kye

Slashcam footage of BMCC 6K FF camera. Looks pretty nice to me! Especially the anamorphic images 🙂

Recording externally via USB-C seems to be an emerging standard across lots of devices, so rather than have associated products and discussions embedded in various camera threads, I figured a new topic might be easier where all the options can be compared in one place. To kick things off, this popped up in my feed, and happens to work well for iPhones but might also be good for other solutions too. It's a paid video (AD), but product looks interesting, as do the cables, if you're trying to make a compact setup.

Are you sure? I'm going to have to think about that for a couple of weeks and I'll get back to you then.

Since computers transitioned to SSDs, the memory swapping mechanisms (where they put a chunk of memory onto the drive and load another one into RAM when needed) have become very useful and quite high performance. If FCPX is running out of RAM, I wonder if your SSD is running out of swap space somehow. I'm not exactly sure how you'd check these things, but that's worth looking into. A few suggestions.. if your SSD is close to being full, try clearing some extra space and seeing if that helps performance. I'd suggest reviewing FCPX settings to see if there's a setting that limits how much swap space it is allowed to use. Do a google on FCPX RAM usage - I just did a search and there's lots of threads of people talking about it, so there might be good tips in there that might help. Also, some folks report that having files (like source media, proxies, etc) on an external drive rather than the internal drive can improve performance too. I'm not sure of your current setup but that's something to check. Good test, but it might not be enough. With all these companies building in AI to their software now I suspect that RAM requirements might gradually balloon over the next years as they integrate this stuff. This is likely to be the same with CPU/GPU as well. Resolve has quietly introduced a number of AI functions for things like NR and frame interpolation and audio processing etc, and I can tell you these things absolutely KILL your computer. I'm talking where you have it on "Faster" and you can do 24fps, on "Better" and it can do 20fps, and then "Enhanced" (which is the AI one) and it can only manage a frame every few seconds, and will crash 8 times out of 10 trying to export the project. If you can, I suggest giving yourself some head room in the specs, especially as you can't upgrade them later!

I suspect that another variable in your observations will be the particular implementation of the system. One thing that frustrates me no end is how these stabilisation systems are specified. The "5 stops" rating is a measure of how much vibration is left in the image compared to how much the camera was subjected to. That's fine, but in my experience this isn't the factor that is dominant at all. My experience is that if you are able to hold a camera steady and only have a small amount of shake then they all do a good job - where the residual is practically invisible. The situation where the system isn't able to do such a great job is where the amount of camera shake is more than the mechanism is able to eliminate - i.e. the mechanism reaches the mechanical limits of its movement. These are the situations that show the limits to the mechanism, and (I suspect) one of the reasons that MFT is often reported as being better than larger sensors. This would make sense as to get the same range of stabilisation an MFT sensor would only have to move half as far as a FF sensor, and do so with a quarter the weight. So, if the systems reveal themselves at their limits, another aspect is how they behave at their limits. No manufacturer implements this in a way that absorbs all movement and then instantly transitions to having zero impact when the mechanisms boundaries are met, and this would be a terrible design choice, so there needs to be some sort of transition zone between full stabilisation and none, like a 'knee' or 'rolloff', and the characteristic of this function would likely impart some sort of 'feel' to the user I would expect. Also, each of the mechanisms is likely to be quite benign when they are operating well within their range of stabilisation, but the further towards their limits you push, they will reveal various artefacts. The combination of IBIS and ultra-wide-angle lenses is well known with the corners wobbling about like jelly on a rickshaw, but I would imagine that there are other combinations that all have their own issues with differing aesthetics too. The sad thing is that with a very rudimentary testing setup, someone that had access to all the cameras (like DXO or some other tester) would be easily able to measure these things, and give some data around them. I think it would be exceedingly useful to know that one camera was able to stabilise twice as much movement as another camera, despite both of them having the same ratio of reduction ("stops") when both are well within their functional range. There are also other aspects that bear mentioning that just aren't well known, for example that IBIS and DIS stabilises camera roll motion, but OIS does not. With all the reading I do, and with the huge emphasis that my own shooting places on stabilisation, I never knew that and learned it the hard way on a shoot of my own.

A few points.. I have no idea how much is spent, but it's enough for them to shoot everything in 1000fps if they wanted to, but they don't The thread is literally titled "24p is outdated", so translating that to be "old" and "wrong" isn't too far a stretch The firestorm is likely a reaction to the increasing popularity of 30p and 60p video, which just looks awful to many of us... perhaps a good parallel was if restaurants all over the world all started cooking every dish with Kale (or some other ingredient that many people hate) - lots of people who don't mind Kale would just say "eat somewhere else" but it would really be a bad outcome. and then the Kale Lovers Association starts a thread on their cooking forums saying that everything except Kale is crap... that is the response here. The thread didn't "degenerate" into a discussion about art, it actually elevated because creativity and art is the whole point of film-making, which is a point you seem completely impervious to hearing Yes, the goal of the industry is to make money, but in case you didn't notice, emotion is what drives sales. Newspaper headlines, clickbait video titles, social media feuds, etc - these all generate the most sales and clicks because they stimulate the human emotional system. You can think what you like about rationality and realism, but it's not how the vast majority of humans work. A great movie is one where the reaction is "I laughed, I cried", not "I found it to be sensible and factual". Movies make us feel to generate sales. To make money, movies make us feel. To make us feel, movies have to be emotive. To be emotive, they use all the tricks.. one of which, out of literally hundreds that are in use, is 24p.

