Tzedekh

It's not just sensor size but also construction. There are smaller sensors with greater dynamic range, such as the BAE CIS2051 used in the BMCC. With any luck, some sensor manufacturer wil come up with a light-sensitive, HDR, 1-inch 4K sensor. With a removable lens and a SpeedBooster-type focal reducer, you can have shallow DOF if you need it, and deeper DOF without it. Sounds like a good solution, but probably not a challenge any sensor or camera manufacturer will take up soon.

Any mention what form 4K will take on the X70? Will it be UHD, or considering that the X70 is supposed to be the "pro" version of the AX100, could true DCI 4K be possible? Also will it be 8-bit 4:2:0 UHD XAVC-S used on the AX100, or the 4:2:2 10-bit long-GOP 4K XAVC Sony announced nearly a year ago?

Or they could implement a global shutter and move the global-shutter circuitry to another layer, like Alternative Vision Corporation. Isn't Sony a leader in stacked-sensor design?

Unnecessary. With a stacked sensor design, the global-shutter circuitry could be on a separate layer. Read comments by Alternative Vision's David Gilblom on this Image Sensors World post about the Epic Dragon sensor. Apparently, with global shutters, you can have your cake and eat it, too.

In the future? It's already been used for at least 60 years. Watch old Warner Brothers cartoons. Frequently, when a character swings a club or bat, it curves.

AVC-Intra Class 4:4:4 supports up to 12 bits.

Maybe I'm wrong, but it seems that global-shutter mode has been removed.

The older sCMOS chip essentially had these same specs -- including dual 11-bit ADCs. The dual readouts are zero-gain and high-gain, and the two are combined to form one 16-bit signal. The Arri Alexa's sensor has a conceptually similar dual-gain architecture.

I suppose Olympus didn't think it necessary. How long did Sony drag its feet before including 24p in a consumer or prosumer video camera?

The CMOSIS CMV12000 sensor alone probably costs BMD at least $1,000 per unit. I like its global shutter. I'm not nearly as sanguine about its dynamic range -- I suppose we'll have to wait and see. I expect to see improvements in sensor technology involving stacking (i.e., the global-shutter circuits are moved to a different layer, thus allowing larger photodiodes) and dual (normal and high-gain) readout like that used in the BMCC's Fairchild/BAE CIS2051 or the Arri Alexa's sensor.

You said that the Bolex was not a Super 16 camera. The significance of Bolex doing the conversion (i.e., it wasn't simple a third-party, aftermarket service) is that, in effect, there was a Super 16 Bolex offered by Bolex.

Except that Bolex, not a third party, did the conversion and pretty loudly promoted it once Super 16 became somewhat popular. You could say it was the very, very late '90s -- in fact, 2009.

A couple of corrections: The Bolex H16 was convertible to Super 16, and round-trip conversion between regular 16 mm and Super 16 could be done in the field. Super 16 has an aspect ratio of roughly 1.66:1 and blow-ups have usually been to spherical widescreen 35 mm. Vantage's Hawk V-Lite 16 1.3:1 anamorphic lenses allowed a 2.40:1 image to be captured on Super 16.

Not so. Indiecam's indieGS2K records CinemaDNG raw over 3G-SDI to the Hyperdeck Shuttle. Whether raw could be sent over HDMI is another matter. And it's probably less likely that 5K or even 4K could be sent over HDMI. I never said it had to be CinemaDNG, though. If the Bionz X processor can encode H.264, it certainly has enough power to encode 4K CineForm RAW, which requires something like 50-60 MB/sec at 24 fps, about half of what 2K CinemaDNG requires.

What would be cool is if someone hacked the Bionz X image processor to get 5K raw out of it.

I think it was the "sheep like" comment, which seemed derogatory, whereas Kays Alatrakchi was making the point that using industry-standard cameras isn't passively conformist but rather professionally prudent.

In many ways it makes life a little easier. It's much like as it was with 35-mm film (and still is, to some degree), when most productions were shot on either a Panaflex or an Arri 35BL as the "A" camera. (Is anyone still using the 35BL?)

BMD and Kinefinity can't compete with Sony, Red, Arri, and Canon at the high end, but they can carve out niches for themselves at the lower end. Unfortunately, they have only a relatively small window of opportunity, as 4K will fairly soon invade the consumer market. I can envision an NEX-VGxx form-factor 4K camera from Sony within 2 years that will produce stunning images, particularly if it includes an OLPF and NDs. HVEC will be practical by then, and even if Sony hobbles it with 8-bit 4:2:0 encoding (which I hope they don't), someone may hack 10- or even 12-bit 4:2:2 out of it, or maybe even raw.

It could also give us 12-bit 4:4:4 8K (8,192 x 4,320) at up to 120 fps.

It would be interesting if someone could reprogram (yeah, hack) the Bionz X image processor to take the 5.4K raw and convert it to 4K CineForm RAW and then output it to the SD card. I would think that the processing power is there, as AVCHD encoding is probably considerably more processor intensive. The system bandwidth, particularly to the card, might be too slow, though.

  That's the problem of citing full-frame 35-mm equivalents -- people start thinking the lens actually is 24-200 mm. Its actual focal length is 8.8-73.3 mm. The low end is like 16 mm or 17 mm on APS-C. While not a fish-eye, that's pretty wide angle.

I'm a little underwhelmed by this sensor. The Aptina sensors have better dynamic range.

  With CineForm RAW, the data rates are comparable to those of ProRes, yet the benefits of raw remain.   My concerns are unknown sensor and unknown build quality.

  This is a trenchant observation. One can ask, "Who else is doing this at this price point?" The answer is, "No one. Even BMD is barely doing it." A company can promise all it wants, but unless it can deliver -- in reasonable quantities (buyers don't have to wait weeks or months) -- it doesn't matter all that much. Has BMD even filled all the backorders yet?

  Remember, the GoPro Hero 3 Black, which uses the Ambarella A7 camera sysyem-on-a-chip, can do 4K (well, UltraHD 4K) at 15 fps. The next-gen chip, the A9, supports 4Kat 30p. No reason that the Hero 4 (or whatever it's called) will cost much more than $400. If you want an interchangeable-lens mount and more I/O ports, well, they would add to the price, but all should still cost less than $1,000.