kye

Oh, the other thing I recommend is to test things yourself. I've proven things wrong in 5 minutes that I believed for years and never heard anyone challenge or question. Most of the things that "everyone knows" online is pure BS, and the ratio of information to disinformation is so small that if someone is disagreeing with the majority of people, then the majority is probably wrong and maybe the minority right. Oh, and if someone tells you something is simple, they just don't understand it enough.

YouTube 4K is barely better than decent 1080p.. here's a thread talking about this very topic: In terms of why your 200D 1080p is soft, that's a Canon DSLR, and nothing to do with 1080p. My advice is this: Watch a bunch of videos ON VIMEO to see what cameras are really capable of in 1080p - you'll be amazed. If you decide you want/need 4K then so be it, but do yourself a favour and try and actually look at images instead of brand names. And yes, "4K" is a brand name - just how the manufacturers of TVs marketed it to people to get them to replace their perfectly good 1080p TVs. Most movie theatres have 2K projectors, so lots of marketing was needed to get people to buy a TV that has 4 times as many pixels as a movie theatre. Forget about Canon, or be willing to pay the Canon Hype Tax. The internet is full of people who think that Canon is the king and everything else is second class. These people are fools who don't know how to tell if a camera is any good or not so they just check what brand it is and then go hang with the people they know will make them feel better. Canon has great colour science - so do most other brands. Canon has great AF - so do many other brands. Canon cripples their products because the fanboys and girls will buy whatever they're selling anyway. Go to the ARRI website, the RED website, and the BM website, download their sample clips and have a look at how plain they are. Try and colour grade them and see what you get. This should show you that the glorious images that you are seeing online from the cameras that you're lusting after, the Canon ones especially, are due to the skill of the operator in post, rather than the manufacturer who designed the camera. Good luck. My journey started with me wondering why my 700D 1080p files looked so bad and thinking I needed Canon colour science and 4K to get good images. I've now deprogrammed myself and use neither Canon equipment nor 4K, but I've spent a lot of money on glass. Good luck.

I've heard that VFX is a different thing entirely, and that you want RAW and at a high resolution as possible. The RAW is because you want clean green-screens without having to pull a key and be having to battle with the compression that will blur edges etc. The resolution is so that the tracking is as accurate as possible so that when you composite 3D VFX into the clip the VFX parts are as 'locked' to the movement of the captured footage as possible - VFX tracking has to be sub-pixel accurate so that the objects appear like they are in the same space as the footage. Screening in 2K is probably an advantage as well as it would mean that there is a limit to how clearly the VFX will be seen, so in that sense 2K probably covers up a bunch of sins.. like SD (and then HD) hid details in the hair/makeup department work that higher resolutions exposed.

@Neumann Films I posted it here because this thread is about the poor quality of the YouTube codec, rather than any commentary about the R5 or the economics or artistic value of vlogging. We all have opinions on a range of matters, but I just thought it was an interesting example, considering that Mattis audience is tech / image / camera centric, and yet even on such a channel a workflow upgrade / resolution downgrade wasn't really noticed. Even for Matti, who even the nay-sayers suggest would be heavily preoccupied with his social media engagement and comments section opinion on various camera products. For reference, I watch his content on a 32" UHD display, in a suite calibrated to SMPTE standards (display brightness is 100cd/m2, ambient light is ~10% that luma with a colour temp of 6500K, and viewing distance within the sweet spot of viewing angle, etc). Maybe some are watching on a phone, but not everyone - one of the first things those who are caught up in the hype of things like Canon CS / 4K / cinematic LUTs / 120p b-roll etc would acquire is a big TV and/or 4K computer monitor. I know that because when I got into video that's one of the first things I did, as that was the prevailing logic online, and I definitely wasn't the only one. Threads like this one are part of my journey of un-learning all that stuff the internet is full of and is often flat-out wrong. For anyone delivering via a different platform other than YouTube it's a different story. Of course, not that different if you do a little reading about how many features have been screened in 1080p in theatres and the film-makers didn't get a single comment about resolution, but that's a different thread entirely 🙂

Matti Haapoja got the Canon R5. Matti Haapoja shot in 4K and 8K. Matti Haapojas computer absolutely choked. Matti Haapoja went back to 1080. Apparently, no-one noticed.

