Full Frame Foveon sensor 2020 (Pt. 2)

Started 7 months ago | Discussions thread
joe173 Regular Member • Posts: 445
Re: Full Frame Foveon sensor 2020 (Pt. 2)

TN Args wrote:

absquatulate wrote:

TN Args wrote:

absquatulate wrote:

TN Args wrote:

absquatulate wrote:

TN Args wrote:

target5 wrote:

For me, the fundamental thing is that is based on 1:1:1 original classic three equal layer Foveon tech, not Quattro and not Bayer.

The fundamental things are actually resolution and noise. Never worship a technology: just gauge results.

There are other fundamental things, but they relate to performance, and this forum is famous for not giving a toss about performance.


It's not a case of not giving a toss about performance, more a case of accepting the trade-off's of the foveon architecture because the benefits are worth it. I'm sure everyone would welcome improved noise performance, but not at the expense of resolution and colour, if I want that I can just shoot a bayer sensor and be done with it.

By "performance" I meant "camera performance", i.e. things other than IQ off the sensor.

I was using terms in the same way as many camera reviews, which have a section for "Image Quality" and a section for "AF and Performance" or similar.


Sure, the same applies, you don't get foveon IQ without a hit in performance due to the amount of data being handled, it would be nice to have faster cameras, but not at the expense of IQ, that's the area where bayer cameras excel.

You brought it back to the sensor's 'performance' instead of total camera performance, e.g. AF accuracy/speed/flexibility, camera responsiveness, boot-up times....

The two are intrinsically linked that's why, you can't look at either in isolation.

So sad that all the camera reviews do exactly that, but you want to score a point, so won't acknowledge the basic issue I am discussing.

As for "the amount of data being handled", this has never been true, and it sure won't be true of the FFF. There have always been equivalent Bayer cameras with just as many photodiodes of data as their Foveon contemporaries, hence just as much data to handle. And Sigma's latest camera is a mid-MP Bayer, and I see reports that the AF is borderline dysfunctional.

It obviously is true, just by looking at the sensor architecture alone. The camera has to combine the data of the three different layers to produce a file, rather interpolating a single layer. The battery size of the SD Quattro is comparable to a standard DSLR battery yet you'll be lucky to get more than 100 shots out of one on a good day. That's because the lengthy write times drain the battery, that's clearly a result of the computational requirements involved.

Absolutely not a necessary consequence of any of those factors you raise. If you have inefficient computation due to lack of resources to get it done right, then the process takes longer and flattens the battery. Similarly, if your hardware chip development (and LSI) is miles behind the competition, due to lack of resources and investment, your cameras are slow and chew up the battery, unnecessarily.

In the Q & A speech with the CEO, he admitted that Sigma cameras were less efficient in processing files due to software emulated rather than hardware ASIC support. This was because they could not use off-the-shelf chips. If I have some time, I will find his answer.

The actual issue is that Sigma may well be a first-tier lens company, but they are a third-tier camera company, and their cameras are non-performers. Their camera and sensor divisions are the CEO's 'hobby', and you can imagine their staffing, research and investment levels compared to Canon or Sony. I am grateful that the cameras aren't flimsy and unreliable, however basic they be. But if Canon or Sony had bought Foveon and pressed the sensors into their cameras, the cameras and the sensors would both be at a higher level than they are today.

If Canon or Sony had bought out a foveon sensor camera they may be able to improve the performance but you can bet your life they'd be charging a hell of a lot more for one. Either way this is all conjecture and we'll probably never know.

So, "10x to 20x the resources and investment gets better results" is unknowable conjecture? And yet, a big price hike is a dead certainty? Ever heard of economies of scale? Methinks you have the probabilities all back to front.

I stand by my claim that anyone who cares about camera performance simply isn't here discussing Foveon cameras, because the cameras wouldn't keep them here, so guess who's left?


Sigma produces cameras capable of fantastic image quality, which are well designed and built, at bargain basement prices.

The current models represent best resolution LPWH per dollar at base ISO. That's about the extent of the "fantastic IQ" equation: on most other IQ measures they trail.

The trade off is that you get cameras less capable in some areas than most other cameras. There isn't much point comparing their performance to a bayer camera, they're completely different beasts that are niche and best for specific things, not a jack of all trades. In terms of operational performance it's like comparing a standard saloon car to a forklift truck, utterly pointless.

Not pointless at all. When I am ready to buy into FF bodies, I will certainly be comparing. I'll bet I will not be alone in so doing. Not comparing your options, now that is for the birds.

I'd also say that Sigma have been producing cameras since the 1970's, so not exactly new to it. I'm sure people would love their Sigma cameras to perform as fast as the best bayer cameras, the reality is that it's never going to happen for the reasons I've stated. To suggest otherwise is to believe that Sigma deliberately cripple the speed and operational performance of their camera's, that's for the birds as far as I'm concerned.

It's never going to happen for the reasons I've stated. As I said right at the start of this sub-thread, "The fundamental things are actually resolution and noise. Never worship a technology: just gauge results. There are other fundamental things, but they relate to performance, and this forum is famous for not giving a toss about performance." You wanted to correct that statement, but not a single word stands corrected.


Post (hide subjects) Posted by
Keyboard shortcuts:
FForum PPrevious NNext WNext unread UUpvote SSubscribe RReply QQuote BBookmark MMy threads
Color scheme? Blue / Yellow