Ideal resolution for a full frame Foveon sensor?

Started Mar 4, 2012 | Discussions
Roland Karlsson
Forum ProPosts: 21,095Gear list
Like?
Re: 32 MP and binning to 8 MP and 2 MP !
In reply to DaSigmaGuy, Mar 5, 2012

FF - 32 MP and binning to 8 MP and 2 MP ! All x3 of course.

That would be cool.

But then Sigma really has to design some high quality FF lenses.

-- hide signature --

Roland

support http://www.openraw.org/
(Sleeping - so the need to support it is even higher)

X3F tools : http://www.proxel.se/x3f.html

 Roland Karlsson's gear list:Roland Karlsson's gear list
Sigma DP3 Merrill Sigma DP2 Quattro Sony Cyber-shot DSC-RX100 III Pentax K-3 Pentax smc DA 15mm F4 ED AL Limited +8 more
Reply   Reply with quote   Complain
Roland Karlsson
Forum ProPosts: 21,095Gear list
Like?
Re: Ideal resolution for a full frame Foveon sensor?
In reply to SigmaChrome, Mar 5, 2012

SigmaChrome wrote:

Do you actually know if the sensor is still in production or if Sigma is just using up old stock?

Good question. It of course depends on how well the cameras sells. And ... that we dont know. Because Sigma keeps it secret.

-- hide signature --

Roland

support http://www.openraw.org/
(Sleeping - so the need to support it is even higher)

X3F tools : http://www.proxel.se/x3f.html

 Roland Karlsson's gear list:Roland Karlsson's gear list
Sigma DP3 Merrill Sigma DP2 Quattro Sony Cyber-shot DSC-RX100 III Pentax K-3 Pentax smc DA 15mm F4 ED AL Limited +8 more
Reply   Reply with quote   Complain
Roland Karlsson
Forum ProPosts: 21,095Gear list
Like?
Re: Ideal resolution for a full frame Foveon sensor?
In reply to mike earussi, Mar 5, 2012

mike earussi wrote:

I agree. I'm tired of Sigma of Sigma trading IQ for absolute resolution. I'd be willing to bet that a 15-20mp FF sensor would produce much better results than what I've seen from the SD1 so far with all its shadow banding and noise at iso 100.

Yes, a 15-20 MP FF sensor would be much better. There are some potential problems.

  1. Cost

  2. Heat

  3. Lenses

-- hide signature --

Roland

support http://www.openraw.org/
(Sleeping - so the need to support it is even higher)

X3F tools : http://www.proxel.se/x3f.html

 Roland Karlsson's gear list:Roland Karlsson's gear list
Sigma DP3 Merrill Sigma DP2 Quattro Sony Cyber-shot DSC-RX100 III Pentax K-3 Pentax smc DA 15mm F4 ED AL Limited +8 more
Reply   Reply with quote   Complain
SigmaChrome
Veteran MemberPosts: 7,191Gear list
Like?
Re: IMHO...
In reply to zodiacfml, Mar 6, 2012

zodiacfml wrote:

Maybe, they actually are, they could even have working samples only limited to R&D department.

I feel FF is the next iteration of the Foveon sensor to keep up with competition in the future. Imagine the FF with an option similar to the D800 that can switch into DX or APS-C format.

That would be absolutely brilliant! We could then still get some use out of our DC lenses.
--
Regards,

Vitée

Capture all the light and colour!

http://www.pbase.com/vitee/galleries

 SigmaChrome's gear list:SigmaChrome's gear list
Sigma DP1 Sigma DP2 Merrill Sony RX100 Sigma DP3 Merrill Sigma SD14 +13 more
Reply   Reply with quote   Complain
Lin Evans
Forum ProPosts: 15,819Gear list
Like?
Re: Don't understand...
In reply to Rawmeister, Mar 6, 2012

I was responding to this:

but I'd also be happy with one increased to 1.5 crop factor.

The SD1 sensor "is" a 1.5 crop...

