kye

For our personal lives, we really only need food, clothing and shelter. We got by just fine even before we had language, so I'm pretty sure that there is no minimum number of megapixels in that equation... But, assuming that you're asking if you can print things and enjoy them, then I'd say there's no lower limit except what limits you would have to your own enjoyment. After all, mosaics have been a form of art for thousands of years and no-one has started complaining that they don't look photorealistic, and people aren't complaining about paintings not being photorealistic either. It's controversial, within online photography communities anyway, but here is Ken Rockwells thoughts on the matter: https://www.kenrockwell.com/tech/mpmyth.htm A few highlights include the idea that the larger you print something, the further you tend to stand from it: Today, even the cheapest cameras have at least 5 or 6 MP, which enough for any size print. How? Simple: when you print three-feet (1m) wide, you stand further back. Print a billboard, and you stand 100 feet back. 6MP is plenty. and also practical experience: Even when megapixels mattered, there was little visible difference between cameras with seemingly different ratings. For instance, a 3 MP camera pretty much looks the same as a 6 MP camera, even when blown up to 12 x 18" (30x50cm)! I know because I've done this. Have you? NY Times tech writer David Pogue did this hereand here and saw the same thing — nothing! Joe Holmes' limited-edition 13 x 19" prints of his American Museum of Natural History series sell at Manhattan's Jen Bekman Gallery for $650 each. They're made on a 6MP D70. I would argue that what is visible will depend on the printing technology you use. A canvas print will hide any blurriness or detail that is in the image, but a metal print will accentuate it. I have visited a house with a couple of large metal prints (maybe 3' x 5') of very high resolution shots of a rainforest with deep DoF and dappled light filtering down to the ferns and rocks by a stream. The resolution was palpable, so much so that both myself and the other non-photography people were all impressed by the print. The discussion included 'oohs and aahs' around how much it cost, how difficult it was to hang, how they had to get little custom lights to shine on it like in an art gallery, how impressive it was, etc. What was missing, however, was any discussion or appreciation of the image. All discussion was of the technology. The sharpness was distracting.

12MP is 4K in 4:3... with a good enough codec, that's 1.5K more than most people need!! The OG Alexa was 2.5K and most cameras these days still can't create an image as good....

Also worth pointing out (and you may already know this) but the IS Boost mode does a spectacular job of shots with gradual movement. I've used it to get buttery smooth pans before. The feeling is kind of like moving through treacle, you have to pull the cameras image where you want it to go but it's super smooth.

I can understand that it's an emotional thing, but I feel I have moved past attachment to any particular format or technology, and have become focused on the experience and the outputs. MFT is great, I have 5 MFT cameras and some really nice lenses, but it's a means to an end. I recently discovered I prefer the size of the GX85 to the GH5, and happily took the loss in image quality for the benefits in the shooting experience. If someone created a camera that was slightly smaller than the GX85 and had equal or better image quality and features, I'd consider switching, not giving the slightest consideration if the sensor was 1/2.5" or Medium Format. The S5 is a great camera, take lots of great images and enjoy it! Enjoy the images you captured with the MFT gear too. Cameras matter, but the better you become at using them, the less they restrict the quality of the work.

In terms of assessing the maximum size for printing, that's another whole debate, but a way to test it is to pull the image up on a 4K display, and then zoom in to the image until you start seeing the video compression artefacts, then reduce the size down a bit, and then measure how large the image would be if you printed it that large. Remember also that the larger you print an image the further back you will tend to view it from. A friend of mine won a certificate and got one of those canvas prints about 2' x 2' across. She sent in an early iPhone photo that was a close-up of one of her kids faces (it was a lovely photo) and the printing place said it wasn't "high enough quality" (read: not enough megapixels) and she had to insist they print it. She hung it high up on their photo wall and due to how the furniture was arranged you could see it throughout their open-plan living/dining/kitchen area, but you wouldn't have looked at it from closer than about 5' away, and even then you'd be looking up at it significantly. It looked great. Had it been taken with a high megapixel camera would it have been sharper? Sure. Would she have ever been able to afford a high megapixel camera? No. Would you ever take that kind of photo with a "real" camera? Unlikely.