Agreed. Like everything in film-making, it's great if you can make it look nice, but if it gets in the way of the story or content, then it's wrong. I remember a wedding photographer talking about taking group photos once, and how it is critical to get everyone on focus in a group shot, which if it's a large group of people can be tricky as the people at the edges are a different distance away. They also mentioned that the most important people at a wedding are the bride/groom, but the second most important people are the oldies as its a very common situation where "the last nice photo of grandma" was taken at a wedding, so making sure to get them in focus should be a huge priority. It's easy to forget, and make the film equivalent of a cake made entirely of icing.

The more I pull colour science apart, the more I realise that companies like BM and Nikon have colour science either just as good as Canon or within a tiny fraction, and also that ARRI colour science isn't perfect and there are things about it I don't care for. I know that this will get me ejected from the 'colour science bro club' but I don't care about being popular and fitting in, I care about colour science and good images.

There's a solution to that - it's called Manual Focus. ....and if you upgrade to Manual Focus, then basically Panasonic cameras are best-in-class! Seriously though, probably the biggest challenge with the AF crowd is that they think that PDAF / DPAF are near perfect and that DFD is unworkable. The real situation is that neither are fit-for-purpose yet. AF will be great, once it reliably focuses on where the director wants it to be focusing, and does so with the right transition speed and doesn't pulse when it's there.

Having something that was continuous would be great, like the Sony one, and the JVC and FS5 ones. It's not that difficult a thing to implement technically, so it's cool that some manufacturers did it. The thing to watch is that when your digital zoom is more than a 1:1 crop you're actually upscaling in-camera, and the quality falls off a cliff at that point. I tried 2x, which is a 1.35x downsampled image, but the 4x is 1.5x upsample so it's taking about 1.3K and upscaling it to a 2K image, which does not look good! Really, the proof is in the resulting image, so just give it a go and see how your camera performs. Absolutely, you're zooming into the noise and the softness of your glass. Given that most lenses soften up when stopped down a bit from their widest aperture, you may find that a general rule-of-thumb might be to stop down a bit when using the zoom function. When you're keeping the digital zoom in the range where it's still downsampling then you do get the NR benefits of rescaling. We saw that the benefits of a 5.2K image downscale to 4K was a decent improvement on things like noise, and that's only a 1.35x downscale (on the GH5) so it doesn't take a huge amount to smooth out the noise, assuming that the noise is pixel-sized, rather than the Canon noise that spans multiple-pixels for some unknown reason.

I've recently "discovered" some of the benefits of using the in-camera digital zoom. These won't apply to everyone, but it's worth considering and I don't hear a anyone talking about it. Let me illustrate by taking my GH5 as an example, but the principle applies much wider, especially as cameras go towards 8K and beyond. The GH5 has a sensor that's 5184 pixels wide. When you're shooting 4K, the camera downsamples the 5.2K to 4K, giving a higher quality image due to the benefits of oversampling. The benefits of oversampling are well known, and many cameras have this. The GH5 also has an ETC mode, which essentially does a 1:1 crop into the middle of the sensor. This is a common feature across manufacturers. If you use the ETC function in 4K, you get an additional 1.3x crop, and if you use the function in 1080p then you get a 2.7x crop. Both of these modes are shooting a 1:1, so you get a tighter FOV, but the image is no longer oversampled, so the quality goes down. Enter the digital zoom. In 1080p mode, the GH5 allows a 2x and 4x digital zoom. The 2x digital zoom is less than the 2.7x 1:1 zoom from the ETC mode, so (assuming that the cameras image pipeline is designed well) the resulting 1080p image should be an image that is taken from the middle 2.6K pixels and downsampled to 1080p. In other words, the 2x digital zoom is a way to punch-into the sensor but still keep the quality of an oversampled image. This principle will occur any time that the digital crop is less than the crop of going to a 1:1 area on the sensor for whatever resolution you're shooting in. The Sony ClearImage Zoom comes to mind here, where (I think?) you can zoom in by a lesser amount than the 4K 1:1 crop (which is something like 1.5x?). Perhaps other manufacturers have similar functionality too. Of course, the lower the resolution you're shooting then the more likely this will be available on your camera. For those still shooting 1080p then this is worth looking at. For those shooting 4K, this will increasingly be useful as sensors creep up towards 8K and beyond. For me though, I've now abandoned the idea of having to buy an 85mm prime, because I can simply do a 2x digital zoom with my 42.5mm lens and get the same FOV and (basically) the same image quality, and if I happen to have my 7.5mm lens on and want to quickly grab a 15mm FOV shot then I know I won't be sacrificing quality to get one. Had I known that this feature delivered such good results I may have actually bought different lenses, so it's not a trivial thing, and it's worth giving a quick go if your camera supports it.