Lin

Rawmeister wrote:

Yes, but the Sd1 sensor is not FF. It's 23.5mm across, or 1.5 crop. The original post was about what we would like to see for res on a FF Foveon at 36mm across.

The idea is that if we had it full frame size it would give better performance with respect to image quality and be less demanding of lens resolution because the pixel pitch would be much larger.

-- hide signature --

learntomakeslidshows.net

Reply   Reply with quote   Complain
Rodrigo Cunha
Contributing MemberPosts: 698
Like?
Re: IMHO...
In reply to SigmaChrome, Mar 6, 2012

And it would cost only $19999.95 with a bundled 24-70

I mean, if you have that kind of money, you could get yourself a Leaf Aptus-II 12 with 80 high quality megapixels today!

Reply   Reply with quote   Complain
petr marek
Regular MemberPosts: 174Gear list
Like?
Re: Ideal resolution for a full frame Foveon sensor?
In reply to DaSigmaGuy, Mar 6, 2012

I would be happy with 8Mpx (x3) Foveon with larger photosites. What is fundamental is absence of typical Foveon noise with higher ISO, low light sensitivity and widest possible dynamic/tonal range. The size of the sensor should reflect this. Also, the bigger is the sensor, the bigger is the problem with chromatic aberration with standard lenses, as we can see with SD1.

Also I would be overjoyed for Foveon full HD video. That would be groundbreaking. And it needs just 2Mpx on the sensor... And 8Mpx is great for pixel binning...

From my perspective I don´t need full frame and extreme resolution, I would be happy, if Foveon delivers just the best 8Mpx I can get of any digital camera.

Reply   Reply with quote   Complain
Mostly Lurking
Senior MemberPosts: 2,103
Like?
Re: Don't understand...
In reply to Lin Evans, Mar 6, 2012

An SD-15 sensor increased in size to 1.5 crop. Yes, the SD-1 is a 1.5 crop sensor, but it has a much greater pixel density than the SD-15.
--
William Wilgus

Reply   Reply with quote   Complain
mike earussi
Veteran MemberPosts: 5,008
Like?
Re: Ideal resolution for a full frame Foveon sensor?
In reply to Roland Karlsson, Mar 6, 2012

Roland Karlsson wrote:

mike earussi wrote:

I agree. I'm tired of Sigma of Sigma trading IQ for absolute resolution. I'd be willing to bet that a 15-20mp FF sensor would produce much better results than what I've seen from the SD1 so far with all its shadow banding and noise at iso 100.

Yes, a 15-20 MP FF sensor would be much better. There are some potential problems.

  1. Cost

  2. Heat

  3. Lenses

-- hide signature --

Roland

support http://www.openraw.org/
(Sleeping - so the need to support it is even higher)

X3F tools : http://www.proxel.se/x3f.html

There are always technical problems with any new concept or invention. The key is whether you are willing to pay the cost in R&D to solve them (and actually, lowering the pixel density may also help lower the heat).

Reply   Reply with quote   Complain
Mostly Lurking
Senior MemberPosts: 2,103
Like?
You mean I should throw away my SD-14 and SD-15?
In reply to Rodrigo Cunha, Mar 6, 2012

That's funny, especially if you look at this SD-10 shot:

I still have the SD-10 as well.
--
William Wilgus

Reply   Reply with quote   Complain
Rawmeister
Regular MemberPosts: 272
Like?
Re: Don't understand...
In reply to Lin Evans, Mar 6, 2012

No, the SD15 is a 1.7 crop factor. 20.3mm across

Reply   Reply with quote   Complain
Usee
Senior MemberPosts: 2,059Gear list
Like?
Re: Ideal resolution for a full frame Foveon sensor?
In reply to DaSigmaGuy, Mar 6, 2012

DaSigmaGuy wrote:

I just did a calculation to work out what the resolution of a full frame Foveon sensor would have, if it has the same number of photosites/mm as the SD1 sensor has...It works out to be about 34.6mp x3...Or as Sigma would say, 103.8mp!

