Sigma SD1 Merrill instead of 6D or 5D mark III Locked

Started Jan 31, 2014 | Discussions thread
This thread is locked.
Jay Ell Regular Member • Posts: 256
Re: No - you make basic errors regarding optics, artifacts etc.

Kendall Helmstetter Gelner wrote:

Jay Ell wrote:

Kendall Helmstetter Gelner wrote:

But the Sigma would be better for landscape, by far.

No it would not. It would have one possible advantage of maybe slightly higher resolution - with aliasing as drawback. Apart from that it is hard to see why it would be better and the colours can be a big issue.

The lack of color moire (aliasing) is a plus.

Alliasign and color moire are not the same.

Which is why I separated the concerns.

You said: "color moise (aliasing)" - is certainly indicates that you think they're the same.

<...>But the other luminance aliasing is the whole REASON you use Sigma cameras over other cameras - because the aliasing mimics sub-pixel detail

No it doesnt. Aliasing does the exact opposite.

Not in actual pictures.

I disagree.

You apparently do not look at those I guess'

I do, no need start insulting. Or is that a moderator priviledge?

, again it's the appeal to many of the cameras. In a receding railing on a building you might see seven vertical lines where there are really nine - but the point is you still see vertical lines, and any inaccuracy is at the level of a pixel. So to the eye the scene is consistent.

I of course agree this is one of the things which appeal to many or most Sigma users. Perfectly fine

<...>No, this is not true, but it creates the impression of artificial order in the image.

Correct, which is the desired effect by those that like the camera. There is no difference to the eye between the artificial order and the real, because they mesh perfectly well in the image in the transition from real to artificial order.

Or course there is a big difference - human eye does not think that ordered lego bricks are natural detail. And in human made environment aliasing can run rampant and for many in disturbiing way.


You can easily blur something after the fact

Nonsense. Aliasing is difficult to cure - moire is the easiest subset, but regular jagged edges need heavy blurring to be fixed after the quantification - it is much less problematic to do it before sampling.

Real life use of images disagrees with you. Even just the natural slight blurring in real prints mostly dispenses with jagged edges.

Well, most images are never printed, but just looked on screen so unless you blur the aliasing remains. Also, I disagree with your statement - aliasing manifests itself in many ways in images - some may blur away to a degree in print, some do not.

Anyhow, interesting how now you say that the resoution advantage (per pixel) of Foveon is not important at all when you print as if there is enough blur to remove aliasing after they've created, then there is enough blur to remove any and all Foveon per pixel resolution advantage as well.

I will try to remember now for future discussions that when it comes to print you do not consider Foveon to have resolution advantage as that is the only logical conclusion from your words.

, you cannot take a smoothed image and provide more than a pale imitation of detail, like a watery diet coke next to the Real Thing.

I am sure you see nightmares then when 99.999% of all the images you see in the world are made with cameras using conventional image sensors if you indeed see all those images as pale imitations of detail. Funny how you see clear digitali artifacts, aliasing, as "Real Thing" or good representation of it.

Since mostly people post Bayer images at tiny fractions of the original size it's not really an issue.

Prints in magazines, advertisements,...

Now large (or even small) bayer prints, yes you can easily see the problems they introduce.

In large printes aliasing becomes apparent no matter how you try to disagree.

So I don't go to photographic galleries often unless the photographer uses medium format or a Sigma camera.

Extremely weird reason in my opinion. I go to galleries to see interesting or beautiful shots, not to see shots made with some particular camera. What the camera is doesn't interest me at all unless it ruins or makes the photo.

Sometimes I am in gallery show and often in other prints you can detect lots of issues... but that doesn't stop me from enjoying a good image, just a little sad it wasn't also technically better.

And this makes the reason above even more odd... isn't the work of art more important than the tool which was used to make it? I guess we all have our own little peculiarities.


I can understand how everyone does not like detail

You mean artifacts. Detail and artifacts are not synonyms.

They are in a Foveon sensor.

No they are not. When you get jagged edges or jagged details the do not look at all like details, but like errors and sometimes they are very irritating.

That's what you cannot grasp. But that is a large component of why people that have the cameras like the results so much.

