kershner / screenBloom

Fake Ambilight for Philips Hue via Python

Home Page:http://www.screenbloom.com

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

[Feature] Turn light off below threshold

MattTS01 opened this issue · comments

I'm a big fan of the project and love the atmosphere that can be created when the lighting matches up well with what's in the film. Fire, neon lighting, sunsets etc. There are some scenes where it seems even minimum brightness of 1 isn't quite dark enough to create the right effect.

Scenes showing the darkness of space or where the lights suddenly go out feel like they could be greatly enhanced.

Is this feasible to achieve or does it add further lag when turning the lighting back on once the scene is brighter? Is there any risk of damage from rapidly turning on/off the bulb?

Hey there!

Thanks for the feedback. I get this question a lot, and for obvious reasons. It'd be badass if the lights could completely cut off during a particularly dark scene or something. I've tried various implementations of this idea throughout the years with varying degrees of success, but none of them are really viable.

Basically, there is a bit of start up latency when the lights are coming back on. It's really not noticeable for one-off events like coming home and turning your lights on, but when you're flitting back and forth between "on" and "off" states potentially dozens of times per minute, it becomes untenable fast. Not only that, but the "on/off" would need to be sent for every light being addressed, which gums up the general command pipeline and could cause more backups/delays.

I'm not actually sure if turning them on/off rapidly would risk any damage to the lights. I'm guessing no, but would need to do more research.

It's also worth mentioning that I haven't done any more experimentation with this idea since I moved up to the V2.0 bridge, which is much faster and efficient than the older one. It could potentially be more viable, I'll try and investigate that at some point.

Thanks for the fast reply! It sounds like what I'd expected. It isn't really something that the technology is designed for so would require Philips to have a separate 0 brightness state separate from off. I don't suppose the new Entertainment update changes things at all?

I'm tempted to give it a go myself but I can't imagine I'll come up with any radically new ideas. I've thought about trying to apply a few ideas from computer vision to the image processing side of it but that could add a lot of extra complexity without necessarily gaining too much. I can easily see that making something work better for some rare situations will totally break it for others.