The same calculation but with the same photosites/mm as the SD15 returns a much more modest figure of 14mp x3...Or as Sigma would say, 42mp...Less than the resolution of the SD1 with a 1.5x crop sensor!

Clearly the photosite size of a full frame Foveon sensor could be made much larger and hence less noisier, with a much higher dynamic range and much better high ISO capability, whilst still allowing a sizeable increase in resolution over the SD1.

Personally, I'd be more than happy with a 20mp x3 full frame Foveon sensor because of the fantastic image quality it would give compared to the SD1.
For IQ and filesize sake, I think Sigma should not go above 25mp x3.

So what resolution would like to see Sigma use for a future full frame Foveon sensor?

Beside the pixel size, one should also take in account that the layered sensor has a certain depth...

...thus one needs a suited light path (not only in the center, but also in the corners of the frame) to maintain pixel seperation and to avoid smearing and vignetting.

Therefore I think, that FF would only make sense (without further compromising the Foveon look) if the mount is big enough to allow telecentric lens designs,
like they probably use in the Merrill DPs...

...for the actual SA-mount one shouldn't go further than this:

http://forums.dpreview.com/forums/readflat.asp?forum=1027&message=38916088

... if uncompromised image quality really matters.

-- hide signature --

It doesn't matter what You do - it matters how You do it, to get satisfaction!

Envy is the highest form of recognition.

Stop to run, start to think.
Think twice - that doubles the fun!

Your world is as big, as Your mind.
Avoid to have only one point of view!

Uli

 Usee's gear list:Usee's gear list
Samsung NX1 Samsung NX 12-24mm F4-5.6 ED Samsung NX 45mm F1.8 2D/3D Samsung 16-50mm F3.5-5.6 power zoom Samsung 50-150mm F2.8 S
Reply   Reply with quote   Complain
zodiacfml
Regular MemberPosts: 489Gear list
Like?
Re: Ideal resolution for a full frame Foveon sensor?
In reply to petr marek, Mar 6, 2012

They should be now developing full frame sensors for production in the future just to stay relevant. I doubt Foveon can have improvements comparable to Sony's back illuminated sensors. As I recall from an interview, they're developing software fixes for lens problems since they predict there's still an MP race in the future, after all, the SD1 needs it.

If you ask me, a full frame with 15MP would be an ideal upgrade to their current sensor but I doubt we will get that at a reasonable price and pixel count, marketing it again as an MF competitor. They could do HD if they can quadruple the processor which is possible with a smaller chip manufacturing process which smartphones/tables today are fiercely competing on. That would be nice to speed up a Sigma camera at any resolution.

petr marek wrote:

I would be happy with 8Mpx (x3) Foveon with larger photosites. What is fundamental is absence of typical Foveon noise with higher ISO, low light sensitivity and widest possible dynamic/tonal range. The size of the sensor should reflect this. Also, the bigger is the sensor, the bigger is the problem with chromatic aberration with standard lenses, as we can see with SD1.

Also I would be overjoyed for Foveon full HD video. That would be groundbreaking. And it needs just 2Mpx on the sensor... And 8Mpx is great for pixel binning...

From my perspective I don´t need full frame and extreme resolution, I would be happy, if Foveon delivers just the best 8Mpx I can get of any digital camera.

 zodiacfml's gear list:zodiacfml's gear list
Sigma DP2
Reply   Reply with quote   Complain
cptrios
Senior MemberPosts: 1,347
Like?
Re: Ideal resolution for a full frame Foveon sensor?
In reply to DaSigmaGuy, Mar 7, 2012

DaSigmaGuy wrote:

The resolution of the SD1 sensor is 4800x3200, which works out to be exactly 200 photosites/per mm on both the x and y axis. 200 x 36mm = 7200. 200 x 24 = 4800. 7200x4800 = 34.56mp, or roughly 35mp rounded up. So the resolution of a full frame Foveon sensor with the same amount of photosite/mm as the SD1 sensor would be about 35mp.

