Bayer - Foveon comparison

Started Nov 23, 2008 | Discussions
Arky Regular Member • Posts: 433
Bayer - Foveon comparison

Sigma SD14 vs Canon 50D (15.1 MP and 1.6X crop factor). The same lens was use for both cameras, a Canon 200mm f/1.8 L stopped down to f/4.

SD14, out of camera JPG (extracted from X3F).

Full size available at:
http://www.pbase.com/bigflat/image/106242475
RAW X3F file available at:
http://cid-709471d4dadefb7b.skydrive.live.com/self.aspx/shared/SDIM3020.X3F

50D, out of camera JPG.

Full size available at:
http://www.pbase.com/bigflat/image/106242504
RAW CR2 file available at:
http://cid-709471d4dadefb7b.skydrive.live.com/self.aspx/shared/IMG | 3026.CR2

Feel free to download, modify, print, crop, re-post and point out the obvious deficiencies of these comparison images.

 Arky's gear list:Arky's gear list
Sigma SD1 Merrill Sigma DP1s Canon EOS-1D Mark II Canon EOS 50D Sigma SD14 +14 more
SigmaChrome Forum Pro • Posts: 11,180
Re: Bayer - Foveon comparison

Pretty much what I would have expected.

Foveon: Rich, full and ripe colour and tones. Sigma: Really need to nail the exposure accurately. In this case, it seems pretty good.

Bayer: Acceptable but flat colour and tones. Canon: Really need to nail the exposure accurately.

Did I say the same thing about exposure for both Canon and Sigma... Well... I did too. Have a look at the rendition of the sky in the Canon 50D shot: Flat, dull and boring - But it's a 14bit, 15MP file (in RAW) - so there is plenty more information that can still be recovered. The SD14 picked up all that info - and more without having to look for it.

That's MHO, for what it's worth.
--
Regards,

Vitée

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 dp0 Quattro +31 more
mike earussi Veteran Member • Posts: 8,171
Re: Bayer - Foveon comparison

Standard conclusion:

Foveon is prettier and more dimensional, Bayer has higher resolution. Now if Sigma would just make a FF chip...

AdamT
AdamT Forum Pro • Posts: 58,986
Be careful what you ask for

otherwise you MAY get it ......

Most if not all of Sigma's standard zooms aren't pretty on full frame unless Heavily stopped down, the edges on the likes of the 24-60, 24-70 and 28-70 EX DG Macros don't get reasonable until F8 or smaller is achieved and the typical Sigma sample variation omn these lenses means that you'll be trawling through crates of the things to find one which is any good full stop - the wider EX primes are just as bad or worse on FF (the 24 F1.8 EX needs F16 and the 14mm never makes it) .. Unless Sigma plan on using the EOS mount or even Nikon so you can use their glass, I really hope they avoid full frame like the plague ..

What they would benefit from is using a 1.5X crop as all their APS lenses handle that crop factor and it gives a bit more real estate.

-- hide signature --

Please ignore the Typos, I'm the world's worst Typist

 AdamT's gear list:AdamT's gear list
Canon PowerShot G1 X Sony RX100 III Sony RX10 III
jonny1976 Veteran Member • Posts: 4,269
Re: Bayer - Foveon comparison

the sky is simply terrible...it has an awful green cast. probably the real sky was dull as in the canon sample. i like the sigma shots more, but the sky is terrible, not rich or others.

SigmaChrome wrote:

Pretty much what I would have expected.

Foveon: Rich, full and ripe colour and tones. Sigma: Really need to
nail the exposure accurately. In this case, it seems pretty good.

Bayer: Acceptable but flat colour and tones. Canon: Really need to
nail the exposure accurately.

Did I say the same thing about exposure for both Canon and Sigma...
Well... I did too. Have a look at the rendition of the sky in the
Canon 50D shot: Flat, dull and boring - But it's a 14bit, 15MP file
(in RAW) - so there is plenty more information that can still be
recovered. The SD14 picked up all that info - and more without having
to look for it.

