance the computer screen light with the room practicals; otherwise it's going to be a nightmare to grade. Much easier to fix these things in pre.

I like the full frame look, but having shot films full frame since 2009 I get how S35 is more practical. It's a good lens; it sux at f/1.8, but it gets better from f/2 and is very good at f/2.8. It's as sharp as my Leicas. Bokeh isn't bad either. Focus throw is very short, but I don't do much focus pulling at 20mm.

The crop's about the same as the BMMCC with a speed booster. The crop alone gives the MkIV a 2 stop low light improvement over the MkIII. Where I'd normally be shooting at f/5.6 on the MkIII, I'd be shooting at f/2.8 on the MkIV to get the same DOF. It could compliment the 5D MkIII raw well, depends on Andrew's log profile and noise. I've already got a Tokina 11-16mm and a Sigma 20mm, so the wide end is covered.

Filmconvert 5D MkIII Resolve ML raw BMD Film profile with a few tweaks. Resolve grade.

Looks ok to me.

Size matters but it's more about not needing 4k for narrative work, and the extra post expense that incurs. Both the MkIV and MkIII with ML would make a great 2 camera narrative film set up. Now I'm running with the MkIII and BMMCC as a b-cam. The D750 is great, but something smaller with an EVF suits my needs better, something I can pocket and have on hand for production stills and BTS video. I've got my eye on an X-T2 23mm f/2 combo.

I don't care.

ML have been doing Digic6 development for a while, it may happen sooner than you think. When I was recording 2 seconds of raw video people said continuous raw would never happen. A week or so later I was recording continuous raw. It isn't cheap, but there's not a lot of 1080p full frame raw options. The street price should be around $4500; the Terra and the Ursa mini 4.6k are $7000 AUD and they don't do 2k full sensor raw. I'd have to invest another 6 grand in hard drives and tapes to shoot my film in 4k. I'm praying to Kubrick, God of DSLRs and debauchery. Right now there's only 2 cameras I really want to shoot with (besides my MK3): a 5D Mark IV with 1080p raw, and an Alexa Mini. Screw 4k. A D750 with raw would be even better than the 5D, but the guy who was trying to hack it hit a wall. I prefer mirrorless for stills, why I still have the A7s, possibly soon to be replaced with the new Fujifilm.

I'm messing around with different film stock LUTs (I've got a smallHD 501) and custom camera profiles, but the Canon standard profile works well enough for monitoring. I create whatever look I want in post, but I've been thinking about making custom LUTs from graded camera tests.

No, but it's 1s and 0s; anything is possible, somebody just has to figure it out.

Lulz! I knew it. The BMMCC takes up fuck all room in my bag, even so its worth putting it there just so I never look like a complete fuckwit.

Having slept with a Mk3 for 5 years I can already see from the DP Review test the potential of the Mk IV with ML raw, it's not a "game changer" but I'm pushing the Mk3 to its absolute limit all the time, so an incremental improvement: take my money. 4K raw doubtful, but compressed 60p 1080 is a maybe. I shoot everything with a 2.39:1 frame which is just over 60MB/s. 5D Mk3 2.39:1 1080p raw is about 62MB/s. At that aspect ratio 4k 3:1 compressed raw is doable on a CF card, but I'd still shoot 1080p raw because it has enough resolution for broadcast/theatrical and there's no crop.

Ewwww. Hell yeah. You 3:1 compress it.

You know how it is, it's all about the click bait these days. But even so, it was DP Review that alerted me to "ISO invariance", which makes a big difference to shooting technique.

I haven't used mine much, but if Zak got an OLPF that's probably a good indication that it needs it.

We're not. It's interesting to look at to get a bit of an idea of signal to noise ratio improvements. Sensor dynamic range tests are what they are: sensor dynamic range tests. There's lots of variables like testing methodologies, subjectivity, in camera processing, codec, log gammas, and post processing techniques that all together give the true available dynamic range. Most people believe the A7s and BMPCC have more dynamic range than the 5D Mark 3 raw mostly because of published figures they've seen. Back in the real world the 5D has the most finished graded range of the 3, and I only know that because I've extensively tested all 3.

I know what it's like to get all emotional about a new camera release; in my case it's expensive. I don't know if it's the encroaching grey hair, I got a life, I learned to love the bomb, or I just don't give a fuck anymore, but nowadays I'm a lot more focused on creativity than pixels. Pixel peeping is still a very important activity, especially for a freak like me with the audacity to shoot Hollywood money shots with no budget. So as you may have heard Canon have ramped up the sausage.. erm marketing machine and spat out another 5D. What Canon has delivered is pretty much what was to be expected: another incremental upgrade, nothing pixelorgasmicly™ groundbreaking, but there's a few things to consider before totally writing the 5D Mark IV off: 1) It's better in low light than the Mk3 2) There's more resolution in the DP review tests than the Mk3 which MAY produce a 1080 image something akin to the A7s, not quite ML raw resolution, but on par with the BMMCC. (not sure if it has an OLPF? Aliasing?) 3) the Mark IV looks to have a 1-1.5 stop dynamic range boost (waiting to see a trusted test like DXOmark) 4) 1080p h.264 HDR mode produces a nice dynamic range boost and could help a lot to produce a decent 8bit H.264 image (as far as 8bit H.264 can be considered decent). 5) Has anybody crying over the 4k crop ever heard of the Tokina 11-16mm? or the Sigma 18-35mm? Would any of that make it worth $5000 (AUD) strictly for filmmaking? No fucking way; I can get a C100 MKII here for that money, but moving along.... 6) Magic Lantern 1080p raw is a possibility (albeit not a probability). If Magic Lantern eventuates, with the aforementioned improvements it would definitely be worth the $5000 to me. The two main issues I have with the 5D Mark 3 raw are the 11.7 stops of dynamic range, and noisy digital gain over 1600 ISO. Despite those issues there is no raw camera on the market that can match it in that price range. So I'm going to wait and see if the Magic Lantern developers can do their magic; if not: too bad so sad.

Even though they are ISO invariant, you wouldn't do that with the Nikons either.

The Alexa Mini can do internal raw. 30k for a camera that can't do what a hacked 5D MK3 can do, what a joke Canon is.

Noise looks to be on par with the 1DxII up to 6400 ISO. I'll wait for the DXOmark figures, but I suspect a 1-1.5 stop DR improvement over the MK3, which would make it a great upgrade should Magic Lantern eventuate. It looks like it will be much better in low light than the Mk3. 5D Mk4 ISO 800 pushed 3 stops

A7s 100 ISO pushed 6 stops A7s 6400 ISO

Erm 5D Mk3 100 ISO pushed 6 stops 5D Mk4 100 ISO pushed 6 stops At 6400 ISO they look about the same except the MK4 is sharper than the MK3 and the D750. While it's not ISO invariant like the Nikons, it is a huge improvement over the Mk3, and that noise isn't hard to clean up in ACR. It's almost as resolute as the D810. D750 6400 ISO 5D Mk4 6400 ISO If you need to push any DSLR video 6 stops good luck with that.

I'll write something up and send it to the forum admins when I get a chance.