Hacker News new | past | comments | ask | show | jobs | submit login

You can https://www.e-consystems.com/13mp-autofocus-usb-camera.asp

I bought one of these for using with my lab microscope since the alternatives from thorlabs was 10x more expensive for no reason. It worked fine for a year then it stopped working. Maybe I spillled something on it? Who knows ¯\_(ツ)_/¯




thorlabs cameras, and scientific cameras in general, have very different response functions. useful if you want to make quantitative measurements, accurately represent a B&W scene, etc. I use a BlackFly (FLIR) camera on my scope, but have been looking to do something similar to you, since the camera API is not V4L2, but rather a custom framegrabber SDK.


I just wanted something to capture transmitted light images on a real-longterm setup. Turns out the best method is to wheel a simple tissue culture microscope to a warm room and put a cell phone on the eyepiece :) made some cool movies for sure!


Currently I'm playing with this: https://github.com/bionanoimaging/UC2-GIT which does a lot of that (mainly for educational purposes- I think most scientists should focus on getting grant money to buy professional equipment).


That's a candidate for sure, my efforts slightly predate the 3d printed microscope revolution slightly, and I had to MacGyver my way to prove I can image cell cultures long-term without spending 2-300k (since I was doing wild goose chase hypotheses) and hence this solution.

Are you playing with it in a garage or in an academic / industrial setting?

I'm not yet ready to start but my goal is to try and set up an actually productive garage lab - maybe use some neighbouring univ core facility on occasion but not to set up a lab there. Not a fan of academia in general and hope to not contribute money/ideology towards perpetuating that ponzi scheme!


Garage. The goal is to build a prototype which could then be scaled to a warehouse-sized robotic biological experimental system. But, I have also worked with such things in commercial settings.

I'd say that for most real scientists, it makes more sense to raise the funding to buy a professional scope because a lot of the dumbness is engineered out so scientists can just sit down and be productive.


Yeah I have one of these, the picture is nice with good lighting (I have a black and white one, with an antique 16mm film camera lens). Going to write a OBS plugin for it when I have time.


I'd be curious about this. I think what would make sense is to write a V4L2 driver for it, then any app that uses V4L2 could use it (on Linux). Or, is it far easier to write an OBS plugin than a driver?


Yeah thats another option. I haven't found one, but I do have userspace code that works.


Thorlabs don't make their own cameras by the way, they're white-label IDS Imaging products (at least they used to be).

Most of these [machine vision] cameras will communicate using genicam, so you can use libraries like Aravis and maybe there's a way of getting that to talk to V4L.


$250 is deep into mirrorless camera territory though.


If you by new $250 isn't close to getting you anything decent which also works for streaming as far as I know.

Plus for many cameras you need to capture their video output, i.e. USB is not enough (through that currently changes). So you need a capture chip with low latency or else things are bad for "life" usage. This is often another 60-130€ for something decent if brought anew.

Plus they don't have a monitor clip so you need buy an additional stand.

Sure you can buy used parts but then you also need to compare it with used part prices for webcams which normally (i.e. non Covid19) are also lower.

I mean a Sony a5100 with usable objective currently sells new for ~US$700 but you can probably get it new for that price with capture card and stand if you buy clever and we ignore Covid19 for fairness.

(And even used you are still far above 250€ for any recent decently usable mirror less camera with objective in my experience.)




Or GoPro, which you don't need a USB interface for. Works all the way back to hero 4 apparently.

https://gopro.com/en/us/news/how-to-use-gopro-for-webcam


A GoPro performs horribly in bad light though. And it costs serveral hundred dollars as well.

If you already own one, then it might be an option. But buying a GoPro to get a better webcam doesn't really make sense.


Neither of those things is true. You might have to mess with it a little bit in truly dark rooms, but if you can read comfortably you'll be just fine. And far better than any webcam.




Join us for AI Startup School this June 16-17 in San Francisco!

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: