Video Input without blackmagic
@brunomartelli the camera i posted outputs SMPTE 1.5ghz standards over SDI, so 1080i 50/60, 1080p 25/30 720p 50/60.
It seems like you are using a tower PC with pcie slots, you can get a decklink pcie model like the mini recorder. But this has limitations you might regret in the future. The Blackmagic site has precise capture specs for all models and you can decide to pay a little more for 3, 6 or 12g capability for future proofing (as long as you have the appropriate pcie slot free), there is even a drop down filter to filter by the bandwidth in the site. 1.5ghz is generally SMPTE 259M and 292M compliant. Thunderbolt and USB models also have quite a lot of choice and price range.
Thanks! this is very helpful!
I can report BlackMagic Decklink Mini Recorder with SW 10.11 works with its own software and Isadora and SDI CCTV cam, but video source not recognised by other apps e.g VLC. Blackmagic Technical support says they stopped supporting DirectShow.
So anyone want to buy a BM Decklink Mini recorder?
Or know of any other way to connect SDI CCTV cam to a PC? (i have a custom machine learning app that is wanting the CCTV cam feed)
How has the ML app been made?
@brunomartelli ffmpeg can be compiled with the Blackmagic SDK, openframeworks, processing, cinder and vvvv can all utilise the SDK for capture, there is also blackspout that will output spout from Blackmagic capture. There is even a python module for capturing through the SDK. There is usable capture support from so many different approaches, it is pretty trivial to add it to your custom software.
ffmpeg can be compiled with the Blackmagic SDK, openframeworks, processing, cinder and vvvv can all utilise the SDK for capture, there is also blackspout that will output spout from Blackmagic capture. There is even a python module for capturing through the SDK. There is usable capture support from so many different approaches
I feel the need to point out that any commercial software utilizing ffmpeg is doing so illegally.
"Common Question: Why won't Isadora 3 support FFmpeg?
The following quote FFmpeg's site should give anyone pause:
Q: Does FFmpeg use patented algorithms?
A: We do not know, we are not lawyers so we are not qualified to answer this. Also we have never read patents to implement any part of FFmpeg, so even if we were qualified we could not answer it as we do not know what is patented. Furthermore the sheer number of software patents makes it impossible to read them all so no one (lawyer or not) could answer such a question with a definite no, those who do lie.
That's a risk I'm not willing to take yet, even though other companies seem to be doing so (Resolume, MAX, etc.) All it would take is one lawsuit to kill Isadora completely.
I will revisit FFmpeg as a possibility after Isadora 3's release. Perhaps if the end user installs the codecs, we'd be safe. That might be one route
I hope that clarifies things.
Best Wishes, Mark"
@dusx Hi the ML app is basically Python script running some OpenCV and Tensorflow
You will have to make some changes to the code to use this library, but it should not be too hard.
This thread is not really related to Isadora, it seems you were after advice for capture hardware for a python based machine learning tool, but asked in an Isadora forum and followed advice given for Isadora, not python. In the future it would be good to ask your python questions in a python forum where people have the experience to help you.
@woland as OP was asking about ffmpeg for a single bespoke tool this legal warning is not relevant. I am also still wondering how so many large commercial softwares use ffmpeg without worry.
Thanks for this link Fred,! (I'm not really a python guy, so I'm passing it up the food chain to Tom).
(The reason the thread is still here, not in Python thread, is the project began as Isadora to getting CCTV feed to Isadora then morphed into Isadora smoothing OSC values from separate app etc etc, it seemed more useful to feedback info on the Blackmagic.) I really appreciate the help from this forum.