I'd suggest an alternative approach. You should really be buying 2 x S5iiX cameras, Angenieux EZ-2 15-40mm T2.0 lens, and an Apple Studio with M2 Ultra 24-Core CPU, 64GB Unified RAM, 1TB SSD, which totals USD$21,500 so if you ONLY buy the things you've listed above then you're SAVING a ton of money! I've been watching fashion influencers on TikTok and I'm pretty sure that's how financial management is done these days. If you do enough of these deals, the reward seems to be a big house with sports cars in the driveway.

I'm not saying that any of these are better or worse overall, it's just that they work differently, and for any given situation the pros and cons of one method might be more beneficial than another method. There will be situations where IBIS is the best, situations where OIS is the best, and situations where EIS / DIS is the best. The same will be true for all the combinations too. Throw in gimbals and sliders and cranes and you will still find that the answer of "which is better" will still depend on the circumstances. I just posted because Ty mentioned he wasn't aware of the situations when IBIS is better than DIS, and so I provided an example. There are examples where DIS is better. It's like anything, there's nuance and the more you can understand this then the more you'll be able to optimise your own work. Yeah, ROI doesn't really factor in to these discussions.. if it did, the answer would almost always be that the "Buy nothing" option would be best! Or, if there is money to be invested, it would be better put into marketing or some other skill that can actually move the needle on the financial performance of your business. The cold reality is that the quality of the product isn't actually the limiting factor in most small businesses, which is why the world is full of smooth-talking people selling things that are mediocre at best and driving around in luxury cars.

Like @mercer has said, this is ridiculous. There's two people in this thread who are making post after post about how 24p is old and wrong etc, and then there's the entire industry spending trillions of dollars a year shooting films and who have access to any camera they want and could shoot in any frame rate they liked, and yet they select 24p. So, who is right? Two internet keyboard warriors, or the entire worlds entertainment industry? Please. We did a survey, and you didn't win this "debate".

Time is more relative than you think. For example, there are lots of political parties that still think it's the 1950s!

A realist... I like it!

The key difference is that OIS and IBIS will stabilise during the exposure, and digital IS stabilises afterwards. I did a test some time ago to show what this looks like. This test is deliberately with a long lens, so is somewhat exaggerated, but should be indicative of the issue. If you are using very short shutter speeds then it doesn't matter, it's only when using a 45-360 degree shutter angle that these effects would be visible. Once you know how to recognise it, you see it in YT content every so often, so it does happen in real-life. The combination of action-camera and low-light is particularly succeptible. Digital IS is a great tool, and for some people it's sufficient. Use whatever works for you 🙂

NICE! That sounds like a great way to really help out the folks downstream in the image pipeline without much extra effort. I've never heard of anyone doing that before, and yet it seems so obvious..

Semi-recently a certain cat-lover on YT who shall-not-be-named did a video talking about variable ND filters, the different types, and the different issues that vNDs can have. I was surprised at how complex the situation is and how many things can go wrong when using them. In the tests he showed, which included a range of vNDs, performance didn't seem to correlate much with price, but that might also have been intentional as I think he was launching his own line and therefore had a vested interest in not showing any good ones that are out there. The take-aways that I got were: there's lots happening in there, even from a single vND changing when you turn it lots of things can go wrong, depending on the type used and how they're used better to use fixed NDs or simply to accept the errors and fix what you can in post

Interesting link - I wasn't really that aware of the history of ambient music. I guess many people participate in an art form without really understanding the history and foundational concepts. Thanks for sharing 🙂

I'd settle for people that understand: The equipment has a purpose other than to have steadily increasing specifications That purpose has very little to do with "reality" (whatever the hell that means**) **I think we'd also all benefit from knowing a bit more about the human visual system, which (to be frank) is so bizarre that it's a wonder we can see at all, and (also being frank) people seem to demonstrate virtually no understanding of.

This is all true, but only looks at the creation side. Without humans to experience it, there is no art. Or, at least, not any kind of art that has existed so far. Maybe the robots will love strings of prime numbers or something, who knows. But this is what I base my comment on - that film-making (and all forms of art) are for human consumption. Humans are more than rational beings, we are emotional beings, and perhaps more than that too. So to reduce things down to what is objective is to throw away the entire purpose of art. Creation without subjectivity is science or engineering, not art.

That was from 2012. I've had a number of replacement computers since then! At the time it seemed fine, but that was my first machine with an SSD, so it wasn't hard to impress me back then 🙂

To clarify, in the above I meant that science and everything else does a poor job of making sense of the meaning and morality of life. Obviously, science does a pretty good job of a lot of things, which is why we're able to chat on the internet about film-making and are not rural farmers lamenting the fact that half our children died before the age of 5.