Makes sense. Getting a 2.8 that is sharp from f2.8 up would probably have been more expensive. Of course, if you're cropping into the image circle then you're really only interested in the performance of the lens in the middle of its image circle. Unfortunately that's data that places like lens rentals publish, and not for that many lenses, as it's obviously very laborious to gather. I ended up with the Voigtlander 42.5mm f0.95 prime. Not sharp wide open, but almost as sharp as the competition when stopped down to their widest apertures, and can be used wide-open in a pinch. Subjectively a bit of sharpening goes a long way to matching things, especially if you're degrading the image in post as well with halation or other softening effects.

lol, the only place I didn't look was in the thread title! 😂😂😂 There's a pretty well established principle that lenses sharpen up when you stop them down the first couple of stops. As mentioned before this has exceptions, but those exceptions seem to be just that, exceptions. DXOMark has a lot of interesting lenses and you can look at the graph of sharpness (perceptual megapixels) vs aperture to see how much it sharpens up when stopped down. Typically they look like this: Where wide open they aren't as good as when closed down a couple of stops. So in that sense, the target aperture really matters, as the best way to get a sharp image at f2 might be to buy an f1.2 lens. The exception is normally much slower lenses that have the same kind of optical performance as the above, only don't open the aperture as much.

Absolutely. I have f0.95 primes, so you're preaching to the choir here! Actually, they didn't. Or at least, when I just re-read this thread five times looking for it, I couldn't find it! True. The issue I took, and thus my slightly sarcastic reply, was that this is viewed far too simplistically by people. Take two hypothetical lenses, one is f1.4 and the other is f2.8. Let's say that the f1.4 one isn't so sharp wide open, but the f2.8 lens is. The traditional, one-dimensional, thinking is that if you want sharp images then the 2.8 lens is the one to go for because "the f1.4 is soft wide-open but the f2.8 lens isn't", end of story, and mostly, end of how deep the persons knowledge is about the subject. The problem with this thinking is that the f1.4 lens might be sharper at f2.8 than the f2.8 lens is, but the sharp-wide-open one-dimensional thinkers don't go that far.

Why, f4 lenses are sharp wide open, that was your criteria. People often make the mistake of judging lenses wide open, but the problem is that not all lenses have the same largest aperture. Sometimes the best way to get a lens that's sharp at a particular aperture is to get a lens with a wider aperture than your desired one.

This seems to be the download page from Cinematography.net: https://cinematography.net/raw-and-exr-camera-rushes.html I haven't downloaded anything from there yet, but will check it out. I'm also unsure of the licensing, so if you explore these then make sure you're clear on that before proceeding.

I have consistently failed to make my own footage as good as I was hoping for, so have decided to try my hand at grading the best footage available, the sample shots from the manufacturers themselves. I've done this in the past but got sidetracked, but am now back on it. I figure that until I can make ARRI or RED footage look good, there's no point criticising my own footage - maybe the footage is fine and it's my grading skills that are solely to blame. My previous attempts, however short, made me realise a few things: flat ungraded footage from any camera looks dull there was a lot more noise in the RED and ARRI footage than I was expecting - ie, A LOT more Overall, the biggest surprise was how average the footage looked. Since then I've learned a bunch and hopefully will have more success. Anyone want to join me? Arri footage is here: https://www.arri.com/en/learn-help/learn-help-camera-system/camera-sample-footage RED footage is here: https://www.red.com/sample-r3d-files Komodo footage from Seth Dunlap is here: https://www.dropbox.com/sh/cfxdizrgt3aa4ml/AADPxF6XZIs2_uB7Xt_piHUya?dl=0 BM UMP 12K footage is here (underneath the "Generation 5 colour science" heading): https://www.blackmagicdesign.com/products/blackmagicursaminipro BM Pocket camera footage is here: https://www.blackmagicdesign.com/products/blackmagicpocketcinemacamera/gallery Sony Venice frames here: http://www.xdcam-user.com/tag/footage/ but I couldn't find any official footage for download I've downloaded footage from https://raw.film before too, so that might be worth checking out. As a reality check, I'm also pulling in the ML footage I've shot back when I was playing with it on my Canon 700D, and also the test shots I've been shooting with the BM Micro Cinema Camera. I suspect that the footage I shoot is much better than I think, and that I'm much worse at colour grading than I think. If anyone is interested in joining me that would be fun, and if there are other places to download well shot sample footage that would also be very interesting.