It should be pointed out that it is not mandatory to have a short shutter speed for photos. Let's review two examples. The first is this guy, who is obviously playing at an extreme level here, with huge energy and drive: It is an incredible photo, there's no doubt. But does it convey the sense that he's pushing himself to the limit? The more I look at it the more it looks like it could be a still life, maybe he was on wires and it's a setup. In a sense, it implies motion but doesn't actually express any. Contrast that feeling to this image: There is no denying this. Not only is it a great photo, and not only does it show that world-class people are pushing themselves to the limit (the three guys on the right definitely are!) but it shows the results of that effort. Now scroll back up to the first image - what is the energy level of the first image now? I would suggest that the obsession with short shutter speeds is part of the same obsession with getting "sharp" images, which is driving lenses to be clinical, sensors to be enormous, megapixels to be endless, computers to be behemoths, and images to be soul-less. To be a bit practical, there are likely limits to how much blur you want in a scene, and it's relative to the amount of motion involved, which often varies quite significantly, even from moment to moment in a lot of sports. Some more examples of varying amounts of blur.. This one is a great image, but if it was important who the defender was then it's too blurry. If not, then maybe not.. Sometimes it's not important to get anything completely sharp.. images without feeling are of limited value, and I find that motion is full of feeling. Here's one that is full of emotion (especially if you knew the subjects): Also worth mentioning is that you don't always have to have the subject still and the background blurred, it can be the other way around, but it changes the subject of the image somewhat. and with video you're taking lots and lots of photos so the creativity is endless... Remember - don't let the technical aspects blind you to the point of capturing things in the first place.

I did a "Send Feedback", which thanks to the ability to attach a screenshot was pretty straightforward in showing the issue. Interestingly, there's no ticket number or anything - you pitch your feedback into the void and that's it apparently. As a paying customer normally that buys you some kind of human support, I guess not with YT? Or maybe not for me as I'm in the family account but I'm not the account that paid for it.

Yeah, that guy isn't the most thorough reviewer, but of course he's probably a better barometer of what the YT Camera Bros think of it than someone like Gerald.

It seems that it's only being tested on web, and that it still needs more work... I click on it, it says I need Premium membership (which is odd considering I'm logged in and the YT logo says "Premium"), then when I click on the "Get Premium" button it takes me to a page that says the offer isn't available and I need to go into settings to manage my subscriptions and purchases. I'm part of a YT Premium Family Sharing account, so maybe that's the issue.

Just saw a new option appear in YT.... Some quick googling revealed that it's only for YouTube Premium (paid) membership, which I am a member. Anyone else seen this? Do we know what is going on under the hood?

The few YT camera hipsters I've seen review it had positive things to say... TL;DR: Stabilisation - "gone are the days of needing a gimbal" Colours and DR - "Super easy to handle" "easy to colour grade and get a nice aesthetic" (buy my LUT pack) AF - "GH5 was terrible" "S5 is so reliable" "it sticks onto him the whole time" "the AF is unbelievable" SSD recording - "prores is easy to work with" "speed up your workflow" "This is one of the sexiest looking cameras out there" "I would highly recommend checking it out".

Are you perhaps thinking of the thread where we posted our submissions to the EOSHD $200 camera challenge? or the thread with the results? Submission thread: Voting / results / discussion thread: I love that my butchered GoPro is the thumbnail of that thread! Not sure if everyone else would see the same shot, but I'm pretty sure that the camera gods would be happy knowing it died trying to be converted to a D-mount camera.

I agree, zooming definitely increases the visibility of the noise in the image. One thing I think worth remembering is that cinema cameras have traditionally been quite noisy and noise reduction used to be a standard first processing step in all colour grading workflows. It is only with the huge advancements in low-noise sensors, which allowed the incredible low-light performance that high-end hybrid cameras have today, that this noise has been reduced and people have gotten used to seeing clean images come straight from the camera. I just watched this video recently, and it is a rare example of high-quality ungraded footage zoomed right into. He doesn't mention what camera the footage is from, but Cullen is a real colourist (not just someone who pretends to be one on YT), so the footage is likely from a reputable source. It's also worth mentioning that the streaming compression that YT and others utilise has so little bitrate that it effectively works as a heavy NR filter. This really has to be seen to be believed, and I've done tests myself where I add grain to finished footage in different amounts, uploaded the footage to YT, then compared the YT output to the original files - the results are almost extreme. So this also plays a role in determining what noise levels are acceptable.

Interesting comparison. For anyone not aware, it's worth noting that since the iPhone 12, they quietly upgraded to doing auto-HDR and the h265 files are also 10-bit, as well with the prores files being 10-bit of course. It is interesting how little they promoted this - I have an iPhone 12 mini and didn't even know about it for the first year of owning it! This is what really unlocked the iPhone for me - I have used iPhones since the 8 for video but prior to the HDR / 10-bit update it really wasn't a competitor to anything other than cheap camcorders, as it lacked the DR to record without over-contrast or clipped highlights and bit-depth to really be able to manipulate it sufficiently in post to remove the Apple colour science and curve.

