This topic is in the 2012 archive. You can read it and like its posts here. It is read only.
Archive 2012D800 spanked the 5DIII in every DXO category!
p.8 #1D800 spanked the 5DIII in every DXO category!
That's weird. I found the ZF.2 21mm needs almost no additional sharpening in most landscapes.
Up close though, not that strong.
Keep in mind that Canon TS/E is 24mm, a lot easier to make/design a 24mm lens than a 21mm...
Up close though, not that strong.
Keep in mind that Canon TS/E is 24mm, a lot easier to make/design a 24mm lens than a 21mm...
Apr 24, 2012 at 01:09 PM
p.8 #2D800 spanked the 5DIII in every DXO category!
FlyPenFly wrote:
That's weird. I found the ZF.2 21mm needs almost no additional sharpening in most landscapes.
Up close though, not that strong.
Keep in mind that Canon TS/E is 24mm, a lot easier to make/design a 24mm lens than a 21mm...
It's also easier to design a high performance 21mm lens than a 17mm but somehow Canon has managed to make it exceptional.
Just glad I have both.
That's weird. I found the ZF.2 21mm needs almost no additional sharpening in most landscapes.
Up close though, not that strong.
Keep in mind that Canon TS/E is 24mm, a lot easier to make/design a 24mm lens than a 21mm...
It's also easier to design a high performance 21mm lens than a 17mm but somehow Canon has managed to make it exceptional.
Just glad I have both.
Apr 24, 2012 at 01:29 PM
p.8 #4D800 spanked the 5DIII in every DXO category!
Tariq Gibran wrote:
https://www.fredmiranda.com/5DIII-D800/index_controlled-tests.htmlThanks for the link. That's an impressive demonstration of how good the Sony sensor is performing in the D800. Exciting to know that this level of performance is available now.The crazy thing is that I just checked what the NEX-7 could do with shadow recovery - shown below. Frankly, I was not expecting much at all but damn if it might be even better with the shadows than the Canon 5D MIII! I believe this example is even more extreme than what FM showed in his example.It "looks" more extreme because your original frame shows the Blacks slider set to "5". Set this to "0" and the Fill Light set to "0" and show us the screen capture of that alongside the other shot with Fill Light set to "80".Nope. It is extreme and a setting of "5" for the black point is the default setting for the NEX-7 by the way (and is not that much). Here is what changing the default of the black point to 0 gains you 
http://www.gibranstudio.com/o.jpg
That's a significant difference. I'm just saying, setting the black to anything above 0 will clip information to black and make it appear as there was no visible information to begin with. As shown here, there was.
All shadowed data is clipped in your other sample:
http://www.gibranstudio.com/drn7a.jpg
AngryCorgi wrote
Tariq Gibran wrote
AhamB wrote
Tariq Gibran wrote
Fred Miranda performed a really nice comparison between the D800 and 5D MIII which seemed to really pinpoint the pro's and cons's of both bodies. What really got my attention was the amazing way the Sony sensor in the D800 handles noise when recovering shadow detail. The Canon sensor is simply pathetic by comparison. Canon has a looong way to go beyond just MP's to catch up in sensor tech.https://www.fredmiranda.com/5DIII-D800/index_controlled-tests.html
http://www.gibranstudio.com/o.jpg
That's a significant difference. I'm just saying, setting the black to anything above 0 will clip information to black and make it appear as there was no visible information to begin with. As shown here, there was.
All shadowed data is clipped in your other sample:
http://www.gibranstudio.com/drn7a.jpg
Apr 24, 2012 at 02:44 PM
p.8 #5D800 spanked the 5DIII in every DXO category!
AngryCorgi wrote:
That's a significant difference. I'm just saying, setting the black to anything above 0 will clip information to black and make it appear as there was no visible information to begin with. As shown here, there was.
All shadowed data is clipped in your other sample:
The difference is visible and, of course, the blacks do get clipped above 0 but in the context of the comparison (in which I believe I pointed out that defaults were used), the less than .5 stop difference you are pointing out in the shadows sort of pales in the face of a 3 stop push. It simply does not change the point of the comparison by any significant amount.
