Red(now Nikon) doesn't own the patent to RAW recording. They own the patent to compressed mosaic raw recording; however even that is debatable as it's been shown that the technology both existed before they claim inventing it, and Red failed to file their patent within the legal time frame required for it to be legitimate. While a lot of people think Nikon won their lawsuit by simply buying RED. The flaw with that argument is that Nikon wasn't the first to beat RED's patent, Jinni Tech was. Jinni Tech managed to get red to drop their patent suit before it ever went to court; however, Jinni Tech was also so small that they ultimately went under from the post expense of the battle, but they did manage to get Red to drop the law suit before that happened. AKA they technically won. Nikon was literally the first big company that was taking the Jinni tech law suit approach for their battle as well, and many felt that they were likely going to win, and assumed that was why RED dropped the law suit, but then it was assumed it was because Nikon bought RED instead; however, Nikon bought RED for much less than many thought it was worth, and because no one who has used the Jinni Tech argument has ever made it to court the patent still remains in place, at least until it either expires in 2028 or Nikon attempts to renew it. That said it would be hypocritical of Nikon to pursue the patent given their arguments prior to the acquisition of RED; however, they might want to try and recoup some of the expense of the purchase and hope no one really tests the patent all the way to court, as a functioning patent generating money is better than testing it in court only to have it fail.
@@easystreetphoto2401 That isn't entirely true. Supposedly a patent owner can modify their existing patent giving it some new patent-able feature, and tie it with the existing patent thereby extending the patent's life. That is exactly how RED claimed their patent was valid, because there were aspects of the RAW format that we not originally in their patent application, but they extended it to include it. The only potential issue is that they had announced and started publicly selling the potential for the hardware over a year before applying for the actual patent, and never disclosed that fact during the patent application. That is the number one reason both Jinni Tech and Nikon were arguing the patent was invalid; however, RED had never let either of those cases go to court, and no one prior to Jinni Tech attempted to use that point to fight the patent. Other attempts to counter the RED patent had mostly been based around the idea of how obvious compressed RAW was to engineers who had already been using mosaic RAW for stills, and that didn't fly in the courts.
The end result with the lego camera is actually pretty damn impressive. I love that modularity in tech is improving and I'll be honest, I didn't even think about how useful it'd be for things like homemade movies on the cheap. Great vid
Okay, so the crop of the sensor, if I've got the numbers correct, is 6.92. That means that 50mm lens FoV equivalent of a standard Super35 camera would be roughly the same as a 231mm super telephoto lens. That isn't in the ideal range of most cinema glass even for even portrait footage. One is going to want an 11mm for typical 50mm FoV super35 equivalent portrait lens, 7mm to 8mm for a typical 35mm FoV standard lens, a 6mm for roughly a super35 wide angle 28mm equivalent, a 4mm for a 18mm-19mm super35 super wide angle equivalent. Interestingly there is a single lens that can cover almost every one of these focal lengths, is pretty fast with a fixed focal length of up to f/1.6; however the lens doesn't have a filter ring so one would have to built them self a "matte box" equivalent solution and add some IR cut along with ND to get something usable out of the lens unless they want to crank down their aperture size. Also there is very little know about how soft this lens is wide open, but it's probably gong to need to have it's aperture dropped a bit to get a decently sharp image, but for the budget cinematographer who is learning to shoot with this budget camera it is the most ideal solution. Just as a thought on a filter solution, one might try getting a box of step up ring adapters and seeing if one can slide over the outside of the lens, even with a little black electrical tape this could get one a filtering ring solution for roughly $10-$20 and one could match an ir cut and maybe variable ND for whatever happens to be the best budet price. Anyhow, here is the best budget single zoom lens for the entire range: 4-12mm = Arducam 4-12mm f/1.6 - ~$50
@@Andoonline 5.6 * 16 = 89.6mm on Full Frame, and is definitely on the upper end of what's considered a portrait lens. Typically they range between 75mm-100mm with 80mm-85mm often considered one of the most flattering ranges. For Super35mm (1.5 crop) equivalence a 16mm 5.6 crop lens is more like a 60mm lens on that format. That's a good lens option to have.
