[LFLee]

Just as an example. The following picture is of two birds at near the top of a 60' Gum Tree.

D800E AF-S 300/4 and TC 14E II in 1.5x crop mode: Scaled back to 2000 pixels on the long side

Wow... thanks Bossa. I have a similar set of Pentax gear like you and have been thinking switching to get Nikon FF. Just not sure how much I would 'lost' in this process. If Pentax FF is a sure thing, I will wait. I would even buy a K5IIs in the process of the wait. Right now I just haven't bought any Pentax stuff in half year (just bought some old film cameras) already. Is a tough decision because I really like Pentax and don't like the idea of holding a camera everyone else did.

Impressive images.

[bossa]

I've spent quite a bit of time trying to establish a relationship between LW/PH and IQ or apparent resolution in the real world, and as you know, I've been pretty much unsuccessful. The D800 just flat out kicks butt, but so many other FF systems don't. When you're talking lenses, I know I can see the difference between lens separated by 200 LW/PH , on the same sensor, but I can't seem to find a convincing argument for seeing a difference of LW/PH on sensors. For me the D800 stands out in a class of it's own, and no other DSLR camera seems to offer significantly more than a K-5. The only possibility I think of right now is that the D800 is the only FF camera that fully exploits the capabilities of an FF sensor using current technology, the difference being the MP count. I can't find a theory to explain it, I just know what I'm seeing.

If the D800 didn't come along I'd still have a K-5 and a Pentax system.
If Pentax had released a 24MP APS-C camera (k-3?) I would still have a Pentax system or at least a FA*300 or DA*300.
But the deed is almost done as I have three remaining pieces of Pentax kit to sell...

In a years time (after I have recovered from this switch) I would like to buy another Pentax camera (if they have their act together) and a few Limiteds along with a DA*55. I've always had Pentax cameras and whilst it's getting easier to relate to the fact I have a Nikon system now it still feels strange to me. I know that sounds ridiculous but there is a certain degree of emotion in all of this.

PS: Even if Pentax had just released a firmware upgrade to add focus peaking to the existing K-5 I'd I would have retained that body and a few lenses. They are just not 'giving' enough at the moment and I was feeling like I was being chained up in a corner with no other direction to go.

[bossa]

This is how I see these issues.

If a FF has a 4000 LW/PH resolution in the center of the frame and you 1/2 that sensor size and that system has a 2000 LW/PH then surely that is the same resolution [at the pixel level]? This is where that way of measuring things confuses people I think. If you remove the sensor size and use the LW/mm I believe you get a figure which allows an easier comparison across formats as long as you have an idea of the pixel density.

If the FF has a 'resolution' of 1800 LW/PH in the corners and the smaller sensor (1/2 size) has a 'resolution' of 900 lw/ph then that is the same resolution in absolute terms as well.

It's only when you strive for 'equivalence' that the resolution is less [with the smaller sensor of equal pixel density]. i.e. When trying to fit the same FOV in an image 1/2 the size. But in absolute terms the resolution at the pixel level should be identical. IOW, the ability of the system to resolve a double star should be identical.

A Practical Example: I have some panoramas that I did with my K-5 that get up to 150MP and they leave my D800E for dead.

[Allison]

Just as an example. The following picture is of two birds at near the top of a 60' Gum Tree.

And I thought I was okay with my decision to stay with Pentax and wait for unicorns...you are making me want to sell of and go Nikon again.

[bossa]

And I thought I was okay with my decision to stay with Pentax and wait for unicorns...you are making me want to sell of and go Nikon again.

An K-5, FA*300 (or DA*300) and a Tamron 1.4x TC should get identical results, if not slightly better (due to the extra million pixels), than these pictures. Only the TC might lower the quality a tad.

But as Norm has said, this is a D800 and it's 'also' a FF camera and not just a 15MP APS-C camera as seen here.

[Class A]

If you remove the sensor size and use the LW/mm I believe you get a figure which allows an easier comparison across formats as long as you have an idea of the pixel density.

However, this assumes that one always print/view images with a size proportional to the sensor size.

Your "same resolution" argument is valid if you print the image from the sensor that is twice as large, double the size of the other image.

In practice, that is not a reasonable assumption. While larger sensors enable larger output sizes, typically people use the same output sizes, regardless of sensor format (in which case the lines/picture-height metric is the relevant one).

Pixel peeping at 1:1 resolution has its place but I believe it is the wrong approach to evaluate (practical) resolution differences.

[bossa]

However, this assumes that one always print/view images with a size proportional to the sensor size.

Your "same resolution" argument is valid if you print the image from the sensor that is twice as large, double the size of the other image.

In practice, that is not a reasonable assumption. While larger sensors enable larger output sizes, typically people use the same output sizes, regardless of sensor format (in which case the lines/picture-height metric is the relevant one).