That's a significant difference. I'm just saying, setting the black to anything above 0 will clip information to black and make it appear as there was no visible information to begin with. As shown here, there was.
All shadowed data is clipped in your other sample:
The difference is visible and, of course, the blacks do get clipped above 0 but in the context of the comparison (in which I believe I pointed out that defaults were used), the less than .5 stop difference you are pointing out in the shadows sort of pales in the face of a 3 stop push. It simply does not change the point of the comparison by any significant amount.
Apr 24, 2012 at 04:54 PM
Guest
Offline
Guest
Details
p.8 #6D800 spanked the 5DIII in every DXO category!
wayne seltzer wrote:
Let me know if you think this too much sharpening.
Just a bit oversharpened for my web taste. Try to reduce radius in USM - seems to be somehow "fat" sharpened.
Let me know if you think this too much sharpening.
Just a bit oversharpened for my web taste. Try to reduce radius in USM - seems to be somehow "fat" sharpened.
Apr 25, 2012 at 04:30 AM
p.8 #7D800 spanked the 5DIII in every DXO category!
Wayne, it does look "crunchy" and processed. Are you not able to use the raws from the D800 for some reason?
Apr 25, 2012 at 05:39 AM
p.8 #8D800 spanked the 5DIII in every DXO category!
^Heavy blocky posterization in the water. Looks like the tonal gradations have been destroyed by jpeg compression.
Apr 25, 2012 at 07:38 AM
p.8 #9D800 spanked the 5DIII in every DXO category!
Wow the right side of the pic (not sure if this corresponds to which side of the lens because of the orinetation) is super soft unless there's something wrong with the mount alignment on the lens or camera. The focal plane seems whack.
Also, way too much sharpening.
Also, way too much sharpening.
Apr 25, 2012 at 08:38 AM
p.8 #10D800 spanked the 5DIII in every DXO category!
The D800 should be able to recover quite a bit more than the NEX-7, and with better tonality (14-bit vs 12-bit ADC). The D7000 using the 14-bit option recovers lots of detail and with very good color accuracy:
Underexposed 6.5EV (blacks = 0)
http://farm5.staticflickr.com/4103/5149397045_86be3cdeba_z.jpg
Lifted in Raw Photo Processor (no NR at all)
http://farm5.staticflickr.com/4106/5150004530_7c3dfbf779_z.jpg
I would expect a BETTER performance (slightly) from the D800. Especially in regard to noise grain.
Underexposed 6.5EV (blacks = 0)
http://farm5.staticflickr.com/4103/5149397045_86be3cdeba_z.jpg
Lifted in Raw Photo Processor (no NR at all)
http://farm5.staticflickr.com/4106/5150004530_7c3dfbf779_z.jpg
I would expect a BETTER performance (slightly) from the D800. Especially in regard to noise grain.
Apr 25, 2012 at 12:50 PM
p.8 #11D800 spanked the 5DIII in every DXO category!
Tariq Gibran wrote:
All shadowed data is clipped in your other sample:The difference is visible and, of course, the blacks do get clipped above 0 but in the context of the comparison (in which I believe I pointed out that defaults were used), the less than .5 stop difference you are pointing out in the shadows sort of pales in the face of a 3 stop push. It simply does not change the point of the comparison by any significant amount.
To be honest, it doesn't really surprise me that the NEX-7 has better shadow recovery than the 5Diii. At this point, just about all of the cameras with Sony aps-c or 35mm sensors are essentially ISO-less, and you can shoot them at or near base ISO all of the time, and just adjust exposure in the converter. Of course, there are a lot of variables, especially in raw converter choice, but ISOs are starting to become irrelevant...unless you're using an EVF/LCD based camera, like the NEX-7, and you need to see the jpeg preview to shoot.
Ironically, even though these cameras are more forgiving with exposure, my exposures have been more consistent than ever with the live preview histogram.