That's MHO, for what it's worth.
--
Regards,

Vitée

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

-- hide signature --
 jonny1976's gear list:jonny1976's gear list
Nikon D2X
Martin Digital Regular Member • Posts: 173
Re: Bayer - Foveon comparison

i played around in CS3 acr with cs2 raw file. i was trying to emulate the sigma look, kind of.

probably best viewed in a color profile aware browser such as Safari, IE doesn't handle embeded color profiles.
here is my take

Martin Digital Regular Member • Posts: 173
Re: Bayer - Foveon comparison

I forgot to mention, I took as the reference the X3F file as opened in SPP and then selected x3f settings.
--
http://www.flickriver.com/photos/binarymartin/popular-interesting/
http://www.flickr.com/photos/binarymartin/

Bodoh Contributing Member • Posts: 552
Re: Bayer - Foveon comparison

Pretty much sums up my standard complaint about many images I see here. The un-natural turquoise skys! What's with that? Maybe it's the famous SD-14 green cast. I think it's awful - unless of course you like that interpretation. Fortunately I've never had that from my DP-1.
--
http://www.pbase.com/knobby

 Bodoh's gear list:Bodoh's gear list
Sigma DP1 Sony Alpha NEX-7 Sigma 30mm F2.8 EX DN Sigma 19mm F2.8 EX DN
I.B. Regular Member • Posts: 435
Re: Bayer - Foveon comparison

Humm...jaggies, of corse, resolution maybe, but DR...

Compare at 200% the upper side (the wheels section and the wire netting).

In SPP2,5 histogram and color correction, zero sharp, saved in tiff 13,9Mpx"Large prints" then converted to 1,5Mb jpeg in PSE.

The whole Sigma image: http://img509.imageshack.us/img509/7345/sdim3020biguz8 . jpg (Copy taking away the space before jpg)

-- hide signature --

Greetings from Italy.
Italo B.
http://italo.fotopic.net/

 I.B.'s gear list:I.B.'s gear list
Sigma DP2 Merrill Sigma DP3 Merrill +4 more
abarth
abarth Regular Member • Posts: 309
Re: Bayer - Foveon comparison

Ok, here's my try after 30 secs in CameraRAW/Photoshop.

I selected a point on the bridge for a custom white balance in CameraRAW and then applied auto color in PS. The difference is not that big anymore. Except for the leaves, which are more yellow on the Sigma.

Your Canon shot:

The PSed Sigma Shot:

 abarth's gear list:abarth's gear list
Sigma DP1 Sigma DP1 Merrill Sigma SD14 Sigma SD1 Merrill Sony a7 III +12 more
SigmaChrome Forum Pro • Posts: 11,180
Re: Bayer - Foveon comparison

Okay, there is a little green cast in the original SD14 picture... one click later and it was fixed:

1. Loaded into PSP PX2 (could have been PS or a number of similar programs).

2. Auto colour balance applied.

3. Click Apply -> Fixed.

Some folks can also easily do this in SPP, using the Color Wheel.

Probably the sky was NOT as dull as the Canon sample.

Why? Because the colours in the rest of the image indicate a bright day with a clear blue sky. Look at the reflections on the water. Foveon cannot make up ALL the colours in an image using pure speculation.

jonny1976 wrote:

the sky is simply terrible...it has an awful green cast. probably the
real sky was dull as in the canon sample. i like the sigma shots
more, but the sky is terrible, not rich or others.

-- hide signature --
 SigmaChrome's gear list:SigmaChrome's gear list
Sigma DP1 Sigma DP2 Merrill Sony RX100 Sigma DP3 Merrill Sigma dp0 Quattro +31 more
Ghibli Regular Member • Posts: 162
Re: Bayer - Foveon comparison

Sigma WB is way off (I suppose chronic WB issues got misnamed as "Green Cast" notwithstanding sometimes the error is heavily onto Yellow; this time the err was on the side of Cyan). Once corrected the image is fine. When both files were enlarged to A3+ and seen at actual print size (about 31% onscreen) Sigma still, by a small margin, made a better image (impression of acutance, 3D and colour fidelity) IMO.

Thank you very much indeed for generously posting the RAW files.

SandyF Forum Pro • Posts: 14,941
SDx detail

one of the first things I noticed on the photos is the greater, clearer detail in the foliage... the tree on the left, the background underbrush, etc.
More branches, leaves, in the underbrush, less yellow mush.

