Version 3.0 for release June 20, 2009
Premise:
Unless Jobs is an idiot—and he's not—this will happen.
Exactly as described here. Just not any time soon.
Bugger! At the WWDC 08, the Video iPhone was NOT introducded, after all. Geez.
iPhone 3.0 will be the first hand-held videophone. Two cameras make it possible. One points away from the view screen, the other points at the caller as they gaze at the view screen. Tricky image management gives other iPhone users both views.
Cameras these days are tiny. The 2MP camera in the original iPhone is quite small, but the 1.3 MP camera in the MacBook Air is far smaller.
iPhone 3 is about identical in size and heft as the Gen 2. The iP3 has two cameras. One faces the user, to capture the face of the caller, and the other one points straight out from the iPhone's back to catch whatever is there. Each camera has its own auto exposure and white balance system so the caller can be in a dark room looking out to a daylight vista and both images will appear in balance.
Face Size Games.
The field of view for the caller's camera (right of, and in line with the earphone aperture) is wider than the image captured.
Face recognition technology is used to scale and crop the image to just larger than the caller's head, showing this image as a 64 x 80 pixel reference image in the upper right corner of the screen while the forward camera sees a modestly wide-angle view scaled to just under monitor width.
OK, how about lower left corner...
The face recognition system will recognize up to four faces at once, scaling the "MyPic" and expanding up to 160 pixels wide if a horizontal line of faces is accounted for. As people step into frame, the MyPic window smoothly scales to include them.
Received images flip the relationship. The distant connection's face is flush left and slightly lower, initially.
Image stability.
Setup preferences give you several MyPic / MyView / YourPic / YourView options and exact formats are not yet finalized, but pre-production samples keep the MyPic image smoothly centered, even as hand motion is irregular.
This is a form of image stabilization found on many camcorders in which the pixel map of the final image is indexed from different starting places on the image chip's larger map of total pixels.
As the camera moves or wiggles, motion detectors tell the image system to pick each cropped frame from a different position on the chip, offset by the amount of instability, giving the illusion of a rock-steady video phone camera.
At the same time, the MyView camera is similarly stabilized. Transmitted MyView images are either 120 x 160 or 240 x 320 depending on frame rate and preference choices. The stabilization may be switched off in preferences, but why would you?
Frame rates.
Transmitted frame rates are variable. A good connection with simple subject matter and low contrast lighting will raise frame rates to 12 fps under ideal conditions. With a "Two Line" option, this can be achieved more consistently, but the service will likely add about $20 to each phone's monthly fee.
Frame rate can drop as low as 1 fps under difficult circumstances. Priority can be assigned to MyPic or MyView, but that's it. Nothing you can do from the iPhone can assure live frame rates. Still, in practice, even frame rates as slow as 3 fps feel more like a better connection than just voice alone.
Each local camera can be switched off with touches to the screen.
Design points.
In typical Apple fashion, the layout and control of the potential four sources involved in a video phone conversation is intuitive. MyPic and MyView images overlap a bit and are tied together. Touching either allows it to show to the receiving phone and activity is confirmed with a small green + in the corner. Touching either dims the view and turns the + into a red slashed
.
The received image(s) are under the outgoing set when the iPhone is held vertically. It can be moved to the top position with a finger flick, but the only control over this image set is the Record Mode.
Read my clips.
Record mode may be activated at any time with icons. Record from MyView will be at the full 12 fps and can capture both cameras at once. Apparently the recorded image is a composit of the screen layout, not individual camera feeds. Record mode that catches received images always captures whatever was sent.
For security, date and time of every frame is buried in the metadata of recorded files, perhaps promoting the iPhone 3.0 to become an evidence-gathering device in the hands of some users.
Offline video recordings can now be made with a MyView motion recording format incorporating MyPic or not. Up to four files prepared this way can be arranged along the bottom of the screen as tiny icons, then dropped into the MyView image area during a call, transmitting immediately to the recipient. Apple currently calls these DropIns.
A special MyView Xmit mode allows pristine MyView files to be sent to the recipient as instant messages. This takes longer than live DropIns, but files received this way run at the full frame rate when viewed at the destination.
Deaf users will now be able to transmit in a iQuiet mode for improved frame rate at a resolution making lip-reading and/or American Sign Language a functional reality.
Can speech to text be far behind?
Too good.
iPhoneChat will doubtless become the most prevalent video phone mode. When no MyView or YourView images are involved, both MyPic images move to the top of the screen, offset and slightly prioritized. The other person's face is larger than yours and faintly reflected in an infinite black glossy environment, similar in principle to an iChat session on a laptop. If three-way video phone capability is in the cards, it was not demonstrated on the current test articles.
If you or the person you are calling switches off either MyView shot, the screen images slide smoothly into an accommodating format. The experience feels just like you are on a newscast with a clever director manipulating images of both host (you) and a distant correspondent (them).
Today's prototype seems quite polished with relatively few hiccups. Occasional image dropouts and slow responses are fairly rare. Speculation is that MacWorld this summer may be the first time it is shown publicly.
Way too good.
The impression this gizmo makes is strong, to put it mildly. Why didn't anybody think up anything like this in the past? They had me at iPhoneChat. But the feeling that you are part of a telecast as you use it, the opportunity for the caller and callee to show you their world as they maintain eye contact with you, the ability to record all the sound and video data any time you wish—it's mind blowing. Or mind-melding, if you are a trekster.
Everybody will HAVE to have one. (Just like now.)
Accessories will abound. Remote BlueTeeth cameras, security systems, cross connection to Apple portables (iPhoneChat blows both ways) and sheer world-wide telecommunication options will make this the single most desirable communications tool ever. Ever.
We expect Apple stock to plummet precipitously in the hours after its announcement. As usual.
All concepts on this site subject to the crushing presence of realities
beyond our control. ©2008, Peter iNova, all rights reserved.Steve Jobs: Don't you dare steal these ideas without big
payments and credit where it is due. 2/25/08 is when the basic operational ideas expressed on this page went public. Either you, or some of your guys, read it here first.
