Home

Please support mflenses.com if you need any graphic related work order it from us, click on above banner to order!
SearchSearch MemberlistMemberlist RegisterRegister ProfileProfile Log in to check your private messagesLog in to check your private messages Log inLog in

Sigma SD14 and Canon 5D comparison
Post new topic   Reply to topic
Home   Digital SLR Cameras
View previous topic :: View next topic  


PostPosted: Wed May 09, 2007 6:42 am    Post subject: Sigma SD14 and Canon 5D comparison Reply with quote
I came across very interesting comparison of SD14 (Foveon chip) and Canon 5D.
http://www.ddisoftware.com/sd14-5d/

Man.. I have a dream from last 5 years to have Foveon chip based camera. Seems like I am dreaming right stuff... Smile

Check the results from other guy,
http://www.flickr.com/photos/alroe/460007414/
http://www.flickr.com/photos/alroe/460013685/
http://www.flickr.com/photos/alroe/460006602/

I am looking forward for frank views with open mind. We all know canon/nikon/pentax all made best cameras Twisted Evil
But whats your take for these results...


PostPosted: Wed May 09, 2007 9:52 am    Post subject: Reply with quote
I don't have experience on the Sigma camera so I can not comment on it or on a comparison.

The comparison with the 5D seems a bit odd to me, I think the Sigma camera should be compared with another crop camera like the 400D.

The specs of the Foveon sensor appear to be superior to those of the Bayer sensor, we must however see how they translate into actual cameras. This is to say that although number are interesting, I prefer to compare actual cameras with comparable features.

The images in comparison are not good in my opinion: in the review, they are comparing with interpolated images. OK about the fact that the 5D itself interpolates bla bla... but I thought we were comparing the output of cameras, not post-production.
As for the other images, what can I comment? The only Sigma picture has the face of the model completely covered by a veil.
The only thing that I can comment on the second set of images is that I don't like the lighting setup at all, it seems completely wrong for a portrait, and neither I like the poses.

I am very interested in Foveon sensor cameras. I tend to agree with the reasoning that if I want to blur a picture I can do it myself.
I am surely waiting for Foveon cameras to fall down in prices. I will surely consider one the day they will arrive in the same price league as the Canon cameras.

However the full frame of the 5D is slowly spoiling me (true Michael?), and although I love my 400D, I don't know if I will want to buy another crop camera as third body. So I hope that Foveon sensor cameras can become full frame and in the same price league as the 5D. I have no idea on how much time it may take though.


PostPosted: Fri Jul 13, 2007 10:59 am    Post subject: Reply with quote
Hello,

I didn't knew that streching (enlarging) the SD14's image to a 12Mpx dimension (5D) the quality almost remains the same. What this means ?
Is the same situation with the FujiFilm FinePix S5 for example ?
Thank you.


PostPosted: Fri Jul 13, 2007 11:48 am    Post subject: Reply with quote
montecarlo wrote:
Hello,
I didn't knew that streching (enlarging) the SD14's image to a 12Mpx dimension (5D) the quality almost remains the same. What this means ?


Where did you hear this?

I think it may refer to the fact that Beyer sensors need stronger antialiasing, so the loss of definition caused by the interpolation necessary to enlarge the Sigma picture to the output size of a 5D, could be balanced by the stronger AA that the 5D applies to the recorded image.

I do not really agree with the statement however. First, the optical quality of the lenses changes: one thing is to reach the field of view of a 80mm focal lenght by using a 80mm lens on a full frame camera; another thing is to reach the field of view of a 80mm lens by using a 50mm lens on a 1.6 crop sensor camera.
The two images will cover the same field of view, but the image taken with the crop camera and the 50mm lens will have a deeper depth of field.

Another difference is that while antialiasing blurs the content of the actual recorded pixels, interpolation creates pixels where they did not exist. So no matter what some people says, the interpolated picture will not be as faithful to the subject as a non-interpolated picture, even an AAd one.

Having that said, I would not give a **** about these sophisms and if I had the money, I would enjoy a Sigma Foveon camera, regardless of how much smaller it is compared to a 5D.


PostPosted: Fri Jul 13, 2007 12:00 pm    Post subject: Reply with quote
Hello,

Orio wrote:
...

Where did you hear this?

...