Well yeah and thats obviously where things like focal reduces help to an extent or using broadcast lenses or old super 8 and super 16 Bolex lenses and alike. Also the crop factor sort of depends for how you shoot, aspect ratio and alike. Not that the crop factor itself changes but how you implement it changes
First time I've heard Scanazon Crinme 😂 going to use that from now on. Awesome video, I've been wanting to make a CinePi for awhile, I'm waiting to see AltCine's version!!
I heard the reason that blackmagic RAW is allowed is because BMD proved that BRAW was made before RED got it's patent, which just makes me question how RED (now under Nikon) got the patent in the first place
4:31 This is funny, Sony, ARRI, Blackmagic and others all have RAW codecs and or encoder packs old high end pro kit from the 2000/2010s still has massive life in it, and Cinema DNG 12-bit support, there is now also FFV1 from FFmpeg for a lossless compressed up to 16-bit 4:4:4 which is actually a bit better then CDNG, but cost of cameras becomes a rounding error when storage costs are factored into the operations.
The moment you said Magic Lantern I immediately wanted to tell you to try crop mood on T4i/t5i or an EOS M. It's legitimately mindblowing how well it works. I can record 2.5K RAW in 14bit with real time preview
This project is awesome! Building your own cinema camera with off-the-shelf parts is super cool. I had no idea you could do so much with a Raspberry Pi. And calling Amazon Prime ‘scamazon crime’ - that really cracked me up! It’s great to see such an involved community and the Discord for discussing and improving these builds. Open-source projects like this really make advanced tech accessible to everyone. I’ve got a few 3D printers and would be happy to print a frame for free to keep the project going. Just let me know if anyone needs help with that!
Great video! As someone trying to build a camera from scratch, and having done a bit of research, most sensor manufacturers do not provide data sheets for using their sensors, but companies like omni-vision, on semi do. A lot of great sensors and they use standard MIPI or parallel etc., communication protocols which are well documented. The hard part (& expensive) is getting an FPGA and program it to transfer all that data to a memory (for video). There are some great APS-C CCD sensors from retired DSLRs (for example Canon T2i) that are analog out that are very capable. With additional ADC circuitry, we should be able to acquire still images even with STM32H7 series that go up to 500 MHz! My theory is, we actually can make an open-source “cinema camera” that isn’t just a camera module attachment that limits the video quality… some random thoughts 😂
You should check out what both Elphel and Apertus have done, not to forget the Sitina1, all open hardware designs. The main difference with the plug and play mentality is that it would be really raw-raw, handling the noise issues, etc. What most people would want in an API to control their camera, and not doing sensor plus optics theirselves.
Thank you for an amazing video. I used Magic Lantern a few years ago and was quite impressed. I do have a question though: any chance you can share some of the raw data? The MLV workflow I had back in the day was... painful. I'm considering building this CinePi camera, but would love to see some sample footage to get a feeling of the output file format and how I would leverage DaVinci Resolve to color grade it. Thanks again for a high-quality video and you've gained a new subscriber :)
I was looking into this kind of thing a few years ago. Sony has a medium format 6.5 cm x 6.5 cm sensor module that I was really keen on trying to do this with. Instead of a pi, full blown GPU accelerated encode on x86. Basically unless I ordered like 5000 camera modules, they were not interested in selling to an individual. And the re-sellers did not stock that because 1) most of these are used in scientific or manufacturing, where such a giant size is not really that important aside from astronomy. 2) trying to license software that played nice with high bit rates and whatnot was a real difficult task as, yet again, most companies buying these modules for their own uses, has a R&D team to do the software with their own FPGAs. 3) Xilinx technically made some co-processors FPGA thing that the Sony camera module worked nice with, but not the larger size sensors, again, for the aforementioned reasons. Seeing this gives me hope that maybe a DIY medium format sensor with GPIO to x86 PCIe could still happen in the DIY space, but it is still very difficult
In case anyone finds this and is curious: the sensor is Sony IMX411 BSI . 151 MP. But if you look at the data sheet, it can only do 6.3 fps in I believe monochrome 16 bit and 3.2/3.1 fps in a 16 bit colour. Ximea sells them with a kind of module interface deal. But yeah, stand alone module to do a high speed full sensor 16 bit depth would be insane
Your chassis makes me think of a software project (if it doesn't already exist): a simplified cad package that takes a lego build and converts it into a 3d printable cad file, so you can prototype the dimensions as part of the design process.
