kye

The other major factor for amateurs is the hidden factor of editing. When you're watching a video on YT, especially about a camera (rather than what the camera is pointed at), you're probably looking at cherry-picked shots. Regardless of if a video is shot outside in available light, is indoor but naturally lit, or completely staged including lighting, the finished video you're watching probably only contains the best shots that the person captured. If you don't know this, then you're going to pick up a camera, point it at things, and then expect to be able to make every shot you take as good as the best-of-the-best shots that have been posted online. It would be a dirty big secret only it's not a secret at all, it's just that amateurs don't know about it. Are you familiar with shooting ratios?

You will get all sorts of answers to this question, but fundamentally, the physics of the way that cameras work means that larger sensor cameras will always be better. Smartphone cameras can (and do) look great in the right circumstances, matching larger cameras. But when you start trying to use these tiny cameras in anything other than the ideal circumstances then they either can't do something (eg, optically shallow DoF, lens choice) or they do it very poorly (eg, low light video).

Interesting comparison of the IBIS on GH6 vs S5, linked below at the timecode: and if you watch a little longer, it just murders the Sony A1. Perhaps the lesson from that is that the GH6 claims 7 stops, the S5 claims 5 stops and the A1 claims 5.5 stops, but the A1 is clearly worse than the S5, showing that the specification is measuring the wrong thing and doesn't actually indicate the relative level of performance in the real world. Instead the limits of performance is how far the sensor shift mechanism can actually shift the sensor (which doesn't seem to be provided or isn't discussed), rather than how good the stabilisation is when the shaking is within the limits of the shifting mechanism (which is the "stops" rating). This is why "but FF claims the same number of stops on their IBIS - therefore they're just as good" isn't an accurate statement.

Thanks for posting, that's a pretty comprehensive review! Yes, the 1080 looks really good. Interesting that it's essentially the same quality up to 120p and then above it looks lower-quality - I guess maybe the downscaling algorithm couldn't keep up and they had to go for a faster one with lower quality perhaps. Also interesting is that it shoots 48p, which is hugely interesting to me. Over the pandemic I've reflected on how I shoot, and I'd actually thought that I might switch to 48p so that I can slow things down in post if I happened to catch something really cool, but I didn't want to have any mis-matches when using most of the footage at 24p. Ideally I'd shoot in 1080p48 in some flavour of Prores that's good enough quality but not too high bitrate, maybe 422 might be a sweet spot.

@SRV1981 Just catching up on this thread and after I had 4 of your posts quoted, I figured I'd just tag you instead. Lots to talk about here, but I think you're just thinking about things wrong. Here's how I suggest you proceed. Cameras do matter. The problem is that most discussions are very polarised either saying they don't matter at all, or they are the only thing that matters. Neither of these perspectives is true, and more importantly, neither is useful. When people talk about WHY cameras matter, they normally discuss the image, but that's actually not the best way of thinking about them either. The best way to think about cameras is that each camera is a combination of dozens of individual features and functions and attributes. Does it have IBIS? How long does the battery last? How good are the internal preamps for audio? Does it have a punch-in feature to focus and is that feature available while recording? How big is it and how much does it weigh? What is the lens mount? How much DR does it have? What codecs does it offer? etc etc. Buying a camera is about getting the best compromise across all the features that matter to you. You might have a camera that recorded a spectacular image, was small and portable, had all the features of a cinema camera, but if the battery life was 15 minutes then it's completely out of the running based on its one fatal flaw. We should be evaluating cameras based on their biggest weakness for how we shoot, not based on their best feature. Skills matter more. The cameras you're talking about are capable of world-class images, including your Canon that you already own. Please don't take this the wrong way, but the problem you're experiencing is that you aren't capable of world-class images and so that's what's letting you down. I'm also not capable of world-class images, far from it in fact, but I'm perhaps down that path a little further than you are. Video is hard and the path to getting great results is difficult. You're not lighting your videos, and you're not designing the sets and locations either. This makes is harder for you than for people who make sets, light them, and then point the camera at them. I also shoot in completely uncontrolled conditions without permission to be where I am (stealth mode as you call it) so size and appearance also matter to me. Unfortunately, not lighting and designing sets makes it harder still to get the kind of images you want to make. Stop spending money on equipment and start spending time to learn. I mean this literally - don't spend another dollar on equipment. Not one. Your current equipment, your Canon and whatever lenses you have (even if it's just the kit lens) is good enough. By far the biggest limitation in what you're doing currently is your lack of skill. So stop spending money and start spending time. This is actually great news for you. IIRC you said that you're a teacher, and I'm assuming you're not getting paid a large hourly rate, so you probably have far more time than you have money to invest. Here's what I suggest - try and replicate other peoples work. Find a video shot on the same camera as you have, find the nicest shot in it, then try to replicate that shot. Alternatively, you could start with a shot from that video that's the most accessible (eg, a shot of someone standing outside during the day) and replicate that. Do it again with another shot. Do it again and again. You're likely to encounter shots where you're not sure how to replicate it and your attempts to do so fail. In these situations you need to experiment. Just think of every step of the process and think "what if I did this differently". Like, when shooting, what if I expose a little darker or lighter, what if I use a larger aperture or smaller one, what if I use one camera profile or another. What if in Resolve I use this control instead of that control. What if I use this LUT instead of that LUT. What if I use a Colour Space Transform instead of a LUT. What if I do it manually using this control instead of that control. Being able to get a good shot is luck. Being able to get good shots reliably requires skill. That skill requires knowing what to do in each situation and why you would do it. This requires you to essentially explore everything it's possible to do and learn what each option does and which ones work in which situations. Unfortunately this isn't something that can be bought, and it can't even really be taught, it just comes with experience. This sounds daunting, but think about it like this. If you'd have started this 6 years ago, you'd have 6 years of experience, when currently, it sounds like you don't really have much at all (apart from looking at videos and buying cameras).