Here ( http://www.ddisoftware.com/sd14-5d/ ) "On the surface, it may seem unlikely that a 4.6 megapixel image upsampled to 12.7 MP can look as good or better than one that started as 12.7 MP but the proof is in the shooting! Even though the SD14 photo on the right above started as a much smaller image, when upsampled to match the resolution of the 5D, it holds up very well, easily matching the 5D in most areas while surpassing it in others. ..."

Sorry. I'm not familliar with technical data about digital sensors so I was quite surprised seeing that a declared 4Mpx sensor enlarged ar 12Mpx (Width&Height of the picture) can produce almost the same quality picture (withount to see the pixels) and because of that the same print size, I think. Because of that I was wondering if that Fuji Super CCD sensor found on Fuji S5 works the same even it is not a Foveon sensor I believe.

Thank you.

Cosmin.


PostPosted: Fri Jul 13, 2007 1:09 pm    Post subject: Reply with quote
I read the article.
It seems to me that the guy is basically trying to make some noise in order to push the sales of his resampling software Qimage.
What he says about the Beyer limitations is totally true. But I don't agree with his conclusion that a 4,6 Mp image will look and print exactly like a native 12,8 Mp image, or better.
The samples he shows may _look_ similar (not all of them), but what about printing?
For good large prints you need much more than what he shows here as 100% crops.
For isntance he doesn't show any 400% pixel enlargements. i bet that if he does, the interpolated (or upsampled as he calls it) image will clearly show it's 4.6 Mp native resolution limits, while the 5D will stand up the pixel peeping much better.
In my opinion he's only proving that an upsampled 4.6 Mp image can _look_ the same or similar on a computer screen than a 12,8 Mp image.
Print enlargements, however, are a completely different story.

_


PostPosted: Fri Jul 13, 2007 11:24 pm    Post subject: Reply with quote
About the Fuji S5 Pro:

I read that it sports the Nikon F mount with limited compatibility. Which means that it can only mount some Nikkor AF lenses - not even all of them - and no manual Nikkors (except for manual, meter-less mode I guess).

This fact also means that it has the same register distance than Nikon cameras. And thus, it's pretty useless for use with manual lenses.


PostPosted: Sat Oct 20, 2007 9:39 pm    Post subject: Reply with quote
Wow LL,
you must know a lot. Thanks for all the information and the time you took to post it. I have really learned something that I didn't know from your message! It's great to have folks like you here who can teach something to the ignorant ones (like me) Smile


PostPosted: Sun Oct 21, 2007 10:53 am    Post subject: Reply with quote
Hi, LL!

Very informative post! Thanks a lot.

I agree to your and Orio's thoughts. I have not used the SD14, but I had the chance to shoot with a SD10 once. I was kind of amazed by the colours and the sharpness, but there was nothing I could not achieve with my EOS 350D, to be honest. (And I do not say that because I only own the 350D. Wink)

Nevertheless, if I could find a cheap SD9 (or even better a SD10) now that the SD14 has dropped in price already, I would love to get one.
Not because it is a better cam than the 350D (it is not!), but because the "feel" of what you get is a bit like shooting on film. That's a very personal view and I don't know if there is a scientific explanation, the picture just resemble film shots a lot.

Anyway, that's fantasy, since I would not be able to spend 400 or 500 Euro on another DSLR - or let's rather say, I would not be able to explain it to my wife. Wink Laughing

Last edited by LucisPictor on Sun Oct 21, 2007 12:58 pm; edited 1 time in total


PostPosted: Sun Oct 21, 2007 1:01 pm    Post subject: Reply with quote
Excellent examples, thanks!

LL, you seem to be really into all these sensor theories. What do you do for a living?


PostPosted: Sun Oct 21, 2007 1:06 pm    Post subject: Reply with quote
LensLunatic wrote:
This becomes particularly important when a sensor captures an object of high "self-smilarity" (eg a radiator, a fine-meshed wire fence, a tiled roof, or a plaid pattern) that is to say an object that resembles the sensor's own photosite array pattern


Now I make a probably very stupid comment: if the problem lies inthe rectangular pattern of the sensor cells, why the producers do not make sensors with cells arranged in a non-linear pattern? E.g. the so called "honeycomb" structure?

Quote:
Below just a "compatibility list" of Leica-R lenses for those who are interested in using old Leica glass. I've gleaned it from an other website some months ago .... Twisted Evil
http://www.wecal.de/MFPICS/Leica-R_List.jpg


Very useful, thank you!!

Quote:
I'd rather recommend you waste your funds on good MF glass.... Laughing Camera bodies come and go, good glass stays - if it has got enough resolving power for the next sensor generation .... Very Happy


Very true advice!


