kye

CA is awful and just reminds me of the horrific edges of people against a digitally clipped sky from a low-DR low-bitrate digital camera. I don't have issues with corner softness or vignetting, but wow does CA make an image scream cheap amateur video to me!

If someone wanted something smaller than an Alexa, cheap enough to crash and better image quality than a gopro then they could have just bought a BMPCC or BMMCC. Sounds like they set out to build a camera for a specific task without first phoning up the camera store to see if one existed already.

I have a few m42 lenses and an m42 0.7x speedbooster and have done comparisons between the speedbooster and the lenses with a non-SB adapter. It's definitely worth doing with your own setup, but I found that the speedbooster didn't really add any perceptible difference when compared to the distortions of the lenses I tried. One thing to pay attention to is that the speedboosted image may have softer corners but that could be that you're just seeing the lens further away from the centre.

Filters and processing in post (as others have previously stated) can get the job done pretty well. Here's a quick attempt to match a Samyang to a Lomo in Resolve - IIRC it was only a few minutes work. Filters would be even better as they emulate flaring from off-camera light sources better, which can't really be done easily in post. Before: After:

hopefully you'll find an adapter in-country. ebay and shipping times from china for accessories and adapters has significantly improved my strategic and long-term planning skills. I've bought new equipment ahead of a big trip and had to sit and very slowly and methodically go through and identify all the adapters and accessories I'd need to buy so that they all arrive in time. I've previously drawn pictures of the setup in order to identify things.... camera body... ordered... <draw a picture of a camera body> .... does setup work yet? No, I need a lens. camera lens... ordered... <draw camera lens next to camera body>... does lens connect to camera body? No, I need adapter. lens adapter... ordered... <draw a picture of a lens adapter joining lens to camera> .... does setup work yet? No, I need a battery. ..... then at the end... I pick up the camera, put the wrist strap around my wrist, turn on camera, it powers on. OK. Light goes from scene through adapter into lens, then through adapter, then onto sensor. I hit record, camera writes, files to memory card.. <ok.. google if my memory card will work with camera. good. what about for the specific mode I'm recording in? excellent. What about a backup plan if card doesn't arrive in time...> etc etc etc I replaced my XC10 with GH5 with only weeks to go until a major trip. Requiring all new lenses, batteries, media, adapters, etc... I had to get a lens shipped to a pickup point near the first hotel we were staying in! Ah, those were the days.

Do RED provide good profiles for ACES and RCM or conversion LUTs? If so, maybe they can be matched in post easily? Things like easy and accurate colour matching is one of the things you pay for when buying an established brand name, and is worth money when shooting with multiple models from one manufacturer, so I'd anticipate that they'd be onto it.

Some months ago I switched from focusing on new cameras to understanding lenses and colour science, and you'd be surprised how many times I was looking for compatibility between some lens format and some other camera and EOSHD links kept coming up in google. If you put in "bmpcc c-mount" into google the EOSHD thread talking about compatibility and vignetting is not only the top hit, but also by far the most useful. It seems that EOSHD is the centre of the internet for a number of very specific things in video-making, especially topics that focus on the benefits and modern use of vintage or older technology rather than just the relentless regurgitation of specs of new cameras. When I post here, I am often thinking of these threads as a panel discussion where there is an audience of people who don't post or aren't registered who are looking for information on the topic.

@Amro Othman yes, the differences in those two images are significant, for lens tint at least. One thing I would caution on is that colour shifts can often occur in camera (or in colour profiles / conversions when processing in post). I have tried replicating colour science across cameras on many occasions and it is very common to process different luminance values in different ways. For example, the old orange/teal grade pushes mids and highlights warmer and shadows cooler, but also many film-look processes desaturate extreme highlights and shadows to create clean and blacks. The list of common processing tricks goes on and on. To indicate how prevalent it is, someone did a colour accuracy comparison and Sony cameras were the most accurate and yet many other brands with less accurate colours are perceived to have nicer colours, so effectively, distortion of colour accuracy is not only common but desirable by the market. Doing a direct comparison with identical exposure would be very interesting to see.

I saw the image above and thought how nice an image it was, then clicked through to the post and saw it was a Minolta 55/1.7 and thought it was even better, then I read the cost.... these old lenses really are something! Nice images

The GH5 shoots 4K 10-bit 422 h265 and 6K 10-bit 420 h265. I also use a laptop for this purpose, despite having a GH5 and shooting the above mentioned h265 modes.... I feel your pain.

Yeah. I've noticed a pattern of cinematographers saying that they got the mini to do a few action shots or whatever and then in doing testing they discovered the mini was really great and ended up using it for the whole production. These comments made me wonder why you wouldn't just do that as kind of a default, and only go for a larger model if there was a specific reason.

Is this file format a new format? I got in touch with BM about getting better performance with playing 10-bit 5K h265 files from the GH5 and they said that AMD hardware acceleration for h265 was newly implemented (at that time) and may not have been optimised yet, so maybe there is further optimisation that might be done?