I really should have just gone to Home Depot, and figured out something there. The Lego was more to convey just how much I don’t have access to a 3d printer,and how unwilling I am to buy one… at the moment 😂
And I only scratched the surface and talked about what is considered the “easy stuff”. Just wait until the pi 5 images come out and the larger inch plus sized sensors become readily available, as well as some off the shelf housing projects it’s gonna be… amazing.
I miss being this excited. I had tried to do what theyve done with cinepi with other machine vision cameras and janky af software. Im glad to see where we have all come to. Cant wait to see updates.
I wish Sony LTY 900 DCG sensor was available to public to purchase. Till then I feel Android RAW video is holding strong as we can encode in ProRes HQ on fly using a powerful Snapdragon CPU encoding.
That is so funny I am literally in the process of ordering my CinePi hardware. A Pi 5 and a Sony IMX585 4 channel CSI-2 from Soho(Japanese Company). Love your video!!! I built a Lego Pi Case too. They Rock!!!
Well it’s possible through building your own sensor board like a starlight eye, but also at some point, I believe an upgrade to the HQ camera module will be releasing so I would wait for that
interesting, I wonder, could you run the sensor further away from a full-frame lens.. and would that result in more light per sensor element? rather than sub sampling a much larger sensor, you'd be capturing all of the light from the lens, might have really nice low light.. maybe not compared to the bleeding edge, but compared to my gear 😆
i've been wanting to get my holga set up with a pi and digital camera. could be fun, but the quality will be dismal, not much that'd fit in the space can beat film.
Hi, what patent protects is _commercial_ exploitation, everyone is free to use one to construct and use a device, what is forbidden is to sale it commercially speaking - this is valid in each and every country that have patents. On the other hand, granting a patent for raw recording just means the patent office is more a psychiatric institution that a responsible administration (not to talk about software patents, which are preposterous and, as usual, a brake oriented toward small and disruptive companies).
4:16 raw is just a digital form of the raw sensor data, I don't recognize any claims a company tries to have over it. Sure they may whine about it but at the end of the day it's your sensor data.
@@Andoonline realistically they can't do much if you don't flat out copy their software implementation, the biggest issue with raw recording regulardless of format is going to be bandwidth overhead and sheer compute, but my area is more the hardware side of things.
@@deltacx1059true but I looked a little into that route and found it rather pricy however I did forget to mention that many public libraries these days do let you do some 3-D printing so that could be a route to look into
@@Andoonline it is pretty strange in terms of pricing, some local places I checked will do it dirt cheap when it comes to cutting plate parts but most services youtubers advertised as being cheap end up being stupidly expensive, sometimes to the point where you could buy a machine and do it yourself for cheaper. (Which defeats the point of said service)
And if I have it my way we’ll be able to just take their sensors off their cameras and get access to their color science and put it on ours… one day… maybe… lol 😂
it's a fun thing to try out i think, but it being able to record raw video doesnt make it a cinema camera... first of all the sensor is too small for that methinks but most importantly you don't talk about the reliability of the camera. will it work at all times or will it randomly stop working or bug, corrupt your files, not save properly etc... cause thats the most important part in a professional tool. you dont want it to prevent you from doing what you want. also you mentionned a usb microphone but that is very much not a viable i/o for professional audio recording. My point is it's a fun project and there's definitely value in it but calling it a "cinema camera" is definitely a stretch. also thats just me but a good lens on a bad camera is miles better than a bad lens on a good camera.
Believe me, I understand everything you are saying about it. But please check out the discord. I could never do this project justice. This project goes way beyond anything I mentioned and anything I included , and the camera can be about as reliable as you can build it, and I really promise that there are configurations that you can 100% call a cinema camera, But also I really do value your opinion on it anyways, thank you.
As a passion project it’s really awesome to be able to build your own video camera from off the shelf parts but it is impossible to compare these DIY cameras with ARRI or RED. Both companies have spent millions on R&D on sensor, colour science, compression, power consumption and more. You’re NOT going to get the same results with resolutions, skin tones, build quality and even compatibility with other industry standard equipment. It’s a great idea but as with most things, you do get what you pay for with film equipment. I challenge anyone to build a camera that would be able to capture the “look” , bit depth and resolution of an ARRI Alexa with a rugged design, a v-lok battery mount and a PL Cine camera mount
"Cinema camera" implies cinema quality. That isn't just resolution, it's image quality, good performance in low light, good depth of field, good color reproduction, the ability to pair the sensor up with good glass. The Pi cam sensor is the same kind of stock sensor from ultra cheap mobile phones, there is nothing cinematic about it.