Maybe it's just marketing. The FF enthusiasts / myopians will look at this and have their confirmation bias renewed and will be left with the impression that Panasonic understands them, sees FF as being the superior sensor size ordained directly by God, and that Panasonic is the brand to get. The MFT enthusiasts will look at this, be encouraged by the frame-rates and remember the track record of the GH line, and hold their opinions for when we see the IQ. Then the GH6 can be released and the MFT enthusiasts will see that the IQ is excellent. The FF myopians won't even look at "toy camera" reviews and will be none the wiser. ....and everyone gets what they want and Panasonic lived happily ever after. *roll credits*

Is this just fantasy? Lockdown in pandemic No escape from reality...

The fundamental challenge you have is the one that we all have - there's no perfect camera. The second, very unfortunate, challenge you have is that the 'simple' task of having a small, light-weight camera that can keep up with your kid and how you live your life. What this translates to, and why I put 'simple' in quotes, is that you want a camera that can: be small and pocketable, likely implying a smaller sensor have autofocus that can keep up with a small child (otherwise known as 'world-class' - watch some AF tests if you're unsure of this) can do that AF in the lighting that you find yourself in, which after the sun goes down, will be extreme low light has the kind of IQ that you deem acceptable Basically, the above is a very difficult set of requirements, unlikely to be met by anything, so you'll have to choose which things you are willing to compromise on. Parents getting a camera for pics of their kids running around think it seems like a reasonable request, but in reality it's kind of like saying you want continuous eye-AF at f1.4 in pitch-black. Cameras are getting way better than they used to be, but we're not there yet.

See if you can borrow one (or buy it with a return option) and do a bit of a stress test. I say this because I have a friend who is into stills and has an old DSLR that he hates lugging around, but he's repeatedly tried going on trips with only the latest iPhone and has kept going back to the DSLR. His reason.... none of the iPhone photos are good enough for him to print and put on the wall. The moral of the story is that the best of the convenient options may still not be good enough, and you might be better off trading some of the convenience in order to get something that will work for you. In the end we only remember the footage, not the camera gear, so as long as the equipment is as minimal as it can be for the required results, and as long as it's minimal enough that you actually take it and use it, then that's the way to go. You haven't talked about what focal lengths you're interested in using, and how you actually shoot. A MILC and pancake prime isn't a bad option and I've done that as a (barely) pocketable option in the past to great results.

Very impressive. Almost didn't watch it as the thumbnail doesn't do it any kind of justice whatsoever!

As a stop-gap measure, could you use a filter on front of the lens. Companies like Tiffen offer various models. Obviously something on the filter stack would be easier, but in a pinch...