Yeah, to me the main price is the size and form-factor, not the $$$$, but it depends on what you shoot.

I was going to say that the reviews I saw of the a73 said it had as good low light performance as the a7s2, but I guess if you're going external then they're not equivalent.

There's definitely a use-case for having a camera to capture a signal as evidence, but I'd suggest that one with a wireless connection that is continuously uploading would also be handy. People that do the wrong thing aren't afraid to confiscate the evidence of it. What's that saying... once you've crossed the line it's hard not to go all the way.

You didn't, I was just having a little fun while also making a point. I think that the RAW vs compressed comparison doesn't align with still images at all because of the way that compressed stills handle DR. Compressed stills clip the DR of the camera, which is why many people shoot RAW. Personally I think that's blindingly stupid when it's just as easy to make a JPG from the whole DR of the RAW image, like any free image editor can do but somehow cameras costing thousands of dollars can't do. If shooting log h264/h265 cut off a couple of stops of DR then we'd all be lusting over RAW and having t-shirts printed about it as well, but thankfully, we're not.

That reminds me of (IIRC) National Geographic which doesn't permit any manipulation of the content of an image. To verify this you would submit your final images (with all the cool colour grading, cropping, etc you did) and also the RAW file and they would compare the RAW file with the edit to ensure you didn't photoshop out powerlines or whatever. It makes sense that the edited version would be the one published but the RAW capture would be retained to prove the content of the output. I definitely agree that digital video will gradually descend into fakery, in much the same way that photography has done with things like Mariah Carey being photoshopped to basically be a different person. The only difference between still images and video is processing power, and things like Resolves Face Refinement is just the beginning. It's interesting to use film as the proof, although if digital is the source then the manipulation could be done before printing to the film, so tech like @KnightsFan mentions would be required.

I have no idea what you're talking about... I was once waiting for some friends outside a movie theatre and a couple of people walked in carrying something that looked very heavy, and after walking over to the front desk and very carefully setting it down they had a quick chat with the staff before leaving again. A couple of the rather young staff then tried to lift the object, but one girl wasn't strong enough to lift her end, so got someone else to come and help carry it inside. Bored and curious, I went and asked the staff what the box was, and it was a roll of projection film being delivered for a movie that was about to be released. I remember it being about 1m tall, 1m wide, and about 30cm thick, with very sturdy looking handles on each side.

Many years ago there was this thing called FILM. It came in rolls that only lasted for MINUTES. You paid to buy it and paid again to develop it. It wasn't even REUSABLE. Madness!! But seriously, if you're having a hard time with RAW file sizes then RAW isn't for you. In fact, RAW isn't for most of the people on this forum. A blind test comparing RAW to Prores would pretty quickly show that compressed formats are good enough for almost everything.

As affordable cameras continually raise the high-waterline on bit-rates, are we getting to the point where the cost of media necessitates bringing back some of the frugality of film? I crunched some very rough numbers, and the cost of SSD is about 10% of the cost of 16mm film for 4K30 in 3:1 RAW, which will be more like 40% of the cost for 8K, but that's re-usable so it's not that direct a comparison. Of course, CFast and other media is more expensive, but you can dump footage to a cheaper storage medium every few hours on set, so that's relatively re-usable too. How are you folks shooting 4K RAW or 6K RAW and salivating over 8K RAW approaching your media management?

Cool article. One part that stood out to me was "During my initial tests I found that the Meike primes performed better optically in almost all areas." That's quite encouraging as the Veydras was optically excellent.

I'm happy to upload a couple of random C-Log shots from my XC10 if anyone is interested....

Nice work on your edit. The image holds up quite well, and the music was definitely a good choice

Thank you. That's probably the best source of high-quality colour grading reference images yet posted online!

I suspect they're the same too, but my point was that I couldn't find any evidence either way. It does make sense though, considering the various commercial aspects and the timeline. Ah, good point. They were for MFT mount, but that doesn't mean the optical design isn't compatible with a larger flange distance or with larger sensors. Other brands like Rokinon make the same lenses in MFT and DSLR mounts, essentially just adding a spacer the same way a dumb adapter is just a spacer. This is the 85mm Veydra: The coloured parts are the MFT sensor and the grey is the extra size of a S35 sensor, so that optical design appears to cover S35 relatively well. The other focal lengths aren't as high performance but they seem to cover it. This is a comparison between a Zeiss and the Veydra: Note that the Zeiss is graphed over a FF sensor (the X axis goes to 20) whereas the Veyrda is only S35 (X axis goes to 14). However, of note is that the Zeiss has degrading performance out to 10 (MFT sensor size) but then stays a lot more constant out to FF, and is of sufficient performance that these are actually used on FF sensors by real cinematographers. Extrapolating this principle, the Veydra may or may not have performance beyond the S35 sensor, so it may cover FF? The other Veydra focal lengths have significantly worse performance at the edges of the S35 sensor, so aren't as likely to cover FF, or at least with any quality: Regardless, the S35 Meiki lenses may have the same optical formulas, although the FF ones probably don't.