You can get nice picture out of it. But so can you with most modern sensors. But I kind of agree that it implies something else. Even if they made an Imax movie with GoPro 11 most people would not call that a cinema camera. Still it is a cool project I just wish it could handle more interesting sensor. What would be nice is an open source external recorder. Or a 3chip camera
I mean there’s the imx 585, or the starlight one inch. movies have been shot on 16 and super 16mm film and those look great and even look great upscaled to 4k from a 1080p remaster as there is plenty of detail to work off of. “Cinema camera” I don’t think is necessarily an image quality standard, as I think it has more to do with flexibility. However quality does play into why people use a camera for shooting a movie at all. But anyways the cinepi community is always adding more feature’s and more support for more hardware. so i don’t see why we won’t see a super 35 sensor on a cine pi at some point, and I think someone has even gotten a sensor close to that working. This video does not do the community justice as things change so rapidly and its goals are sky high.
I was very frustrated with the file mess of magic lantern on my eos M, its what drove me to the Black magic ecosystem having ef lenses for my Canon t6i, Missed a magic lantern rebel camera by this much 🤌
As someone who uses external file sorting software, so I don’t have to deal with exactly that. I understand your pain. Especially since my 7D’s CMOS is dead, sorting by date at the moment is kinda a pain . 😂
Red(now Nikon) doesn't own the patent to RAW recording. They own the patent to compressed mosaic raw recording; however even that is debatable as it's been shown that the technology both existed before they claim inventing it, and Red failed to file their patent within the legal time frame required for it to be legitimate. While a lot of people think Nikon won their lawsuit by simply buying RED. The flaw with that argument is that Nikon wasn't the first to beat RED's patent, Jinni Tech was.
Jinni Tech managed to get red to drop their patent suit before it ever went to court; however, Jinni Tech was also so small that they ultimately went under from the post expense of the battle, but they did manage to get Red to drop the law suit before that happened. AKA they technically won. Nikon was literally the first big company that was taking the Jinni tech law suit approach for their battle as well, and many felt that they were likely going to win, and assumed that was why RED dropped the law suit, but then it was assumed it was because Nikon bought RED instead; however, Nikon bought RED for much less than many thought it was worth, and because no one who has used the Jinni Tech argument has ever made it to court the patent still remains in place, at least until it either expires in 2028 or Nikon attempts to renew it.
That said it would be hypocritical of Nikon to pursue the patent given their arguments prior to the acquisition of RED; however, they might want to try and recoup some of the expense of the purchase and hope no one really tests the patent all the way to court, as a functioning patent generating money is better than testing it in court only to have it fail.
I wish I had included all of this, especially since some of my wrong information was definitely out of date. thank you for that.
patents can't be renewed
@@easystreetphoto2401 That isn't entirely true. Supposedly a patent owner can modify their existing patent giving it some new patent-able feature, and tie it with the existing patent thereby extending the patent's life. That is exactly how RED claimed their patent was valid, because there were aspects of the RAW format that we not originally in their patent application, but they extended it to include it. The only potential issue is that they had announced and started publicly selling the potential for the hardware over a year before applying for the actual patent, and never disclosed that fact during the patent application. That is the number one reason both Jinni Tech and Nikon were arguing the patent was invalid; however, RED had never let either of those cases go to court, and no one prior to Jinni Tech attempted to use that point to fight the patent. Other attempts to counter the RED patent had mostly been based around the idea of how obvious compressed RAW was to engineers who had already been using mosaic RAW for stills, and that didn't fly in the courts.
Paragraphs, man, pa-ra-graphs! It just takes a couple of newlines to turn that rambling mess of yours into something readable.
@@EvenTheDogAgrees Happy? ;)
"building a camera is like building a pc" 🙌🏾🙌🏾 you're a breath of fresh air.
❤
u already kno bout that life😂👌
I mean what I Truly want to build is a camcorder with interchangeable lenses and raw recording roughly the size of a vx1000
Can't wait for the AltCinePi!
The end result with the lego camera is actually pretty damn impressive. I love that modularity in tech is improving and I'll be honest, I didn't even think about how useful it'd be for things like homemade movies on the cheap. Great vid
Bruh I watched through the entirety of this and just noticed this has 900 views. Insanely underrated content. Kerri it up!
Well watching the whole thing will boost it for others to see. Thanks a lot!
I worked very hard thank you
Okay, so the crop of the sensor, if I've got the numbers correct, is 6.92. That means that 50mm lens FoV equivalent of a standard Super35 camera would be roughly the same as a 231mm super telephoto lens. That isn't in the ideal range of most cinema glass even for even portrait footage. One is going to want an 11mm for typical 50mm FoV super35 equivalent portrait lens, 7mm to 8mm for a typical 35mm FoV standard lens, a 6mm for roughly a super35 wide angle 28mm equivalent, a 4mm for a 18mm-19mm super35 super wide angle equivalent. Interestingly there is a single lens that can cover almost every one of these focal lengths, is pretty fast with a fixed focal length of up to f/1.6; however the lens doesn't have a filter ring so one would have to built them self a "matte box" equivalent solution and add some IR cut along with ND to get something usable out of the lens unless they want to crank down their aperture size. Also there is very little know about how soft this lens is wide open, but it's probably gong to need to have it's aperture dropped a bit to get a decently sharp image, but for the budget cinematographer who is learning to shoot with this budget camera it is the most ideal solution. Just as a thought on a filter solution, one might try getting a box of step up ring adapters and seeing if one can slide over the outside of the lens, even with a little black electrical tape this could get one a filtering ring solution for roughly $10-$20 and one could match an ir cut and maybe variable ND for whatever happens to be the best budet price. Anyhow, here is the best budget single zoom lens for the entire range:
4-12mm = Arducam 4-12mm f/1.6 - ~$50
You are actually really close, it’s a 5.6 crop factor. so my 16mm c lens is the equivalent of a 70mm Ef apsc
@@Andoonline 5.6 * 16 = 89.6mm on Full Frame, and is definitely on the upper end of what's considered a portrait lens. Typically they range between 75mm-100mm with 80mm-85mm often considered one of the most flattering ranges. For Super35mm (1.5 crop) equivalence a 16mm 5.6 crop lens is more like a 60mm lens on that format. That's a good lens option to have.
@@billyoung9538 lol I guess my math was a bit off
Well yeah and thats obviously where things like focal reduces help to an extent or using broadcast lenses or old super 8 and super 16 Bolex lenses and alike. Also the crop factor sort of depends for how you shoot, aspect ratio and alike. Not that the crop factor itself changes but how you implement it changes
It looks like a video that may be featured in a future documentary about the beginning of open-source cameras. Anyway, I'm here for it, great video.
Thank you very much for introducing Cine Fox, your CinePI-LEGO is also very cool!😆
First time I've heard Scanazon Crinme 😂 going to use that from now on. Awesome video, I've been wanting to make a CinePi for awhile, I'm waiting to see AltCine's version!!
I heard the reason that blackmagic RAW is allowed is because BMD proved that BRAW was made before RED got it's patent, which just makes me question how RED (now under Nikon) got the patent in the first place
@@bar7381 no the reason is that it isn't actually raw, it has been debeyered in the camera/recorder
What an engaging and confident narration. This guy should be a professional presenter.
Thank you, he’s going to be
4:31 This is funny, Sony, ARRI, Blackmagic and others all have RAW codecs and or encoder packs old high end pro kit from the 2000/2010s still has massive life in it, and Cinema DNG 12-bit support, there is now also FFV1 from FFmpeg for a lossless compressed up to 16-bit 4:4:4 which is actually a bit better then CDNG, but cost of cameras becomes a rounding error when storage costs are factored into the operations.
The moment you said Magic Lantern I immediately wanted to tell you to try crop mood on T4i/t5i or an EOS M. It's legitimately mindblowing how well it works. I can record 2.5K RAW in 14bit with real time preview
Outro goes hard ngl. Also the Lego shell is awesome lol
The song was never meant to be heard by public, but when ando put it there I was cool with it being used
@@YogertPC was it in the discord at one time? I swear I’ve heard it before
@@HowdyYT There was a slightly different vocal-less version in R9 380 cleaning video
You make doing this stuff worth it howdy ❤
@@Andoonline 🥺
This project is awesome! Building your own cinema camera with off-the-shelf parts is super cool. I had no idea you could do so much with a Raspberry Pi. And calling Amazon Prime ‘scamazon crime’ - that really cracked me up!