I was quite excited when Apple introduced prores support, thinking that it would mean that images would be more organic and less digital-looking, but wow was I wrong - they were more processed than ever. I think the truth is hiding partly behind that processing, where the RAW image from them would struggle in comparison. I shot clips at night on my iPhone as well as other cameras, and I noticed that the wide angle camera on my iPhone was seriously bad after sunset - even for my own home video standards. I did a low-light test and worked out that the iPhone 12 mini normal camera had about the same amount of noise as the GX85 at F2.8, but the iPhone wide camera was equivalent to the GX85 at F8! and that was AFTER the iPhone processing, which would be way heavier than the GX85 which had NR set to -5. Apple says that the wide camera is F2.4, so basically the GX85 has about a 3-stop noise advantage. There are lots of situations where F8 isn't enough exposure. I'm not sure where we're at in terms of low light performance, but this article says that this Canon camera at ISO 4 million can detect single photons, so there is a definite limit to how large each photosite can be. Of course, more MP = less light in each one. In terms of my travels, I still prefer the GX85 to the iPhone when I can - being a dedicated camera it has lots of advantages over a smartphone.

One of the (potentially numerous) differences when swapping from 4K to 6K (on a 4K timeline) is that the downsampling goes from being before the compression to afterwards. Compression is one of the significant contributors to the digital look IMHO (just compare RAW vs compressed images side by side), and if you can downsample from 6K to 4K in post then you're downsampling and also interpolating (blurring) the compression artefacts that happen on edges. I would favour a workflow where the image is debayered, downscaled, then recorded to SD with very low or even no compression. It would effectively have the benefits of downsampling and the benefits of RAW, but without the huge file sizes of RAW at the native resolution of the sensor. Of course, no-one else thinks this is a good idea, so....

I find IBIS to be critical to my workflow personally, as I shoot exclusively handheld and do so while walking etc, but I definitely understand that I'm in the minority. But yeah, having that eND would be spectacular - set to auto-ISO / auto-eND / 180 shutter / desired aperture for background defocus and you're good to go!

Lots of them are, but as you brought it up, here are my impressions on the above: It's still far too sharp to be convincing - I noticed this in the first few seconds of the video - TBH my first impression was "is this the before or after footage?".. of course, once you see the before footage then it's obvious, but it didn't immediately look like film either The motion is still choppy with very short shutter speeds - this would require an VND which isn't so easy to attach to your phone I think that people have some sort of hangup about resolution these days and as such don't blur things enough. For example, here are three closeups from the above video. The original footage from the iPhone: Their processed version: A couple of examples from Catch Me If You Can, which was their reference film: As you can see, their processed version is far better, but they didn't go far enough. I've developed my own power grade to "de-awesome" the footage from my iPhone 12 Mini and detailed that process here, but here's a few example frames: I'm not trying to emulate film, I'm just trying to make it match my GX85. Here are a couple of GX85 shots SOOC/ungraded for comparison: I haven't got the footage handy for the above shots, but here are a few before/afters on the iPhone from my latest project from South Korea. iPhone SOOC: iPhone Graded: iPhone SOOC: iPhone Graded: iPhone SOOC: iPhone Graded: I'm happy with the results - they're somewhere in between the native iPhone look (which I've named "MAXIMUM AWESOME BRO BANGER FOR THE 'GRAM") and a vintage cinema look. My goal was to make the camera neutral and disappear so that you don't think about it - neither great nor terrible. Going back to Dehancer / Filmbox / Filmconvert / etc.. these are great plugins actually and I would recommend them to people if they want the look. I didn't go with them because I wanted to build the skills myself, so essentially I'm doing it the hard way lol. The only thing I'd really recommend is for people to actually look at real film in detail, rather than just playing with the knobs until it looks kind of what they think that film might have looked like the last time they looked which wasn't recently... I keep banging on about it because it's obvious people have forgotten what it really looks like, or never knew in the first place. And while they're actually looking at real examples of film, they should look at real examples of digital from Hollywood too - even those are far less sharp than people think. The "cinematic" videos on YT are all so much sharper than anything being screened in cinemas that it's practically a joke, except the YT people aren't in on it.

I have a vague recollection that some cameras were able to tag or put comments onto specific images as you were taking them (or just afterwards) and that the info would be embedded in the images metadata or some such. It might even have been you that mentioned it. Anyway, if you were taking full-res stills and they were being sent instantly to the photo/video village for quick editing and upload to the networks and social media channels, then you could simply tag which images you thought should be cropped? If you are shooting full-res images then cropping in post is the same as cropping in-camera I would imagine?