Why should I grasp a falsehood. I do understand perfectly that many like the crispiness - sure I like crispiness, but it is not a free lunch as you think it is. Maybe you just don't see the aliasing errors, but many do. If you don't see then, it does not mean they do exist.

(well actually I can't but intellectually I observe this to be true), but saying that you shouldn't choose a Sigma camera because of aliasing totally misses the allure of the cameras to many.

I pointed out other reasons too which you conviniently ignored - high noise, inaccurate colours

I try to ignore things you say that are blatantly wrong. I guess I will not ignore it any longer.

Noise at low ISO is among the lowest of any camera around (especially true of the older cameras).

No it is not. Noise has two components: read noise and shot noise. The read noise of contemporary cameras is about 3 electrons. The best Sigmas have about 25 electrons (I've measured as have some others and you can too - use the PTC). The old ones have over a hundred electron read noise. This is per pixel. The shot noise is due to the poisson distribution of photons square root of the number of photons recorded by the pixel. The total noise is sqrt(shotnoise^2 + readnoise^2). In the shadows Foveon collapses. In the midtones and highligts the difference is insignificant regarding this noise. For Foveon next issue is colour conversion noise which is very significant - this is because the colour separation method is extremely weak and different from what human eye colour separation. If you shoot in black and white, outside of shadows Foveon is competetive.

You should really study noise. Emil Martinec's site gives a good introduction.

Yes, they have greater noise at high ISO.

Sigma ISOs are metadata. If the "high ISO" in Sigma is bad, then so must be the shadows in low ISO.

We all know that,

You said above: "Noise at low ISO is among the lowest of any camera around (especially true of the older cameras)." This is so wrong that you certainly can not say that "you all know that", who that "we all" is you talk about - you're an individual, not a spokes person for Sigma users. The new Sigma at image level has significantly better lower noise than the old one. This is a fact due to larger sensor collecting more photons and giving better SNR for midtones and highlights and the much lower read noise helping in midtones and shadows.

Please to not talk about noise if you do not understand it at all.

it's not like you are bringing anything new to the table

To you I'm bring lots, but I've understood many before me have failed to educate you. To me it seems that you do now want to learn - but please think about noise regarding the new Foveon sensor and the old one and tell me one good reason why the old onw would have better high exposure noise?

- but you are attempting to make that statement for all ISO settings which is incorrect.

All images have different tones. The shadow end of Foveons is bad.

If you continue to make that assertion without physical proof it's obviously only to mislead others.

You play the MOD card to silence me? You on the other hand have significant lack of understanding of for example basics of noise, yet you need to give physical evidence. Ok, here is one link for you think think: It's mostly about 2-layer detectors, but in the end Foveon X3 is also included in the tables. Please try to understand it and then accept it.

As for colors, your problem is again a total lack of experience with the subject.

See above link.

I've gone on many photo shoots in groups of people with other cameras. From those results, it's simply inaccurate again to claim the Sigma cameras have inaccurate colors when other cameras are getting colors wildly wrong

Adendotical evidence is not physical evidence. Ever heard of metameric failure? Also, see above link.

, or at least no better than the Sigma cameras I shoot with. There was an example of some Sony A7 same image I commented on where a blue bear was turned totally purple, or vice versa... anyway the Sigma camera got the color exactly right.

Sure. Obscure anecdotes are definative proof.

The Sigma cameras have a better overall color consistency than any other camera.

LOL. Really? Do you really believe in that? Seriously? Have you read this forum at all?

That is to say, if you get the white balance right the color accuracy is really good, better than you can get from most cameras.

Try reading the Eric Fossums link I wrote above. And then search this forum for Fossum and Foveon (use google to search for "Eric Fossum" and Foveon and ""-parameter). Seriously.

That's simply the observation I have from years of shooting them, why do you continue to claim otherwise not having shot any of them?

Have I not shot with any? How did you come into that conclusion? I have given a bit of hard evidence. Please give me counter evidence - not anecdotes or sample images, but something which is scienficly proper.

often hard to process well and so on. Aliasing is one more issue. Yet you choose to make the statement that it's far better option for this purpouse.