PostPosted: Sun Oct 21, 2007 5:15 pm    Post subject: Reply with quote
LensLunatic wrote:
...and I am now working as a teacher..... Very Happy


So am I. "Hallo, Herr Kollege!" Wink


PostPosted: Sun Oct 21, 2007 7:55 pm    Post subject: Reply with quote
Thank you, I have understood Smile

OK then the solution could be a single one giant cell that work like a television with no more cell separations but a single scanning line that goes VVVVVEEERRY fast throughout the surface of the single one cell. Very Happy

Laughing Sorry I know I'lm crazy sometimes Laughing


PostPosted: Sun Oct 21, 2007 8:58 pm    Post subject: Reply with quote
LensLunatic wrote:

That ain't crazy at all - such cameras actually EXIST!


Shocked Shocked Shocked
Ok, now I'm scaring myself. Laughing Laughing


PostPosted: Tue Oct 23, 2007 12:09 pm    Post subject: Reply with quote
Orio wrote:
LensLunatic wrote:

That ain't crazy at all - such cameras actually EXIST!


Shocked Shocked Shocked
Ok, now I'm scaring myself. Laughing Laughing


Indeed, a real One Pixel Digital Camera :

http://news.bbc.co.uk/2/hi/technology/6263551.stm


PostPosted: Sun Apr 06, 2008 9:49 am    Post subject: Reply with quote
I was interested in the Sigma SDxx cams too, but having had a closer look, I found (for me) two major issues:
- the real pixel count is much lower
- the special lens mount is quite limiting (since there are just a few adaptors available)

Cheers,


PostPosted: Mon Apr 14, 2008 11:59 am    Post subject: Reply with quote
kds315* wrote:
I was interested in the Sigma SDxx cams too, but having had a closer look, I found (for me) two major issues:
- the real pixel count is much lower
- the special lens mount is quite limiting (since there are just a few adaptors available)

Cheers,


Hi Klaus,

I've been looking at these Sigmas as well, and your first issue is what I don't like either. The second one however may be less accurate. Several sources claim that the SA mount is mechanically the same as the Pentax K mount, so any K mount adapter should function. M42 and M39 ones should not be a problem. Nikon F will not get to focus to infinity.

http://www.foto.nordjylland.biz/SD14/SD14-NonSigmaLenses.htm


PostPosted: Mon Apr 14, 2008 1:23 pm    Post subject: Reply with quote
Orio wrote:

Another difference is that while antialiasing blurs the content of the actual recorded pixels, interpolation creates pixels where they did not exist. So no matter what some people says, the interpolated picture will not be as faithful to the subject as a non-interpolated picture, even an AAd one.


Yes, but bear in mind that 50% of the green pixels and 75% of the red and blue pixels are interpolated, in a regular (Bayer) digital camera.


PostPosted: Mon Apr 14, 2008 1:34 pm    Post subject: Reply with quote
Orio wrote:
I read the article.
It seems to me that the guy is basically trying to make some noise in order to push the sales of his resampling software Qimage.

I think that is likely, yes.

Orio wrote:
What he says about the Beyer limitations is totally true. But I don't agree with his conclusion that a 4,6 Mp image will look and print exactly like a native 12,8 Mp image, or better.
_


Its important to note that we don't have any cameras which produce a native 12.8 MP image (or any other size).

The foveon sensor comes across as a contradictory beast because its either 4.6 or 18.4 MPx depending how you look at it. It does approximate to RGB capture at each photosite (I say approximate because, like colour film, the light is changed by the upper layers before getting to the lower layer sensors).

We tend to forget that Beyer sensors are also not true RGB. A 12.8 MP sensor has only 6.4 MP of green photosites and 3.2 MP of red and blue photosites. However, these are not co-located; which is a benefit for apparent monochrome resolution and a drawback for apparent chroma resolution.

So we are comparing two different approximations to a genuine RGB sensor, which does not currently exist (except in 3CCD movie cameras and some very old, 3CCD digital cameras).


PostPosted: Mon Apr 14, 2008 2:13 pm    Post subject: Reply with quote
ChrisLilley wrote:
...Its important to note that we don't have any cameras which produce a native 12.8 MP image (or any other size).

The Hasselblad H3D gets pretty close with an interpolated 39 MPix 48mm full frame sensor. Wink
Post new topic   Reply to topic
Home   Digital SLR Cameras
Page 1 of 1