Pixel peeping at 1:1 resolution has its place but I believe it is the wrong approach to evaluate (practical) resolution differences.

I'm not talking about 'image resolution' or the difference in format resolution nor even output resolution but the resolving power of the optics and the sensor at the fundamental level. People were trying to ascertain what the ability of their current older lenses might be on a FF camera and I was thinking this could help determine that.

I realize this has no practical value when dealing with photographs day to day but thought it might offer a way to judge current lenses on future sensors. One would need to know the pixel pitch etc for this to work though.

PS: You are exactly right though when you say that output format is the determining factor when speaking about [image] resolution and this points to the arbitrary nature of the terminology used in these discussions.

---

Last edited by bossa; 11-15-2012 at 08:35 PM.

[bossa]

IS there any information around that tells just how accurately camera lenses are figured? Most telescopes are figured to within 1/8 to 1/10 of a wavelength.

[normhead]

Pixel peeping at 1:1 resolution has its place but I believe it is the wrong approach to evaluate (practical) resolution differences.

From PentaxForums.com: https://www.pentaxforums.com/forums/pentax-news-rumors/201457-ff-under-develo...#ixzz2CLoiefPG

That's just obtuse. If you can't see it in pixel by pixel comparison, in digital imagery, it isn't there. That's what is practical, what are you referring to as practical?

[WPRESTO]

IS there any information around that tells just how accurately camera lenses are figured? Most telescopes are figured to within 1/8 to 1/10 of a wavelength.

I have never seen such values for camera lenses, but I daresay they are not normally measured. For large reflecting scopes, the accuracy of the mirror surface from one side to the other is critical, and high precision manufacture is required to get a curved surface that is less than let's say 1/10 of visible light wavelength out of "perfect" over a diameter of 8 inches or more. But the vast majority of camera lenses are far less than 8 inches across, so it is much easier to maintain grinding accuracy from one edge to the other that is 1/10 wave or less. Collimation of numerous lenses ,up to 10 or more, both precision and especially consistently from unit to unit, is the real challenge.

[bossa]

I have never seen such values for camera lenses, but I daresay they are not normally measured. For large reflecting scopes, the accuracy of the mirror surface from one side to the other is critical, and high precision manufacture is required to get a curved surface that is less than let's say 1/10 of visible light wavelength out of "perfect" over a diameter of 8 inches or more. But the vast majority of camera lenses are far less than 8 inches across, so it is much easier to maintain grinding accuracy from one edge to the other that is 1/10 wave or less. Collimation of numerous lenses ,up to 10 or more, both precision and especially consistently from unit to unit, is the real challenge.

Thanks for the response.I made a couple of telescopes years ago and the testing process for the mirror was quite demanding. I can't imagine lens manufacturers doing that for each and every lens element they make and It's quite remarkable that they are able to maintain the standards they do if that is indeed the case. It would be interesting to see results from lenses made to the same exacting standards as a telescope mirror though.

[LXNights]

If Pentax does come out with full frame, I hope they get the selection of lenses for up to where they had it for the LX, PZ-1 and MZ-S...

[WPRESTO]

That is my gripe and the primary reason I have no interest in a Pentax FF. I retired quite a few good FF lenses soon after purchasing a K10 and many good lenses I now own are strictly ASP-C. How many current Pentax lenses are FF, and how long will it take to develop a meaningful selection, considering how long it takes to issue new ASP-C format lenses? I'd rather see a "K-PRO" ASP-C with 24mp, a tilt/swivel screen and good control of noise up to ISO 12,500. It would also be nice if Sigma were actually to deliver their 150 and 180mm macros in a K-mount. Delivery of the 150mm has been promised ever since that lens was introduced how long ago? 18 months or more?

[vonBaloney]

That is my gripe and the primary reason I have no interest in a Pentax FF. I retired quite a few good FF lenses soon after purchasing a K10 and many good lenses I now own are strictly ASP-C. How many current Pentax lenses are FF, and how long will it take to develop a meaningful selection, considering how long it takes to issue new ASP-C format lenses? I'd rather see a "K-PRO" ASP-C with 24mp, a tilt/swivel screen and good control of noise up to ISO 12,500. It would also be nice if Sigma were actually to deliver their 150 and 180mm macros in a K-mount. Delivery of the 150mm has been promised ever since that lens was introduced how long ago? 18 months or more?

The 180mm is now available or soon available as a special order according to another thread here, and they did put their fancy new 35/1.4 in K-mount, so that's good news...

[Ash]

Tilt/swivel screen isn't a pro feature, or the top end Canon and Nikon cameras would have it. But Ricoh are taking on the FF challenge and in time an upgrade pathway worth sinking our teeth into will be dished out. And so to with DFA lenses.