Man, it's all packed in there. The model comparison (p37) is illuminating and shows how much smaller the bodies of the 35 and Mini really are compared to the previous models. One other thing that I noted from that issue was that they are actively cooling the sensor to maintain an optimum temperature. This makes sense as noise (and probably lots of other things) change with this variable, but AFAIK no other cameras do this? It just shows how much extra stuff is in there compared to MILC cameras, and also the lengths that they're going to for ultimate image quality. I also noted a comment from PotatoJets video where he asks about weather sealing and one of the ARRI guys says that you can basically pour beer into the top of the camera and it won't get to the electronics. Of course that's a joke, but he said that on The Grand Tour they were filming in hugely dusty environments and at the end of the day they'd just put a compressed air nozzle up to the air intake and blow all the dust out of the camera. The fact that didn't completely kill their units really is a tribute to the engineering put into these things.

@SRV1981 Here's the video that Mercer is referring to - it was shot with the BMMCC in Prores. The colour palette on the final film is golden/magenta, but the footage was all over the place with some shots being blue/yellow and some shots being quite green. I've seen footage from an Alexa where the film-makers hired the camera but didn't really know what they were doing (IIRC correctly they self-funded a feature but lived in a town/city where there isn't really any professional film-making done so they didn't have anyone around to learn from). The footage was of a scene that was badly framed, not lit at all (and worse still didn't use the available light in a good way). The result was that the shots looked like bad home video. It literally looked like a random clip of two people sitting in a cafe. The best camera in the world doesn't help you. Perhaps to try and ram this point home, here are some camera tests where they exposed properly and applied the manufactures LUT, but did no colour grading. The USD$16,000 Canon C500ii: The $1500 Sigma FP: The USD$6,500 Sony A1: The USD$6,000 Red Komodo: Notice how they basically all look the same, and how none of them look even remotely like a finished colour graded image? I cannot emphasise this enough, buying "the right" camera and expecting great looking images is like buying "the right" paints and expecting your paintings to be like Leonardo Da Vinci.

Well, I'd agree it's a stress test.... just not of the camera! I'm familiar with Agile in software development and if you're curious I'd recommend you have a read through the sections where the sensor and image processing heads talk about the various processes and the number of variables that go into them getting the best colour. IIRC there are literally dozens of variables (voltages and other electronic aspects) in running the sensor and they tried various combinations to try and optimise the image. I'm guessing that having two-week sprints is what allowed them to go down that path and still have the camera see the light of day - it might have been decades more if it was waterfall!

A simple curve in the highlights will do it, and if the curve requires too much contrast to be added to the highlights then you can augment it by desaturating the highlights a touch too, which is a great trick to blend over the areas where the channels are clipping at different points. Digital sensors all clip very very harshly, it's what the camera/NLE/colourist does after that that creates the rolloff.

I think someone mentioned this in passing earlier in the thread, but I thought I'd link it in case anyone was interested. The FD Times latest issue is 96 pages devoted purely to the Alexa 35 https://www.fdtimes.com/pdfs/free/115FDTimes-June2022-2.04-150.pdf One of the highlights I found was the ARRI head of sensor design demonstrating a new experimental firmware that reduces the weight of the camera for steadicam operators: But seriously folks.... it's a good read.

Are you talking about this one? I figured that I'd have to learn the false colour standard at some point, but seeing that and realising there is no standard, and that some of these are really very unintuitive, maybe I'd just make my own false-colour LUT and get what I need from that!

Highlight rolloff is created in post if you know how to colour grade.

Perhaps a side-note, but on the GH5 that tripod mode is excellent. Not only for completely eliminating motion (as it's designed) but also for doing heavily stabilised pans - It kind of feels like you're dragging the frame to one side, but the results really speak for themselves. If you're using stabilisation for hand-held work then it's worth trying that IBIS mode in whatever camera you have to see if this kind of behaviour is possible.

