Hi James, thanks for the reply on product obsolesces. We actually already over run by a year and we are looking to complete in 2026 because we are a small business and medical devices have too many regulations. It's regulatory gestapo.For example, the Pi 5 does not support MMAL at all (or earlier versions of the OS), so there is no upgrade path to the latest models.
We have one engineer who knows about MMAL, and whilst he has been helping you out and may continue to do so, if he choses not to because he has a lot of other work to do then, as stated, support will be very thin on the ground. Whereas we have many more engineers working on libcamera, and as it is a Linux standard, there are loads more people outside of Pi who might be able to help.
You say your project is running until 2029? You could easily redo the MMAL camera parts in libcamera in a week or two.
To you, an upgrade is easy because it is take Zero 2W, zap a new 64 bit OS, enable certain features on libcamera and voila! I agree with you that in that aspect, upgrade is easy.
But because Pi products in particular Zero 2W and Zero 2 have great form factor, stable OS (at least after some iterations of Buster), nice EMC, we built our electronics and software around it. We developed internals such as IOs, MMAL H264 streaming over to PC (call this the device software) with a PC software and we zap live images of patients to clinicians. It is this device software we will be having issues with because this rewrite is not a simple rewrite.
OK, I admit I do not 100% know the ins and outs, not a linux guru and I am not a programming god. I know bits and bops here and there, using that MMAL to another layer of code that puts all these into place and sends live video through the network to the connected PC. I think we have done well so far doing all these despite not being an encyclopedia.
I do not know that libcamera calls and OK I don't mind learning it then I have to learn and figure out how to rewire all these to that layer of code that puts everything in place. Before I know, time has already flown.
We appreciate 6x9's help and time, and sure, I am not hoping a 1 to 1 tuition. We are just grateful with 6x9, jamesh and pi staff for any help we can get in decanting this lens shading knowledge. We just need to figure out this lens problem because we have to use this custom lens and Pi platform on an actual real-world product and not something on a desk, run once and reboot. Once this product is done, there will always be time for improvements (in this case how to figure out ugprades) but that is not the battle today.
Thank you!
Statistics: Posted by jmnemonic — Thu Dec 11, 2025 11:28 am