It’s great to see such an involved community and the Discord for discussing and improving these builds. Open-source projects like this really make advanced tech accessible to everyone.
I’ve got a few 3D printers and would be happy to print a frame for free to keep the project going. Just let me know if anyone needs help with that!
Glad you liked it, we’d love to receive a frame for the follow up pi 5 version of the video. The link to the yogertPC discord is in the description.
insane that the footage actually looks decent, really hope this becomes the future
Great video! As someone trying to build a camera from scratch, and having done a bit of research, most sensor manufacturers do not provide data sheets for using their sensors, but companies like omni-vision, on semi do. A lot of great sensors and they use standard MIPI or parallel etc., communication protocols which are well documented. The hard part (& expensive) is getting an FPGA and program it to transfer all that data to a memory (for video).
There are some great APS-C CCD sensors from retired DSLRs (for example Canon T2i) that are analog out that are very capable. With additional ADC circuitry, we should be able to acquire still images even with STM32H7 series that go up to 500 MHz!
My theory is, we actually can make an open-source “cinema camera” that isn’t just a camera module attachment that limits the video quality… some random thoughts 😂
You should check out what both Elphel and Apertus have done, not to forget the Sitina1, all open hardware designs. The main difference with the plug and play mentality is that it would be really raw-raw, handling the noise issues, etc. What most people would want in an API to control their camera, and not doing sensor plus optics theirselves.
@ thank you, will check them out!
Whoever was the MC of this video was amazing!!
So glad this project is finally getting attention!
Thank you for an amazing video. I used Magic Lantern a few years ago and was quite impressed. I do have a question though: any chance you can share some of the raw data? The MLV workflow I had back in the day was... painful. I'm considering building this CinePi camera, but would love to see some sample footage to get a feeling of the output file format and how I would leverage DaVinci Resolve to color grade it.
Thanks again for a high-quality video and you've gained a new subscriber :)
When I can get some time, hopefully not too soon from now I will definitely drop a Google drive link
I was looking into this kind of thing a few years ago. Sony has a medium format 6.5 cm x 6.5 cm sensor module that I was really keen on trying to do this with. Instead of a pi, full blown GPU accelerated encode on x86. Basically unless I ordered like 5000 camera modules, they were not interested in selling to an individual. And the re-sellers did not stock that because 1) most of these are used in scientific or manufacturing, where such a giant size is not really that important aside from astronomy. 2) trying to license software that played nice with high bit rates and whatnot was a real difficult task as, yet again, most companies buying these modules for their own uses, has a R&D team to do the software with their own FPGAs. 3) Xilinx technically made some co-processors FPGA thing that the Sony camera module worked nice with, but not the larger size sensors, again, for the aforementioned reasons. Seeing this gives me hope that maybe a DIY medium format sensor with GPIO to x86 PCIe could still happen in the DIY space, but it is still very difficult
In case anyone finds this and is curious: the sensor is Sony IMX411 BSI . 151 MP. But if you look at the data sheet, it can only do 6.3 fps in I believe monochrome 16 bit and 3.2/3.1 fps in a 16 bit colour. Ximea sells them with a kind of module interface deal. But yeah, stand alone module to do a high speed full sensor 16 bit depth would be insane
Your chassis makes me think of a software project (if it doesn't already exist): a simplified cad package that takes a lego build and converts it into a 3d printable cad file, so you can prototype the dimensions as part of the design process.
I really should have just gone to Home Depot, and figured out something there. The Lego was more to convey just how much I don’t have access to a 3d printer,and how unwilling I am to buy one… at the moment 😂
This is rad, i bet you can run it through gyroflow with some tweaks and get it stabilized if you wanted to.
Exactly 😊now you’re getting it
The Mokose C200 comes with some awesome C Mount glass.
very good explanation man, haven't quite understood the project that well until now! Love your lego build.
And I only scratched the surface and talked about what is considered the “easy stuff”. Just wait until the pi 5 images come out and the larger inch plus sized sensors become readily available, as well as some off the shelf housing projects it’s gonna be… amazing.