I would argue that art is in the perception of the beholder. I used to write electronic music and was working with a friend on some hypnotic ambient music and we were sitting in silence and listening to the song we were working on when all-of-a-sudden next door locked their car and it gave the "boop-boop" sound, and it fit perfectly with the track. I mean, perfectly. We both looked at each other and immediately set out to create a similar sound to put into the song at that point. I don't expect anyone here to appreciate this because you weren't there, you probably don't like that kind of music, you probably don't think it sounds like art, etc, but in that moment a completely non-creative event created a very aesthetically pleasing result on two listeners engaged in the creation of something that's only purpose was for aesthetic appreciation. I don't have a good definition of art, but if that isn't art then I would argue that nothing is. What this means for AI art vs anything else, I have no idea, but from that experience, I don't think that art requires intent during creation to be perceived as such by the audience. From a practical point of view though, if an AI generated every possible sequence of 2minute digital sound, the percentage that would be enjoyed by anyone would be so low that it's simply impractical to approach it with anything resembling a brute-force strategy. Science does a very poor job of making emotional sense of the experience of life. Science also does a poor job of making sense of the meaning and morality of life (along with everything else). It doesn't matter that you see the world in a certain way, everyone is entitled to their opinions, but you actively refuse to acknowledge that anyone else is different to you, or if you do then you just assume them to be wrong. This is a form of aggressive behaviour, which is why it makes people disagree with you so much. Even your language in the above uses the word "correct" which implies that everything else is "incorrect" and that anyone who thinks differently to you (which is most people here) are wrong. This is a great way to make people dislike you. The more you post, the most I dislike you. You have a lot of knowledge, but you are needlessly making other people angry - you are not convincing anyone of anything. As a practical suggestion, I would recommend you speak only from your personal experience, and try not to criticise things you don't like or don't agree with. If this was a thread about ice cream, and someone said they liked strawberry, no-one would tell them they're wrong for liking strawberry and that the only correct choice is chocolate. This is basically what you are doing, only you're telling us that we're wrong for wanting film-making to be a certain way. Film-making is a creative process that is designed to be enjoyed by the audience. That is a fact. It CAN be used for factual purposes, but this is obviously not the goal of even the majority of film-making, so to judge it on that basis is ridiculous. Please try to adjust your behaviour to be more tolerant of how others see the world.

Interesting stuff. I can imagine that the 'pure' sources of data that are human-only will be worth more and more. I guess that each type of model will have its own weaknesses and blind spots, and it won't be until we get a unified AI model that is fed all the data in all the formats (across all the senses, all the styles, all the topics, etc) that certain integrated elements would be possible for it to understand. It really is down the rabbit hole that we're going. One area that is fascinating to me is that because AI doesn't see the world how we do, it will notice all sorts of patterns that we either miss, or don't pay attention to, or couldn't ever have found. Potentially it could bring enormous knowledge gains about the world and about ourselves. It has the potential for destruction as well, of course, but so much up-side too.

Thanks for sharing. You can really see how the puppy just didn't want to stop!

Depending on the DR of the scene and the DR of the FS7, you might find that to create silhouettes you need to pull the shadows down, which creates a new challenge, as you'll be adding contrast and stretching the image. If this is the case then you might want to test an ETTR exposure vs a normal exposure vs underexposing so that the silhouettes are down near where you want them and then you get more headroom. All these things really are situational, and would be difficult to predict. In reality, you've probably got a decent amount of flexibility, especially if you're not afraid to use a bit of NR. Internet camera folks seem to be allergic to any form of noise in the image, but professional colourists use it regularly, and cinema cameras are a lot noisier than internet people would even believe.

Unless someone has specific experience of this exact challenge, I'd suggest just doing a test prior. Any high DR environment should do, just shoot both profiles and then bring up the shadows in post and see which is cleaner. If the sun is in shot then it's practically an infinite-DR scene, although if the sun is low enough in the sky then you might be able to expose such that everything except the disc of the sun is below clipping and still get non-zero shadow detail. I've seen the latest high-DR cameras like A74 and FX3 do a passable job of having recognisable humans in the foreground with the sun in shot behind them, but it's still likely a stretch for whatever codec and colour space you're working with.

Agreed. Most of the YouTuber crowd I see editing on laptops have some sort of mechanism to attach an SSD to the lid of their machine for editing from. Velcro appears to be popular. I used to edit on my laptop on the train and the seats are often so narrow that any plug sticking out the left or right of the machine would get bumped as people sit down or stand up from sitting next to you. I solved that issue by upgrading to a 1Tb SSD in my machines, which gives enough room to edit a few projects on the go if I need to. I only shoot h264/5 so files are reasonably sized. Most of my hardcore adventures into SSD optimisation were from the first Mac I ever bought - a 64Gb 11" MacBook Air from ~2012. When I bought it I expected to keep my files on external USB sticks but never considered that I'd need more than 64Gb for the OS and applications alone. I ended up buying a slim-fitting USB drive that basically didn't protrude from the chassis and mapped a bunch of system folders to it. I think I even ended up with the entire /Applications folder on it. It worked well but I made sure to buy a larger drive next upgrade!