Consider buying the Canon for astro work, then a Sigma DP-1M, possibly also a DP-2M for landscapes.

Or maybe Sony A7r or Nikon D800 for both astro and landscape?

Both are not as good as the Sigma for landscape, and from many example photographs I've seen it seems like the Canon cameras are better for astro work.

Both are better than any Sigma for landscape.

Which you have no experience to make such a claim from.

According to which evidence? I may have shot with Siigmas, I've analyzed Sigma raw-files, and the colour issues are a scientific fact which has been pointed out counteless of times by people mcuh more knowledgeable on this than I am (like for example Eric Fossum). You may close your eyes from this, but if you do so, you should also shut up.

Neither has AA-filter by the way (well, D800E doesn't, or the Sony). Both alias, just like you like it.

I don't like color aliaisng.

I don't like any aliasing. I'd prefer D800 over D800E.

Both are better for astro than Sigma.

Not better than Canon though, just from results I've seen.

Canon's are usually good for asto, I agree. They also have very low noise in low light if high-ISO is selected, lower than the competition. Very good pixel design.

They will work, but if you are considering a camera specifically for astro work why not the most appropriate?

Why buy two cameras when for astro Canon and Nikon are within 5% of each other? And Nikon is better than any Foveon?

Or are you saying that 15Mp Sigma aliased image is better than 36Mp full frame aliased image

Yes, because of lack of color aliaisng which is impossible to fix.

Colour moire can be fixed quite easily. Lighroom for example has tool for it. It is a bit of manual labour though.

Of other aliasing the chroma-aspect can usually be removed so we're left with aliasing which is also an issue with Sigma.

when the Sigma has much lower SNR, more noise, inaccurate colours, limited processing support, lower image magnification need,...

Again, all inaccurate statements and not based on any real world results.

They are based on measurements and science. The full framers record much more photons than APS-Cs, foveon or not, the read noise is an order of magnitude lower and the colour conversion of Foveon creates noise. When you're ready for more evidence, after understanding and accepting the Fossum-doc I linked, I can give you more food if you like.

Do you seriously think you're being objective?

Well since I was the one that recommended the Canon in the first place, obviously yes.

LOL. You ignore pretty much all the Sigma and Foveon weaknessess, wether sensor or body and you call that objectivity.

Do you really think you are being objective when you can find nothing positive whatsoever to say about cameras you've never touched?

You read what you want to read - Foveon has nice resolution and local contrast. There just isn't really anything else positive about it as far as I can tell, so should I lie a bit to please you?

Or do I write this much because I somehow am afraid of Sigma cnquerng the world and making some other camera look bad? I'm an adult, not a child.

I have friends with a lot of different cameras - some really like Nikon, some like Canon. It seems way more likely to me that I know more about the totality of systems than yourself, who continues to get basic facts wrong about Sigma cameras.

It is you who have the facts wrong.

Besides, lets say you do want to use them for landscape - what lens could you put on them that would work as well as the DP-1M lens?

Lot's of them work just as well or even better. And have a few more focal lengths.

Name any that are wide and sharper than the DP-1M lens. None of the examples you gave are.

Really? The Voigtländer 21mm allows for higher resolution than DP-1M. Many many Leicas do too.

And if the lenses offer moire and other aliasing, they must be sharp. Heck, 15mm f/4.5 CV can produce color moire wide open on a camera with (weak) AA-filter. Thus it is very sharp.

And you have zero evidence for your claim, but it sure sounds like a lens being Sigma means it's the best there is.

Also personally I think full-frame cameras are nice for portrait and other close scenes, but if you are trying to shoot a lot of stuff with a high DOF then a crop sensor camera really helps you out

Nonsense. The small sensor does not offer ANY DOF anvantage.

Lin corrected you on your technical misunderstanding here.

Corrected? He is clueless about DOF as are you. How about studying it a bit? Or test it:

Take a picure with 50mm lens, crop it to quarter size. Then take a picture with 100mm lens and see for yourself what kind of apertures you need to have the same DOF. Crop camera has zero DOF advantage, any more than large-format has disadvantage. It's all just about the FOV, size of the exit pupil and focus distance.

-- hide signature --

Copyright © Jay Ell

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