KnightsFan

I used standalone fusion for a few projects over the last few years. This week I used the resolve-integrated version for a very simple effect. I love how well it works with resolve, it's a truly remarkable piece of software. Many years ago, when I used blender for everything, I remember talking about how I wished you could put compositions from the node graph into the sequence editor. With resolve, that is finally possible!

What we really need is for a universal system of plug and play parts, like we have with PCs. Imagine if every camera were as modular as a PC! Naturally, building a camera would be a lot of research and some compatibility headaches. But just like with PCs, you could buy an off the shelf, pre-built one with software already installed. What we need is for the Axiom project to work, not just within its own world of open hardware and software, but for their standards to become universal enough that big companies manufacture parts that work with that standard.

That's probably right and in keeping with his philosophy, but I haven't seen anything about it coming to other cameras yet. Free SDK for decoding is very different from it being open source or freely implemented in any camera. Right now, it seems a lot like Redcode, which also has an SDK. Yeah, I agree. I remember Grant talking about how ProRes Raw didn't have enough metadata support back when they announced the Pocket 4K. Now I guess we all know what he was really thinking about! I also think that the whole camera-specific metadata (and possibly encoding as well?) is what would keep BM Raw from being properly implemented by Magic Lantern for the 5D4, even if they could make some post-transcoder.

@webrunner5 Yeah, the decoding part is all open, so it could easily end up in all different post production applications, but I didn't hear him say anything about it ending up in non-blackmagic cameras. Did I just miss that part? It seems like a fantastic format, I'd love for it to become ubiquitous among cameras.

@webrunner5 It would be amazing if that happened. I wonder if Blackmagic has any intention of allowing other camera manufacturers to use BM Raw? Decoding seems open with the SDK and all, but I'm unclear on whether encoding is open and/or free.

@webrunner5 Canon isn't going to help ML in any way, shape, or form.

@mercer Watch the video, Grant Petty talks about it. The Pro must have had the hardware all along. They've been working on BM Raw for 2.5 years, he said. Yeah, I know it wouldn't be in camera, just sort of rambling on about stuff haha. I'm sure there will be ways to convert it in post. After all, apparently you will be able to export videos from Resolve in BM Raw. Of course I'm not sure what the practical implication is, but based on the video, it seems that one of the pillars of design for BM Raw is designing it around the sensor hardware and color science. So maybe ML can make a converter, but will it be any good? I'm sure Blackmagic--a company with actual full time employees--didn't spend 2.5 years just to design some a container format and some .sidecar files. I also doubt they invented vastly improved techniques for video compression. My guess is that a lot of the time was spent tuning algorithms to their hardware and color science, something that the open source community might not have the resources to do. But then again, maybe I'm wrong about all of this!

In the video, Petty said it would come to the p4k.

@mercer it may be possible to create files that follow the bm raw specs either in camera on the 5d4 or in post. So I that sense, its possible. But it seems that blackmagic can achieve such high compression ratios (without artifacts) because they specifically designed their hardware encoder to their sensor. Each blackmagic camera that does bm raw will have a unique encoder. Such an encoder doesn't exist for the 5d4 sensor, so presumably, 12:1 compression using bm raw would introduce artifacts. And even if you got it working (with artifacts), it probably wouldn't be in camera unless they can do that encoding in software on the 5d4's processor. Which is doubtful.

Unfortunately, this doesn't seem very feasible based on what Grant Petty said, since BM Raw relies on sensor-specific hardware and measurements to do its partial demosaicing.

I was going to say that, but in my experience some people don't like manually placing a couple files in the right folder, even to save a few hundred GB of data over the internet... lol

I agree, two cards. you can recor to one while copying files of the other, less chances of total data loss, if you lose one you've got the other, you can shoot on two cameras at once if necessary.