I have played around with various ways to structure a grade, with most of my real projects (not tests like I post here) having 3+ cameras. I tried using Groups, I tried using the Timeline feature, but I found them all to be kind of cumbersome in various ways. Here are two approaches for you consideration. Approach One - Shared Nodes Setup a node graph on the first clip that looks like this: A few normal nodes for this clip Three Shared Nodes for this camera Five Shared Nodes for the whole project The idea is that the normal nodes at the start are to adjust things that are likely changing from clip to clip, like WB tweaks, exposure tweaks, and maybe NR or sharpening (some of my lenses are softer at wider apertures so benefit from a little sharpening to even out the odd fully-wide-open shot). The camera specific Shared Nodes are for every clip from that camera and are typically structured to have the second node do the Rec709 conversion (be it via curves or LGG wheels or CST or contrast/sat/pivot or whatever) and the first one be pre-adjustments and the third be post-adjustments. The global Shared Nodes are to apply a 'look' to the project. Depending on how you like to work you could setup more than five to begin with. It's easy to add more at the start and not use them than have to add them later. You can also choose to add a "scene" set of nodes between the camera ones and the global ones, if you want to have different looks for different scenes. Approach Two - Simplicity Itself Review the footage and find your hero shot (or shots, if you've got multiple scenes), which will act as a reference. It could be a good idea to have this as a wide shot because it will include lots of different luma and chroma values. Grade the hero shot. Take a still as a reference, and make sure you're showing the gallery on the screen so it's visible. Grade each shot, manually, to match the hero shot. It should match closely enough to not look 'different'. Apply a look over the whole timeline, either using a Shared Node, use the Timeline window, Put on an Adjustment Layer, whatever you want. Go into the Lightbox view and look for any clips that don't 'fit' and go and fix them. Done. The second approach works really well if you have a control surface. On well shot footage you can get away with doing very simple adjustments to most shots, and only have to do fancy things to solve problems, like desaturating distracting elements in the background or whatever. Cool workflow, thanks for sharing. There's some things in there I hadn't heard of before, like the colour memories feature. Will have to look into that a bit more. Resolve is so deep....

Depending on how you define "cinema camera package" I would suggest that the BMMCC would likely be the leading contender. I've had significant trouble rigging mine up, considering that all the things you have for rigging and accessories are ridiculously large in comparison.

I think it's about intent. Christian starts his video saying "Most of the people I got to know in the last two years, let it be photographers videographers film-makers whatever, tend to only go out and shoot if they got something in mind to achieve, maybe it's a vlog, maybe it's a client project, maybe it's just a travel film". Trout and Coffee starts his most recent video saying "Hudson and I hadn't been on an adventure together for a year, by now it was the very end of August, I had been spending a lot of time staring at my computer in my office and could tell I needed to break off, spend some time away and do what I love most - make a simple film". So, the whole premise of that video was to make a film, ie, no shot in that was C-Roll. Sure, it's not scripted, but every time he hit record during that trip would have been in the context of "I'm making a film about an adventure with my dog" - ie, he had something in mind to achieve, like Christian said. The shot in the one you posted where he was talking with his dad about fly fishing might not have been scripted, and he might not have even shot it with a specific context in mind, but this is where we get to the tricky parts of YouTube and how it becomes all-pervasive. This is the hidden side that people typically only talk about when they're having some kind of meltdown. This is something that lots of people have spoken about - "and you start losing focus on like, what am I doing - am I doing this because I love her or am I doing this for the video? That's toxic for any relationship" (from the below video about 2:30 IIRC.). Most people put on a facade when they go out in public, or when a camera is pointed at them (which is basically the same except the public are invited into your life via the camera). When you're filming your life it's hard not to start living your life like people are always watching. There are great articles on surveillance and how privacy is kind of related to being able to relax in some way. In a sense, C-Roll might be different to A or B-roll due to the expectation of what will be done with the footage, because that's what matters in the relaxation and the creativity. I make home videos, but no vlogs. I make a video of when my family has a holiday, or goes out somewhere, maybe for a party or something. If I'm doing the dad-with-camcorder thing at some occasion where we all got ready, got dressed, and went out to do something special, then that's not C-Roll. What is C-Roll is the shot I got of my daughter (who loves drama and acting) that starts with me testing mixed lighting in the kitchen/dining room and the kids had been home from school for about 20 minutes, and my son comes through to raid the fridge and my daughter is telling my wife about some thing that happened at school and I kind of cheekily kept rolling and she saw the camera and then started making faces and then came over to the camera, took a breath, held for a beat, then dropped into character and then proceeded to give a highly bumbling and misguided tour of our house from the perspective of the character she was in that didn't understand 21st century western living. That shot is the one that was shot for no purpose, wasn't aligned with any project, was definitely not shot for publication, and I will probably drag out for her 21st birthday 🙂