"Color" is usually easily corrected, but detail captured is... detail... Also the versions show that the processing is key

X3F and auto extracted embedded JPEGs of course are just basic unedited versions. Most photos benefit from some editing of contrast, highlight, satuation, color, fill light, etc, ie all the regular stuff.
Best regarsd, Sandy
sfleischmann@gilder.com
http://www.pbase.com/sandyfleischman
http://www.flickr.com/photos/sandyfleischmann

 SandyF's gear list:SandyF's gear list
Sigma DP2 Sigma DP1 Sigma DP1 Merrill Sigma DP2 Merrill Sigma DP3 Merrill +7 more
DMillier Forum Pro • Posts: 21,151
Re: Bayer - Foveon comparison

Pretty different interpretation from mine.

These sorts of differences in opinion continue to amaze me - and I'm not claiming any special skill in judgement here - but it seems to me that you are plain wrong!

Clearly there is a big colour difference - who can say which is accurate but this could be eliminated or reduced with processing.

But the big shock is the huge resolution advantage of the canon - quite different to the Kodak 14nx vs Sd14 test and quite startling for me.

Rather than argue with words, here's some evidence.

Try this:

  • equalize the file sizes (I used bicubic to 6000px just so both files went through the process)

  • add some substantial sharpening to bring out the differences.

-Compare at 100%

Crops:

SD14 (textured surface):

Canon

Sd14 (foliage - where SD is reputed to be good):

Canon

To my eyes the difference is about as great as the difference between a 6MP and 12MP camera.

Personally, i'd want to verify this with other shots, as the difference is surprisingly out of step with what I'd expect but from what I'm seeing here, I'd have to question DPreview's critcisms of the 50D - in resolution at least, it looks to be in a different class.

I wonder how big a print you would need to show this.

SigmaChrome wrote:

Pretty much what I would have expected.

Foveon: Rich, full and ripe colour and tones. Sigma: Really need to
nail the exposure accurately. In this case, it seems pretty good.

Bayer: Acceptable but flat colour and tones. Canon: Really need to
nail the exposure accurately.

Did I say the same thing about exposure for both Canon and Sigma...
Well... I did too. Have a look at the rendition of the sky in the
Canon 50D shot: Flat, dull and boring - But it's a 14bit, 15MP file
(in RAW) - so there is plenty more information that can still be
recovered. The SD14 picked up all that info - and more without having
to look for it.

That's MHO, for what it's worth.
--
Regards,

Vitée

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

-- hide signature --
DMillier Forum Pro • Posts: 21,151
Re: SDx detail

Sandy

I'm amazed at what you have just said. In these images a half blind person could see that the Canon absolutely trounces the Sd14 in foliage detail. The difference is so huge it's almost a joke.

If you really believe what you just said, then I reckon you also believe the Springboks scraped a lucky win against England yesterday...

SandyF wrote:

one of the first things I noticed on the photos is the greater,
clearer detail in the foliage... the tree on the left, the background
underbrush, etc.
More branches, leaves, in the underbrush, less yellow mush.

"Color" is usually easily corrected, but detail captured is...
detail...
Also the versions show that the processing is key
X3F and auto extracted embedded JPEGs of course are just basic
unedited versions. Most photos benefit from some editing of contrast,
highlight, satuation, color, fill light, etc, ie all the regular
stuff.
Best regarsd, Sandy
sfleischmann@gilder.com
http://www.pbase.com/sandyfleischman
http://www.flickr.com/photos/sandyfleischmann

-- hide signature --
SandyF Forum Pro • Posts: 14,941
Re: SDx detail

DMillier wrote:

Sandy

I'm amazed at what you have just said. In these images a half blind
person could see that the Canon absolutely trounces the Sd14 in
foliage detail. The difference is so huge it's almost a joke.

I see the detail differently, maybe because I'm looking especially at the CENTER area background foliage and underbrush rather than the extreme right where your crop comes from.

Remember the SD14 photo is a JPEG from unedited embedded in-camera JPEG by the way. Not the best result from a SD14; detail from a processed RAW is better.

If you really believe what you just said, then I reckon you also
believe the Springboks scraped a lucky win against England
yesterday...

No idea what you're talking about...

Shows that one/two photos don't prove anything. I see detail I like, time and again from my SD14 and DP1. I made a number of 11x14 prints yesterday.

