(This isn't wholly relevant to the Lytmi Fantasy device, but since we rarely talk backlights here, I figured this chunk might still be informative... even though I'm now realizing I posted most of this same ramble a year ago earlier in this thread...)
What background light needs/needed to really be something is a standardized, authored stream from the source, mastered either by the studio or an AI application for the optimal impact.
So, currently solutions like Ambilight provide their lighting effects by analyzing the video stream as it is provided in realtime. It has to guess what would look best, based on what it sees on screen at any one time. Some scenes are easy; if you see the outdoors and a lot of grass, the lights will be mostly green, if it's in a submarine then probably you'll get a lot of blue lights, if a gun is firing hopefully the lights flash quick brights. There's some thinkiness to its algorithm, as it has to average the color and brightness across the whole scene (otherwise you would just get color extended from the very edges, and most times your Amblight would just be showing shades of blue sky,) and it also has to make some assumptions of averaging across time (otherwise a Michael Bay movie would be a freakout shitshow of random lights dashing and flashing for every fast-cut action scene.)
This promo image for Ambilight is a perfect example of what you want from the system, extending the visually exciting elements out into your whole home...
...it is also
probably bullshit, since the colors it is picking out are so specific and perfect in what they chose to accent. (Especially the reds in the top corners coming out across the corner span from just those two little red streaks, or the lower orange/blue corners being so dominant when the brightness is hardly there in the picture. Also overall resolution and dark spots don't make a lot of sense of how many LEDs your kit would include back there.) If Ambilight can/does do a backlight effect this well, great! But this is
probably done by hand, which is why I'm saying that for it to work this well, it would probably need to be done beforehand.
If there was a standard "Backlight Stream" in a video signal, most products would just go for a generic solution, not too different from what your TV does now by reading the image from the video stream, and that would be fine. However, the few products that were cared for could be awesome, customized multisensory experiences. Also, even those mastered generically would still often be better since the AI wouldn't need to be guessing at light values from images in realtime; it could analyze several seconds of video to choose its lighting effects and bake those down for every scene. Even live feeds could be better, because for example a sports game could stay mostly green for the field, even though its close-ups have many more clashing, busy colors in them.
Having it mastered beforehand (or in games, having it controllable by the source device in game code,) would also allow for custom solutions. So a game could flash when you get shot even if the image doesn't have the flash on-screen. Or there could be just a gray shimmering light in say Silent Hill to accentuate when you are out in the fog.
Razer actually has some of these customized backlight features built into its
Razer Chroma system. Razer also has some extra cool stuff for people with Nanoleaf panels or LED keyboard/mouse devices where they light up with special patterns or colors according to in-game triggers. However, it's proprietary, so few games support it and it doesn't do anything with your PS/Xbox games or your Netflix movies or anything on your TV.
(BTW, since I'm bringing up the loose concept of a backlight standard, there could also have been/be a standardized channel for haptics as well. 4D up your apartment!)
)