I miss being this excited. I had tried to do what theyve done with cinepi with other machine vision cameras and janky af software. Im glad to see where we have all come to. Cant wait to see updates.
This is a really fun and cool video! Keep it up!!
I wish Sony LTY 900 DCG sensor was available to public to purchase.
Till then I feel Android RAW video is holding strong as we can encode in ProRes HQ on fly using a powerful Snapdragon CPU encoding.
Lol, liked the LOTR reference 😂
Good video. I like your style. Dude talking at a camera is hard to pull of and in tech no less.
That is so funny I am literally in the process of ordering my CinePi hardware. A Pi 5 and a Sony IMX585 4 channel CSI-2 from Soho(Japanese Company).
Love your video!!! I built a Lego Pi Case too. They Rock!!!
Lol… “Scamazon Crime”…
Hell yeah man, power to the people.
Great insight I may have to try this!!!
Sweet, so once it's ported to RISCV then it's really free and open source because you'll be free of ARM; open hardware is real and important.
They are always looking for ways to port to other platforms. it’s just they struggle to find support for csi and mipi
Can it run crysis?
this is awesome, def gonna give it a go
So a 12.3 megapixel sensor, more than enough for video but what if I want more resolution?
Well it’s possible through building your own sensor board like a starlight eye, but also at some point, I believe an upgrade to the HQ camera module will be releasing so I would wait for that
But also join the cine pi discord for more information
Excellent video! Ngl… I kinda want one now 😂😂
Still surprised this video hasn't blown up yet
It kind of blew up for us. We haven’t done numbers like this before. Perhaps the next Cinepi video will actually blow up.
interesting, I wonder, could you run the sensor further away from a full-frame lens.. and would that result in more light per sensor element? rather than sub sampling a much larger sensor, you'd be capturing all of the light from the lens, might have really nice low light.. maybe not compared to the bleeding edge, but compared to my gear 😆
LEGO not LEGOS! I had to say that.
Great project.
I’m gonna start saying LEGOEN fight me lol 😂 and thanks btw
YOOOOOOOOOO
only have to wait till 2028 5:17
Thanks bro!
Now im looking forward to cameras released in 4 years time when the patent expires LOL
The CinePI discord appears to be down. Is there an alternative?
discord.gg/6nNEMvkmSt
This is crazy cool. Thank you for sharing
Once they get raw 4k 60 fps I’m all in!
Some people on the pi 5 do have 4k 60 working
i've been wanting to get my holga set up with a pi and digital camera. could be fun, but the quality will be dismal, not much that'd fit in the space can beat film.
Good information my man.
'Scamazon Crime' I laughed to hard
Hi, what patent protects is _commercial_ exploitation, everyone is free to use one to construct and use a device, what is forbidden is to sale it commercially speaking - this is valid in each and every country that have patents.
On the other hand, granting a patent for raw recording just means the patent office is more a psychiatric institution that a responsible administration (not to talk about software patents, which are preposterous and, as usual, a brake oriented toward small and disruptive companies).
I want to build one at some point. Camera prices are too insane for someone who just wants to experiment with it.
Nice!!!
4:16 raw is just a digital form of the raw sensor data, I don't recognize any claims a company tries to have over it.
Sure they may whine about it but at the end of the day it's your sensor data.
Personally neither do I but lawyers may see it that differently
@@Andoonline realistically they can't do much if you don't flat out copy their software implementation, the biggest issue with raw recording regulardless of format is going to be bandwidth overhead and sheer compute, but my area is more the hardware side of things.
I'm starting to like the new YogertPC
what's that acoustic background music you use in all your other videos called? I love it.
@@jihadjohn4119we just call it Jaylin song
My friend, Jalen made it many years ago. It doesnt exist by itself anywhere (yet)
*dies from banger
t2i in 2024 is legendary
I thought about upgrading him to just the t3i or eosM for the meme
8:27 except they do, there are plenty of services that can cut the parts for you.
That wouldn’t be “off the shelf”
@@Andoonline technically no but it is something some can just give a open source file to said service and just get the parts.