Now, I know that a FF cut of the SD1 sensor would have its share of issues - noise, shadow banding, etc. And it would put more stress on lens corners than any sensor in history. But it would be ridiculously great for many things. Bear with me:

• Let's assume that the most common print size is 8x10. So, for the purposes of a 2:3 sensor, 8x12.

• Judging by my own playing with Foveon-produced files, I would say that with a good lens and good PP, it's more than possible to print an excellent 8x12 from a 1000x1500 (1.5MP) crop. Probably even smaller - but let's say that number's safe.

• Using DSG's calculations, one could therefore crop a good 8x12 from 4.3% of this theoretical FF sensor's total area. 20.8% of width and height.

Therefore, based on my rather clueless calculations of FoV (which I'm sure I'm doing incorrectly):

You could slap a Zeiss 21/2.8 Distagon on this camera, and, if you never planned on printing larger than 8x12, you effectively (ignoring DOF and perspective) would have a 21-120mm zoom. A Sigma 85/1.4 (provided you used the center, where sharpness is good enough) would give you an 85-410mm zoom. A good 200mm prime would get you as close as 850mm . A 28-300mm zoom (of which none are of course good enough for the purpose) would give you a whopping range of 28-1200m.

Like I said, I'm sure my calculations are off...but even so, the cropping power you'd have would be out of this world. You would, for example, get a nicely-detailed 8x8 of the Moon, where the Moon filled the entire frame, with a 300mm lens.

And, of course, uncropped with an excellent lens we'd be talking 7-foot prints that would hold up to close scrutiny, and probably 15-foot prints that'd be fine from a few feet away.

Reply   Reply with quote   Complain
DaSigmaGuy
Forum ProPosts: 12,276Gear list
Like?
Re: Ideal resolution for a full frame Foveon sensor?
In reply to cptrios, Mar 7, 2012

cptrios wrote:

DaSigmaGuy wrote:

The resolution of the SD1 sensor is 4800x3200, which works out to be exactly 200 photosites/per mm on both the x and y axis. 200 x 36mm = 7200. 200 x 24 = 4800. 7200x4800 = 34.56mp, or roughly 35mp rounded up. So the resolution of a full frame Foveon sensor with the same amount of photosite/mm as the SD1 sensor would be about 35mp.

Now, I know that a FF cut of the SD1 sensor would have its share of issues - noise, shadow banding, etc. And it would put more stress on lens corners than any sensor in history. But it would be ridiculously great for many things. Bear with me:

• Let's assume that the most common print size is 8x10. So, for the purposes of a 2:3 sensor, 8x12.

• Judging by my own playing with Foveon-produced files, I would say that with a good lens and good PP, it's more than possible to print an excellent 8x12 from a 1000x1500 (1.5MP) crop. Probably even smaller - but let's say that number's safe.

• Using DSG's calculations, one could therefore crop a good 8x12 from 4.3% of this theoretical FF sensor's total area. 20.8% of width and height.

Therefore, based on my rather clueless calculations of FoV (which I'm sure I'm doing incorrectly):

You could slap a Zeiss 21/2.8 Distagon on this camera, and, if you never planned on printing larger than 8x12, you effectively (ignoring DOF and perspective) would have a 21-120mm zoom. A Sigma 85/1.4 (provided you used the center, where sharpness is good enough) would give you an 85-410mm zoom. A good 200mm prime would get you as close as 850mm . A 28-300mm zoom (of which none are of course good enough for the purpose) would give you a whopping range of 28-1200m.

Like I said, I'm sure my calculations are off...but even so, the cropping power you'd have would be out of this world. You would, for example, get a nicely-detailed 8x8 of the Moon, where the Moon filled the entire frame, with a 300mm lens.

And, of course, uncropped with an excellent lens we'd be talking 7-foot prints that would hold up to close scrutiny, and probably 15-foot prints that'd be fine from a few feet away.

It does'nt work that way...You have to multiply the lens focal length by the crop factor (1.5x for the SD1M) to give you the FF equivalent focal length.