While I'm not about to buy, I'm also very interested in the quality of the 1080p on GH6. Double the thanks to anyone who shares this 🙂

Yeah, the FX6 is much smaller and the FX3 is really very small in cine camera territory. Even if it doesn't do RAW as Tom mentioned, for action sequences it's got to be very usable, especially considering that up until relatively recently GoPro footage was used in features... and let's just say there aren't any FX3 vs GoPro comparison videos because there'd be no point! Interesting that you draw the line between 'cinema' and not on the basis of RAW recording. I'm not suggesting some other line, I just hadn't thought about drawing it there, but I can see the logic. I think with the increasingly blurred lines between the categories of old, classifying things is becoming increasingly difficult and some definitions are so vague that they've been swallowed by hyperbole already.

The big fat answer is... it depends. One of the most significant factors is sharpening. All the A/B tests that people do leave the 1080p as-is and don't sharpen it to match the extra perceived sharpness of the higher resolution clip. I highly highly highly encourage everyone to do this comparison for themselves so they understand what it actually happening. Now, a better question is, can anyone tell when it's not a comparison and it's just what your film is shot on? I'd say no, considering that the lust for 4K cameras was quickly followed by a lust for diffusion filters to soften the "digital" look they created. .....and the best question is, does it matter to someone who is just watching and enjoying your content? To this I think the answer is quite obviously, no. Of course, the above is a "all else being equal" type of discussion, but all else isn't equal. When you're shooting or editing 1080p then you potentially have less hassles on-set with changing of memory cards, overheating of cameras, etc, meaning you can use the time you have to get one more take of a performance, or work in a way that's less interrupted. In post it either means you can use a cheaper editing setup or a highly spec'd setup will edit in a snappier way, making the tools nicer to use and you will literally create a better edit, and can have more layers of colour grading and more trackers etc so you can colour a bit nicer too. Not having to store such large files (camera media or SSD editing etc) means you can have more money to put into your film, which if you spend it on adding haze or renting a better lighting package or nicer production design will absolutely add to the quality of your output. There's no way in hell that the difference between appropriately sharpened 1080p and 4K can make anything like the improvement that will be made if you: get more/better takes have better production design make a nicer edit with nicer colour grading Resolution is hugely important if all you do is pixel-pee, but if you're interested in creativity it is a very low priority in making an enjoyable end product. (Standard disclaimer - resolution can matter more if your clients care about the spec, if you're doing VFX work like zoom in post by more than 40%, green screening, compositing, doing VR or AR, etc)

The think the other reason is that YT in 4K is far superior to YT in 1080p, but what people don't realise is that it has nothing to do with the resolution and is simply a factor of the bitrate instead. Mix 1080p footage with 4K on a 4K timeline and upload to YT in 4K and it's really quite difficult to be able to spot which clips are which resolution.

There are still some applications that need to go smaller still.... https://ymcinema.com/2022/05/16/meet-the-cockpit-lens-behind-top-gun-maverick/ I don't think you're getting 4 Alexa Mini bodies into the cockpit of a fighter jet, for example. Besides, the smaller they make the whole camera, the smaller they could make the detachable head, and therefore the smaller the places they could get it compared to the body.

Affordable anamorphics are really starting to arrive too, so it makes sense that this would be a new "standard" feature set for mirrorless cameras.

I have no idea what you mean.... They say "cameras are required to have capability to create more impressive and breathtaking expression" I want to create more impressive and breathtaking expression, I mean, don't we all? Finally a manufacturer has stopped using all these nouns all over their press releases and have started talking sense!

I have an enduring interest in tiny cameras with high image quality and every now and then I see a BTS with an odd compact camera in it, and it is often the head of the Venice mounted somewhere and connected by its umbilical cable. I would think it's too common a requirement, but it's not too uncommon either?

I'm guessing you have found people that you can socialise with during the week? I sometimes have breaks between contracts and found that basically everyone I knew was busy working Monday to Friday before 6pm, and were tired after that. I always suspected there was a parallel world of people who worked evenings and nights and weekends and they all just hung out with each other while the business-hours folks were all at work..

I saw that link, but (and I know this might make me sound a little crazy here....) I didn't think that manufacturers announced alliances by wearing t-shirts and posts on rumour sites. Yeah, I know, crazy talk. I guess if @MrSMW says "Panasonic and Leica made a joint announcement that they are 'furthering' their collaboration." then it must be true - regardless of the source!

This is an excellent question. I have been delving into things and I have come to three significant observations: Camera sensors are Linear measurement devices Normal cameras do not look anything like an Alexa - even when shooting RAW still images that aren't debayered and even well within their best operating area (low DR, base ISO, etc) The OLPF and spectral sensitivity of various cameras doesn't change much The problem with these three statements is that they can't all be correct without there being something else going on. The only way I know how to resolve this contradiction is if there is processing going on before the debayering in the ARRI cameras. I don't think the layout presents an enormous challenge for them to be able to put multiple sensors together like the LF does, but you might be right about it being a long wait. The 35 is named to match the 65, and the introduction of the 35 means that now they have cameras over 4K in all sensor sizes. Obviously the new 35 has greater dynamic range and colour depth than the ALEV3 cameras, but considering how much those already had, you could almost just think of the DR on the 35 like a special application camera (like super slow motion) where a production would just use one for those shots.

This seems like a good overview, and includes a few little sections to make sure social-media-bros now some of the basics...