Hacker News Reader

Comment by watermelon0 | Hacker News Reader

Adding to that, another big reason is the community. Raspberry Pi has the biggest community, and majority of people interested in tech have probably at least heard of it, if not even played with it.

There are numerous blogs posts and tutorials available online, and there are many modules and additional boards made specifically for various Raspberry Pi boards.

View on HN

Then again though, there are lots of things that RPi doesn't do, and I wish some China-based OEM would do. I'm not opposed to people coming up with variants, it's just that the current crop of variants don't fix real problems.

Like fix the stupid I2C clock stretching bug. Reverse engineer the Broadcom chip, clone it, and fix the bug, because the lazy asses at Broadcom won't do it.

Or make an RPi with a rich plug-and-play CSI camera ecosystem with hundreds of sensors and lens mounts to choose from, all connected via the camera connector.

Seriously, screw the RPi "high quality" camera and it's teeny useless sensor -- I want 35mm full frame, 20-50 megapixels, in CSI with a python library for exposure control, and I'm willing to pay maybe half the cost of a full frame DSLR to get that.

Nobody sells such a thing at the moment, and it would be an awesome thing to have.

You may underestimate how expensive imaging sensors are in low quantities, a sensor available in single quantities at mouser is the cmosis cmv1200 apsc sensor for $1600-1800 a piece, full frame cmv5000 is $5000+ afair. A whole ecosystem probably wouldn't get the quantities.

If they want to play that nasty game, maybe the suppliers could buy stock cameras for $1000, desolder them, and reassemble them into hacker-friendly boards for $1200, and then resell the rest of the stock cameras for spare parts (e.g. mechanical shutter assembly) at 1/2 the cost of OEM service. Even if they can scoop up the huge graveyard of e.g. Canon 60Ds that photographers will turn over for $200 a piece, or 6Ds for $500 a piece, reverse engineer them, and recycle them into a Raspberry Pi module, that would be amazing.

Even with the CMV12000 though, someone should make a RPi module that "just works", or an alternative RPi that allows it to "just work", for those of us who want to spend our time working on the actual imaging and software ... it's kind of sad there's this huge graveyard of good sensors out there that just can't be used by imaging experts because electronics people gatekeep the industry and only release them at reasonable prices inside consumer devices, which isn't what those imaging experts want at least in terms of programmability, communication, and mounting options.

Yea, I would love a high quality sensor that's "programmer accessible". Too bad that nobody besides raspberry pi foundation even attempts that.

Try looking at astrophotography cameras. Many of them have ASCOM drivers which allow pretty much full control of everything.

https://ascom-standards.org/Downloads/CameraDrivers.htm

Thanks, I just did!

Looks like a ZWO ASI294 MC could be interesting. It's "only" Micro 4/3 and $700 in the cheapest config but having USB-3 and drivers may be what I was looking for.

Full-frame astro cameras are unfortunately around $5k+ (albeit with a resolution similar to a $3.5k A7RIV).

If you're willing to pay several times the cost of a full frame DSLR and have an order quantity of 100 to 1000 or so, the scientific camera companies (for microscopes, telescopes, etc) are used to that sort of quantity and quality and would probably spin you a custom camera.

It's been a while since I've been in the industry but I can try to hook you up if you're serious. Email in profile.

Unfortunately that's too high of a cost; this would be mostly for serious hobby / side project use and the stock cameras are extremely limited in programmability.

Reply on news.ycombinator.com