Great post - I just wanted to add to this from a colour grading perspective. I did a colour grading masterclass with Walter Volpatto, a hugely respected senior colourist at Company 3, and it changed my entire concept of how colour grading should be approached. His basic concept was this: You transform the cameras footage into the right colour space so it can be viewed, and apply any colour treatments that the Director has indicated (like replicating view LUTs or PFEs etc) - this is for the whole timeline You make a QC pass of the footage to make sure it's all good and perhaps even out any small irregularities (e.g if they picked up some outdoor shots and the light changed) At this stage you might develop a few look ideas in preparation for the first review with the Director Then you work with the Director to implement their vision The idea was that they would have likely shot everything with the lighting ratios and levels that they wanted, so all you have to do is transform it and then you can fine-tune from there. Contrary to the BS process of grading each shot in sequence that YouTube colourists seem to follow, this process gives you a watchable film in a day or so. Then you work on the overall look, then perhaps apply different variations on a location/scene basis, and then fine-tune particular shots if you have time. He was absolutely clear that the job of the colourist was to simply help the Director get one step closer to realising their vision, the last thing you want to do as a colourist is to try and get noticed. It really introduced me to the concept that they chose the camera package and lenses based on the look they wanted, then lit each scene according to the creative intent from the Director, and so the job of a colourist is to take the creativity that is encapsulated into the files and transform them in such a way that the overall rendering is faithful to what the Director and Cinematographer were thinking would happen to their footage after it had been shot. Walter mentioned that you can literally colour grade a whole feature in under a week if that's how the Director likes to operate. I have taken to this process in my own work now too. I build a node tree that transforms the footage, and applies whatever specific look elements I want from each camera I shoot with, and then it's simply a matter of performing some overall adjustments to the look, and then fine-tuning each shot to make them blend together nicely. In this way, I think that the process of getting things right up front probably hasn't changed much for a large percentage of productions.

Very nice!!

I don't have any technical information or data to confirm this, but my theory is that Sony sensors are technically inferior in some regard, and this is what prevents the more organic look from the majority of modern cameras. My understanding is that these are the only non-Sony-sensor video-capable cameras: ARRI (partners with ON Semiconductor to make sensors to their own demanding specifications) Canon BM OG cameras (OG BMPCC and BMMCC) Digital Bolex (Kodak CCD sensor) Nikon (a minority of models - link) RED (there's lots of talk about who makes their sensors, but it's hard to know what is true, so I mention them for completeness but won't discuss them further) What is interesting about this list is that, to my understanding at least, all these cameras are borderline legendary for how organic and non-digital their images look. Canon deserves a special mention as its earlier sensors are really only evaluated when the footage is viewed RAW, either taking RAW stills or using the Magic Lantern hack). There are questions if it could be related to CCD vs CMOS sensor tech. This thread about CCD sensors from Pentax contains dozens/hundreds of images that make a compelling case for that theory. I have both the OG BMPCC and BMMCC as well as the GX85, GH5, Canon XC10, Canon 700D, and other lesser cameras and have tried on many occasions to match those to the images from the BM cameras, and although I managed to get the images to look almost indistinguishable (in specific situations) they still lacked the organic nature that seems to be shared across the range of cameras mentioned above. I regularly see images from my own work with the Panasonic cameras and my iPhone, as well as OG BM camera images and RAW images from the 5Dmk3 and Magic Lantern (courtesy of @mercer) and once you get dialled into the look of these non-Sony sensors, it's hard not to notice that the rest all look digital in some way that isn't desirable.

I started lots of threads about the non-camera aspects of film-making. Minimal engagement at best.

I agree with you, but I would like to clarify that my problem isn't that we talk about the tools, it's that we ONLY talk about the tools.

I also wish they'd continue pursuing the look. It wouldn't be too difficult either, as you can design a camera that has both looks - modern and classic. It's been shown that the Alexa is simply a high-quality Linear photometry device, and that the colour science is in the processing rather than in the camera. This leads me to think that you should be able to design a sensor to be highly-linear and then just have processing presets that let you choose a more modern image vs a more organic one. Depending on the sensor tech, you could even design a sensor to have a 1:1 readout for the modern look, and some kind of alternative read-out where the resolution is reduced on-sensor to match the look of film, which might even be a faster read-out mode and give the benefit of reduced RS. Obviously I'm not a sensor designer, so this might be a technical fantasy, but it would be worth exploring. Otherwise, cameras like the GH5 have shown that you can re-sample (either down or up) in-camera, so that sort of processing could be used to tune the image for a more organic look.