DCI:
Absolutely. Getting into the technology a bit, can you indicate how you see the new Bionz image processor being different than the Real Image Processors, and how is it a better fit for an SLR?
Mark Weir:
If what youre saying is the real imaging processor specifically, I would say that probably the biggest difference is operational speed and also what its doing. I mean if you think about it, the main CPU of a camera has similar responsibilities from model to model. But in the case of this processor, I think more effort is put into the image processing and the method by which the image processing is accomplished, then in most other processors. Specifically, part of this is aimed at maintaining very low noise levels even with higher pixel counts and slightly smaller pixel pitch. Maintaining per pixel sharpness becomes more and more difficult to realize as pixel counts go up in APS sensors. But I think that the primary difference is the dynamic range optimization. Which, as far as I know, hasnt really been attempted in SLRs to date. Up until now, and again, Im not aware of each and every product on the market, but up until now, its my opinion that most every image optimization scheme thats been attempted to date has been done after the image has developed.
DCI:
Meaning in the software rather than in the hardware.
Mark Weir:
Well, not only in the software rather than in the hardware, but after the image has been processed, in other words, the process is not applied to the raw data coming off of the sensor. If you think of the shooting process as a sensor capturing the data, the data being processed, and then the data being recorded to media, specifically or perhaps exclusively, all of the optimization has been performed on the image after it has been processed into a JPEG and before it is written to the media. So, in other words its not just post-capture, its post processing.
DCI:
So its basically like applying auto levels to the RAW files?
Mark Weir:
No, its much more than a parameter, we have slides that would probably help explain it in greater detail for you, but I think that theres at least four or five separate parameters which are evaluated and optimized in the process, and to do that on a 10 MP RAW file without negatively impacting the processing speed and therefore without negatively impacting the shooting speed is the real breakthrough.
DCI:
Is there any difference in the way it performs those tasks on the RAW file versus a JPEG?
Mark Weir:
It doesnt perform those tasks on a JPEG at all. Dynamic Range Optimization in the camera is performed exclusively on the RAW data, never on the processed data.
DCI:
Does it boost information in shadows more than highlights or are both given equal weighting?
Mark Weir:
Depends upon the mode. As you know, there are two different selectable levels of DRO: theres DRO standard and then there is DRO advanced. DRO standard is primarily manipulation of Gamma, its not exclusively manipulation of gamma but its primarily that. It covers the entire image, and its primary purpose is to recover or extract shadow detail. DRO advanced can extract detail from the shadowed areas, but it performs the analysis on an area by area basis and it also manipulates additional parameters so it is able to recover the shadow detail, as well as the highlight detail, without affecting the entire image. The primary reason for having separate standard and advanced is essentially the processing time, because we cant execute the advanced process in the time necessary to maintain the three frames per second shooting rate.
DCI:
Is there any way to disengage the Dynamic Range Optimizer?
Mark Weir:
Oh, sure.
DCI:
You can disengage it completely?
Mark Weir:
Oh yeah, you can turn it off. You can et it for standard or you can set it for advanced. Its up to you. We know that there are plenty of people who would say no, no, no, no, no, Im not going to let my camera do that. Im going to do that in post processing in Photoshop. Those people are going to say no, I capture everything in RAW and I process every one of my images by myself, and I perform all of that optimization on my own and were like, fine, no problem. But if you think about it, the time it takes to do thatI mean, thats a very involved process.
DCI:
Absolutely, it becomes much more of a control issue at that point.
Mark Weir:
Yeah, and what were saying is that through the aid of this processor, were able to take an operation that is a minimum of, oh I would say twenty minutes, thirty minutes, and, for some people, hours and reduce it to a near instantaneous process. But we are not requiring it in any way and of course you can always capture RAW and JPEG and leave the RAW data as it is and enjoy the dynamic range optimization in the file that is ultimately processed to JPEG and written to memory. But in no way do we require that, we just think that its a major step forward that a camera could put forward processing at this level of sophistication and do it in the routine operation of shooting. I mean, you can literally enjoy DRO standard while blitzing along at three frames per second unlimited.
DCI:
Yeah, thats great.
Mark Weir:
And that, I think, is the primary accomplishment of the Bionz processor. I dont know anything that can do that.