4 ms·

The official Apple HDMI adapter does the same thing with an SoC in there. The difference is native iOS support instead of a 3rd party app needed to support it.

MBCook 3y ago
I still find it hilarious that that’s how the old cable worked, the iPhone encoded an H.264 video stream and sent it to the dongle, which decoded it and sent it down HDMI. Now that iPhones have USB-C they no longer need a custom adapter. A standard USB-C to HDMI cable is supposed to work. I believe.
- bitwize 3y ago
  Thing is, I thought Lightning had that too? HDMI or DP over the wire? Guess I was wrong...
  - MBCook 3y ago
    Nope. I don’t know if it put itself in some kind of analog mode for composite. I can’t remember if that was a thing around the switch to lightning. But the cool hack has always been how the HDMI adapter worked.
- somat 3y ago
  Probably to work around a usb(lightning is usb2 based right?) link not fast enough to keep up with the phone graphics. However, I note that there are usb graphics card descriptors[1] and I assume usb graphics cards, that is, graphics over the normal usb data pins, not a display port pass through. These descriptors are what I would naively assume a usb to hdmi "adaptor" to be, a usb graphics card. I am finding it hard to hunt down low level information on how these "adaptors" work. Does anyone know what type descriptors they use? and what a iphone does if you plug one in(I am assuming the lightning to usb physical connector is trivial) https://www.usb.org/defined-class-codes#anchor_BaseClass13h https://www.usb.org/defined-class-codes#anchor_BaseClass13h update: I found these reverse engineering documents on synaptics displaylink chips. They appear to be a popular manufacturer of such dongles. And it looks like compression is needed there as well. https://github.com/floe/tubecable/tree/master/doc https://github.com/floe/tubecable/tree/master/doc
  - MBCook 3y ago
    You’re right, Apple never moved lightning past USB-2 speeds. Still seems weird to do things the way they did, unless it was just to reuse some part they already had in another device, thus saving costs. The need for compression is a good point. I hadn’t thought of that. But you’re right other existing parts should have worked if chosen, right? Weird. Just such a fun day on Twitter when it was discovered that what we had all assumed was just a relatively simple adapter was a whole SoC running its own firmware doing this job.

MBCook 3y ago

I still find it hilarious that that’s how the old cable worked, the iPhone encoded an H.264 video stream and sent it to the dongle, which decoded it and sent it down HDMI. Now that iPhones have USB-C they no longer need a custom adapter. A standard USB-C to HDMI cable is supposed to work. I believe.

bitwize 3y ago
Thing is, I thought Lightning had that too? HDMI or DP over the wire? Guess I was wrong...
- MBCook 3y ago
  Nope. I don’t know if it put itself in some kind of analog mode for composite. I can’t remember if that was a thing around the switch to lightning. But the cool hack has always been how the HDMI adapter worked.
somat 3y ago
Probably to work around a usb(lightning is usb2 based right?) link not fast enough to keep up with the phone graphics. However, I note that there are usb graphics card descriptors[1] and I assume usb graphics cards, that is, graphics over the normal usb data pins, not a display port pass through. These descriptors are what I would naively assume a usb to hdmi "adaptor" to be, a usb graphics card. I am finding it hard to hunt down low level information on how these "adaptors" work. Does anyone know what type descriptors they use? and what a iphone does if you plug one in(I am assuming the lightning to usb physical connector is trivial) https://www.usb.org/defined-class-codes#anchor_BaseClass13h https://www.usb.org/defined-class-codes#anchor_BaseClass13h update: I found these reverse engineering documents on synaptics displaylink chips. They appear to be a popular manufacturer of such dongles. And it looks like compression is needed there as well. https://github.com/floe/tubecable/tree/master/doc https://github.com/floe/tubecable/tree/master/doc
- MBCook 3y ago
  You’re right, Apple never moved lightning past USB-2 speeds. Still seems weird to do things the way they did, unless it was just to reuse some part they already had in another device, thus saving costs. The need for compression is a good point. I hadn’t thought of that. But you’re right other existing parts should have worked if chosen, right? Weird. Just such a fun day on Twitter when it was discovered that what we had all assumed was just a relatively simple adapter was a whole SoC running its own firmware doing this job.