Like with ProRes Raw, I was initially kind of skeptical, but I've read a bit about this Blackmagic Raw and it looks phenomenal. Here are my thoughts. - It's not RAW in the way we used to mean. It's compressed, processed, and partially demosaiced. While the terminology doesn't detract from its awesomeness, I feel strongly about calling things what they are. (I said the same thing about ProRes RAW, Canon RAWLite, Redcode RAW, etc.). - It's got a variety of compression ratios. This is VERY important! This means that no matter what quality level you are currently using, you can have the same exact workflow every time. This means the same workflow not just for the image, but the metadata as well. - The .sidecar files allow metadata created in one program to be used in another. Additionally, the .sidecar metadata replaces (without overwriting) the original metadata. You can make color adjustments in software A, then see those same changes in Software B--and then revert back to the original if you want. - Metadata can hold manufacturer-specified color information. This was a criticism Petty made earlier this year against ProRes Raw. Looking back, he was probably itching to blurt out how much better their raw format was! - Metadata can be stored per frame for things like focus distance. This is a must-have for VFX. - Unlike cDNG, the new format is in a single file. Small improvement, but good. - Some processing is done in-camera. I'm very curious how this works, but the promised result is better performance while editing. - And finally, the big one: The SDK is open! I imagine this will quickly find its way into open source programs like VLC, Ffmpeg, Blender, Kdenlive, and more. This won't affect the Big Guys much, but for the low budget folks that will be huge. - Imagine if VLC implemented the SDK and could natively show Raw files, with color correction information. You could send a client/friend a raw file straight from camera for viewing without transcoding or using proxies. - The file size was one reason I haven't really looked at the P4K, and have been so intent on cameras with H.265 (Fuji, Z-Cam). It looks like Blackmagic Raw might bring drastic file size improvements. 46 MB/s for 4.6k 12:1 isn't bad. I'm guesstimating that's like 12 MB/s for HD? That's low enough for me, though I'll have to see how the quality compares to H.265 at the same bitrate.

Exactly! But since the format is open, other manufacturers can get those benefits as well. Only Blackmagic gets the street cred for it, though.

I haven't used it very much, but I've been happy with Resolve's stabilizer so far. And to be honest, I've had very bad luck with the NX1's DIS.

I'm not super knowledgeable about Businesses, but my 2 cents: Unlike many other camera companies, Blackmagic is private. Shareholders require that a large public company GROW in order to make share prices rise, while a private company can stay about the same size without much consequence. So Blackmagic has a lot more leeway to pursue a philosophy, even if it makes them less money in the short term. What they stand to gain in the long term is to make their name synonymous with innovation, the way that Arri or Leica are synonymous with quality.

Exciting! I can't wait to see what the magic lantern people do next.

4:2:2 has double the chroma resolution compared to 4:2:0. How much that matters varies. The best way to see the perceived quality difference is to start with some high quality 4:4:4 footage and then render it as 4:2:2 and as 4:2:0. Try a variety of scenes, try color grading both or doing green screen work. You can come to your own conclusions--always better than listening to what other people claim!

That's what I'm saying--and naturally there's a margin of error, and some outright mistakes will be made, in addition to a switch in methodology (I'm trying to look up when that happened. Perhaps it goes along with the different format of the chart?). But it's outright impossible to compare camera A to camera C by judging BOTH an A vs. B by one blogger, and then a B vs. C by a completely different person. I agree the color so far doesn't look as good as Fuji or Blackmagic. I enjoy color grading, though. Tbh, until about a week ago, I discounted the E2 as a novelty because the graded test footage looked terrible. However, I downloaded some sooc zlog, played around in Resolve and was shocked at how easy it was to get a really good image. Fuji, Blackmagic, and ZCam are my top choices at the moment, but we'll see what other announcements get made. Fortunately, I won't be upgrading until December or January, so I've got plenty of time to see reviews and such! This is a great year to be almost-but-not-quite upgrading.

@sanveer Does anyone do actual scientific tests though? I understand that C5D is suspect, but afaik they are the only ones who do actual lab tests on many different cameras under (supposedly) the same test scenarios.

@IronFilm and @sanveer I meant where do you get your dynamic range info from if not from cinema5d, who? Naturally!

Good to know! Who do you get your info from?

Haha Maybe. What I like about them is they have done multiple cameras with the same methodology, so you can actually compare different brands. Do you know of any more reliable place that does the same? I'd love to see a variety of tests.

H.265, actually (big difference!). You should download some of the footage, it has very few artifacts. ProRes 4k 120 fps would make HUGE files. It's also got some interesting deep learning features that most of us won't use, but looks unique. You can also monitor and control the camera via an iPhone app attached via USB-C, so it may completely eliminate the need for an external monitor. I don't know about the P4K, but the E2 is supposed to use an all-analog gain structure, so dynamic range should by constant across the entire ISO range (though I assume that noise levels will increase). Which brings us to: It's actually about 13.5, which is in line with the other cameras' claims. The 16 claim comes from their WDR mode. I'm not sure how it works, but I think it's doing some fancy double frames or something (I'm sure they explained somewhere on Facebook). They say it can introduce artifacts with moving objects in WDR mode. I hope some third party like Cinema5D does an objective dynamic range test.

I agree... Somewhat. They haven't released any "wow" footage done by pros on a set, colored by pro. However, they have released a lot of sooc files that I think show off what the camera can do, you've just got to do some work on your own color grading. It doesn't look as good as say the p4k, but for a fraction of the data rate it's a fair trade in my book.