@@deltacx1059true but I looked a little into that route and found it rather pricy however I did forget to mention that many public libraries these days do let you do some 3-D printing so that could be a route to look into
@@Andoonline it is pretty strange in terms of pricing, some local places I checked will do it dirt cheap when it comes to cutting plate parts but most services youtubers advertised as being cheap end up being stupidly expensive, sometimes to the point where you could buy a machine and do it yourself for cheaper. (Which defeats the point of said service)
Canon hates this trick
And if I have it my way we’ll be able to just take their sensors off their cameras and get access to their color science and put it on ours… one day… maybe… lol 😂
i really wish there was a magic lantern for sony cameras.
You and I both homie
"scamazon crime" LMAO :D
What happens when you remove the lens to a camera with two raspberry pis ? You get a pipi
do who you love, work everyday of your life
I bet it will boot faster than a red xd those things are abysmally slow.
it's a fun thing to try out i think, but it being able to record raw video doesnt make it a cinema camera... first of all the sensor is too small for that methinks but most importantly you don't talk about the reliability of the camera. will it work at all times or will it randomly stop working or bug, corrupt your files, not save properly etc... cause thats the most important part in a professional tool. you dont want it to prevent you from doing what you want. also you mentionned a usb microphone but that is very much not a viable i/o for professional audio recording.
My point is it's a fun project and there's definitely value in it but calling it a "cinema camera" is definitely a stretch.
also thats just me but a good lens on a bad camera is miles better than a bad lens on a good camera.
Believe me, I understand everything you are saying about it. But please check out the discord. I could never do this project justice. This project goes way beyond anything I mentioned and anything I included , and the camera can be about as reliable as you can build it, and I really promise that there are configurations that you can 100% call a cinema camera,
But also I really do value your opinion on it anyways, thank you.
Hmm any hackers who could help hacking the hvx200 and hpx170, they both run linux haha
Ok ok hear me out… VX pi thousand with Fish eye lens🎉 cue police truck by Dead Kennedy’s
Red sold to Nikon
As a passion project it’s really awesome to be able to build your own video camera from off the shelf parts but it is impossible to compare these DIY cameras with ARRI or RED. Both companies have spent millions on R&D on sensor, colour science, compression, power consumption and more. You’re NOT going to get the same results with resolutions, skin tones, build quality and even compatibility with other industry standard equipment. It’s a great idea but as with most things, you do get what you pay for with film equipment.
I challenge anyone to build a camera that would be able to capture the “look” , bit depth and resolution of an ARRI Alexa with a rugged design, a v-lok battery mount and a PL Cine camera mount
Yet
"Cinema camera" implies cinema quality. That isn't just resolution, it's image quality, good performance in low light, good depth of field, good color reproduction, the ability to pair the sensor up with good glass. The Pi cam sensor is the same kind of stock sensor from ultra cheap mobile phones, there is nothing cinematic about it.
Moar: ua-cam.com/video/0hOvog6dWoo/v-deo.htmlsi=bTStxG0n0qT6WMrf&t=219
@@thehydrobladethe cine pi community to my rescue. I love you guys
You can get nice picture out of it. But so can you with most modern sensors. But I kind of agree that it implies something else. Even if they made an Imax movie with GoPro 11 most people would not call that a cinema camera.
Still it is a cool project I just wish it could handle more interesting sensor. What would be nice is an open source external recorder. Or a 3chip camera
I mean there’s the imx 585, or the starlight one inch. movies have been shot on 16 and super 16mm film and those look great and even look great upscaled to 4k from a 1080p remaster as there is plenty of detail to work off of. “Cinema camera” I don’t think is necessarily an image quality standard, as I think it has more to do with flexibility. However quality does play into why people use a camera for shooting a movie at all. But anyways the cinepi community is always adding more feature’s and more support for more hardware. so i don’t see why we won’t see a super 35 sensor on a cine pi at some point, and I think someone has even gotten a sensor close to that working. This video does not do the community justice as things change so rapidly and its goals are sky high.
this shit so goddamn rad
I was very frustrated with the file mess of magic lantern on my eos M, its what drove me to the Black magic ecosystem having ef lenses for my Canon t6i, Missed a magic lantern rebel camera by this much 🤌
As someone who uses external file sorting software, so I don’t have to deal with exactly that. I understand your pain. Especially since my 7D’s CMOS is dead, sorting by date at the moment is kinda a pain . 😂