It is possible that technology will allow for inexpensive chip fabs, like from a 3D printer. The development is certainly a hurdle.

@endotoxic They also have a lot of experience in the acquisition side of things; the Hyperdeck line seems to have been a significant source of investment for BM over time. We'll have to agree to disagree on the evolution of the lower-power general purpose CPU and its viability for image processing in the next few years.

@tomson8877 I don't think I've talked ill of any brand?

@DrDave To elaborate on what VK said, 3D printers tend to be pretty slow still. Also, if consumer printers are any indication, they require frequent maintenance and troubleshooting. For at least the next few years, they'll continue to be relegated to R&D divisions. I've heard some companies are working on 3D printers capable of creating semiconductors (which is pretty neat!), but I can't imagine using it at scale.

How many transistors you can buy for dollar:

Good illustration of the things described here.

Thanks @agoltz for the link.

Via:http://www.economist.com/news/21589080-golden-rule-microchips-appears-be-coming-end-no-moore

Moore's Law is really more of an economic imperative.

couldn't BMD do the same? Perhaps with additional chips for any encoding duties, if necessary. Or is that not possible with the Milbeaut? Same goes to other brands, could, say Pentax or Nikon use an LSI for image acquisition and debayering, and a run off the mill ARM CPU for the interface and encoding duties? Something from say MediaTek or Realtek. They have got a ton of horse power. Heat could be an issue, but if the CPUs aren't stressed to much...

I hope you are not serious. Using of modern image LSI is possible mostly only via special agreement. Their usage also require vast knowledge and huge investments. In fact, it is why firms so hold to the same image LSI range.

Btw, no one told that Panasonic will use Fujitsu chips, they just made some join effort, we do not know that kind of stuff they are doing exactly.

Intel's SoFIA SoCs designed for entry-level smartphones and tablets will be built using TSMC's 28nm HKMG process technology, said the sources. As for the other series of Atom SoCs codenamed Broxton, Intel will use its 14nm FinFET process to make the chips targeting high-end mobile devices.

Surprise, surprise. 14nm is first process were price is significantly higher and performance increase is very small.

I'm not sure what is up with the doom and gloom in this thread.

This topic is not about doom and gloom as you like to say. It is topic about facts.

I'll be posting much more reasons soon as I'll have slightly more time, stay tuned :-)

“FPGA is what you use when the volume is not large, FPGA is what you use when you’re not certain about the functionality you want to put into something,” he said on the call with analysts. “You can build an ASIC, a custom chip, that obviously can deliver more performance–not 20% more performance, but 10 times better performance and better energy efficiency.”

Head of NVidia.

P.S. FPGA are the core of both BM and Red cameras.

Fpga will never die. But its not energy efficient

You replied to your own though :-) As no one here said this.

Cost of advanced mobile LSI with multiple CPU cores, quite similar to camera chips.

Note that all this years BM failed to move from FPGAs.

@caveport

It'll die slowly and painfully, so I reserved numbers :-)

this thread still going strong. And BM still using FPGA