Best regards, Sandy
sfleischman@gilder.com
http://www.pbase.com/sandyfleischman
http://www.flickr.com/photos/sandyfleischmann

 SandyF's gear list:SandyF's gear list
Sigma DP2 Sigma DP1 Sigma DP1 Merrill Sigma DP2 Merrill Sigma DP3 Merrill +7 more
DMillier Forum Pro • Posts: 21,151
Re: SDx detail

I agree you can't make a final determination from a small sample of photos.

We already know how good the Sd14 can be, so I don't have a critcism of the Sigma here, I'm just amazed at the quality of the 50D which - at first sight anyway - seems much better than I would expect...

I'm printing some sections of these images now to get a better impression.

SandyF wrote:

DMillier wrote:

Sandy

I'm amazed at what you have just said. In these images a half blind
person could see that the Canon absolutely trounces the Sd14 in
foliage detail. The difference is so huge it's almost a joke.

I see the detail differently, maybe because I'm looking especially
at the CENTER area background foliage and underbrush rather than the
extreme right where your crop comes from.
Remember the SD14 photo is a JPEG from unedited embedded in-camera
JPEG by the way. Not the best result from a SD14; detail from a
processed RAW is better.

If you really believe what you just said, then I reckon you also
believe the Springboks scraped a lucky win against England
yesterday...

No idea what you're talking about...

Shows that one/two photos don't prove anything. I see detail I like,
time and again from my SD14 and DP1. I made a number of 11x14 prints
yesterday.

Best regards, Sandy
sfleischman@gilder.com
http://www.pbase.com/sandyfleischman
http://www.flickr.com/photos/sandyfleischmann

-- hide signature --
SandyF Forum Pro • Posts: 14,941
Re: SDx detail

DMillier wrote:
....

I'm printing some sections of these images now to get a better
impression.

I'd suggest at least that you print from a properly processed X3F RAW rather than the embedded default JPEG. We know that the photos from RAWs are better than the JPEG engine.

We also don't know the breeze factor ... there are ripples on the water... presumably the leaves/trees were blowing too.

Best regards, Sandy
sfleischmann@gilder.com
http://www.pbase.com/sandyfleischman
http://www.flickr.com/photos/sandyfleischmann

 SandyF's gear list:SandyF's gear list
Sigma DP2 Sigma DP1 Sigma DP1 Merrill Sigma DP2 Merrill Sigma DP3 Merrill +7 more
DMillier Forum Pro • Posts: 21,151
Some sanity restored

I printed the middle section to an equivalent print size of about 20 inches for the whole image.

On the print to my eyes the Canon image still seems somewhat superior in detail and realism but only by a small amount. You would only see it if you were looking for it and it is broadly in line with the pixel count advantage over the Kodak.

Phew, just goes to show that what you see at 100% rarely means anything in a print.

I'll take your advice and try and produce some more optimal prints as well a double check - perhaps that will equalize things. I'd still be interested to know exactly at what print dimensions become visible.

ps

It's also worth noting that the Sigma's crop factor gives it a noticeable telephoto effect advantage if the lens is good.

SandyF wrote:
DMillier wrote:
....

I'm printing some sections of these images now to get a better
impression.

I'd suggest at least that you print from a properly processed X3F RAW
rather than the embedded default JPEG. We know that the photos from
RAWs are better than the JPEG engine.

We also don't know the breeze factor ... there are ripples on the
water... presumably the leaves/trees were blowing too.

Best regards, Sandy
sfleischmann@gilder.com
http://www.pbase.com/sandyfleischman
http://www.flickr.com/photos/sandyfleischmann

-- hide signature --
jonny1976 Veteran Member • Posts: 4,269
Re: Bayer - Foveon comparison

for me the jpeg showed from the OP told the contrary.

pixel per pixel the sd14 is by far the beter image in the post. everything is better, dynamic , crispness, color, microcontrast. Look at the bottom gray structure for example. the image seems to come out the video the canon instead is simply soft. I don't know if it's a problem of sharpnening.

Sd14 is supposed to be a good 10 million sensor compared to ccd sensor, clearly if you upres it 4 times it will lose against a 15 million pure sensor that you are upresizing not more than 1,2 times!!!

P.s: canon 200 1,8 is one of the sharpest lens ever made from canon, at f 4 is near the sweet spot. Instead also with this lens i'm not still convinced of the imag producd by canon 50d.
--
http://www.pbase.com/jon1976

 jonny1976's gear list:jonny1976's gear list
Nikon D2X
Keyboard shortcuts:
FForum MMy threads