For example, the Zeiss 21mm will give you the same FOV as a 31.5mm lens would do on a FF camera. Cropping reduces resolution though so while the FOV gets less when you crop, sort of simulating having a greater focal length lens, resolution goes down which does not happen when you actually use a longer focal length lens. So cropping should only be used as a last resort, not all the time, and you would be far better off getting some longer lenses and stitching the resulting shots into panoramas to increase the FOV whilst giving you maximum image detail.

 DaSigmaGuy's gear list:DaSigmaGuy's gear list
Fujifilm FinePix F660EXR Sigma SD10 Sigma SD14 +25 more
Reply   Reply with quote   Complain
cptrios
Senior MemberPosts: 1,347
Like?
Re: Ideal resolution for a full frame Foveon sensor?
In reply to DaSigmaGuy, Mar 8, 2012

DaSigmaGuy wrote:

It does'nt work that way...You have to multiply the lens focal length by the crop factor (1.5x for the SD1M) to give you the FF equivalent focal length.

For example, the Zeiss 21mm will give you the same FOV as a 31.5mm lens would do on a FF camera.

We're talking about a FF sensor in this thread, are we not?

Cropping reduces resolution though so while the FOV gets less when you crop, sort of simulating having a greater focal length lens, resolution goes down which does not happen when you actually use a longer focal length lens. So cropping should only be used as a last resort, not all the time, and you would be far better off getting some longer lenses and stitching the resulting shots into panoramas to increase the FOV whilst giving you maximum image detail.

Obviously this is true, but as far as I'm concerned a good Foveon image loses resolution on cropping at a fraction of the rate of a Bayer image. That's the basis of my whole post!

Reply   Reply with quote   Complain
3dreal
Senior MemberPosts: 2,128
Like?
Re: Ideal resolution for a full frame Foveon sensor?
In reply to Rawmeister, Mar 8, 2012

Rawmeister wrote:

Considering the dodgy problems of getting the availalble lenses to resolve the sensors.... I'd say 15MP x 3 giving around 30MP bayer equivelent.

So same file size as SD1 except better ISO performance and less strain on lenses.
I figure you could use any lens u want since the pixel pitch is same as SD14/15.

And a Zeiss 21mm should be spectacular on it!

With unsharp corners.....

Reply   Reply with quote   Complain
Rodrigo Cunha
Contributing MemberPosts: 698
Like?
Re: Ideal resolution for a full frame Foveon sensor?
In reply to 3dreal, Mar 9, 2012

3dreal wrote:

Rawmeister wrote:

Considering the dodgy problems of getting the availalble lenses to resolve the sensors.... I'd say 15MP x 3 giving around 30MP bayer equivelent.

So same file size as SD1 except better ISO performance and less strain on lenses.
I figure you could use any lens u want since the pixel pitch is same as SD14/15.

And a Zeiss 21mm should be spectacular on it!

With unsharp corners.....

The Zeiss Distagon 21mm f/2.8 is extremely sharp, even in the corners, even wide open.

It's not perfect: there is a weird "moustache" distortion. But for nature shots you won't notice that, only the extreme sharpness,

I'm sure it has less resolution in the corners than in the centre, but you would probably need a pixel size of 2 microns or so (educated guess...) to clearly be limited by lens resolution in the corners.

Reply   Reply with quote   Complain
ABMountainHiker
Junior MemberPosts: 45Gear list
Like?
Re: Ideal resolution for a full frame Foveon sensor?
In reply to DaSigmaGuy, Mar 10, 2012

Thanks for that. So if I may interpret your OP suggestion a slightly different way: you're saying the 5 micron square pixel is perhaps too small (giving noise at low ISO it seems?) and that the 7.8 micron pixels of the older sensors (as in DP2) are too large (resolution too low)

You're proposing a FF sensor with an intermediate pixel size, say around 6.5 micron square.
Is that it?

Reply   Reply with quote   Complain
Keyboard shortcuts:
FForum MMy threads