AngryCorgi wrote
That's a significant difference. I'm just saying, setting the black to anything above 0 will clip information to black and make it appear as there was no visible information to begin with. As shown here, there was.All shadowed data is clipped in your other sample:
To be honest, it doesn't really surprise me that the NEX-7 has better shadow recovery than the 5Diii. At this point, just about all of the cameras with Sony aps-c or 35mm sensors are essentially ISO-less, and you can shoot them at or near base ISO all of the time, and just adjust exposure in the converter. Of course, there are a lot of variables, especially in raw converter choice, but ISOs are starting to become irrelevant...unless you're using an EVF/LCD based camera, like the NEX-7, and you need to see the jpeg preview to shoot.
Ironically, even though these cameras are more forgiving with exposure, my exposures have been more consistent than ever with the live preview histogram.
Apr 25, 2012 at 12:51 PM
p.8 #12D800 spanked the 5DIII in every DXO category!
6.5 stops underexposed? This has really stopped being about photography, and is now about something else.
Apr 25, 2012 at 01:44 PM
p.8 #13D800 spanked the 5DIII in every DXO category!
carstenw wrote:
6.5 stops underexposed? This has really stopped being about photography, and is now about something else.
It's still photography. If you're exposing for maximum dynamic range there is no such thing as underexposed.
6.5 stops underexposed? This has really stopped being about photography, and is now about something else.
It's still photography. If you're exposing for maximum dynamic range there is no such thing as underexposed.
Apr 25, 2012 at 02:00 PM
p.8 #14D800 spanked the 5DIII in every DXO category!
If you are exposing for photography, there is no need to push by 6.5 stops. Put your zones where they belong and you will never need to know the behaviour demonstrated above. This is just extreme. 3 stops, yes, but 6.5 stops is just nuts.
Apr 25, 2012 at 02:15 PM
p.8 #15D800 spanked the 5DIII in every DXO category!
carstenw wrote:
If you are exposing for photography, there is no need to push by 6.5 stops. Put your zones where they belong and you will never need to know the behaviour demonstrated above. This is just extreme. 3 stops, yes, but 6.5 stops is just nuts.
+1
Expanding DR past 3 or 4 stops, IMO, always looks unnatural whatever the technique. Paper or screen has less DR than the eye so it is great to be able to get a bit more out of the scene, one way or the other, but adding 6 stops starts looking pretty funky to me.
If you are exposing for photography, there is no need to push by 6.5 stops. Put your zones where they belong and you will never need to know the behaviour demonstrated above. This is just extreme. 3 stops, yes, but 6.5 stops is just nuts.
+1
Expanding DR past 3 or 4 stops, IMO, always looks unnatural whatever the technique. Paper or screen has less DR than the eye so it is great to be able to get a bit more out of the scene, one way or the other, but adding 6 stops starts looking pretty funky to me.
Apr 25, 2012 at 03:09 PM
p.8 #16D800 spanked the 5DIII in every DXO category!
carstenw wrote:
If you are exposing for photography, there is no need to push by 6.5 stops. Put your zones where they belong and you will never need to know the behaviour demonstrated above. This is just extreme. 3 stops, yes, but 6.5 stops is just nuts.
The zones are partially outdated, since digital has a very hard stop in zone 10. You could get away with putting target material up in zone XI and even XII on a negative - which you often do/did with scenes with very large contrasts - as long as you planned for some dodging / burning in the development.
In digital this is just not physically possible. Once the values in the raw file reaches 15,000 or so (different for different cameras) you get a full blank. Nothing. Flatspot with miscoloration in the entry/exit from the blown area.
But from a photographical PoV, the practical use for anything actually requiring more than -3Ev is severely limited. The type of lenses we use limit the usable range of the projected image, and the deepest shadow detail will be very diluted by stray light.
I've seen demonstrations of 18Ev capable surveillance systems, and even though some very competent lenses were used the shadows just got "cleaner" not "more detailed". Contrast was very, very low in the lowest stops. But it's still usable for areas like office entrance surveillance, in-car warning systems and so on.
If you are exposing for photography, there is no need to push by 6.5 stops. Put your zones where they belong and you will never need to know the behaviour demonstrated above. This is just extreme. 3 stops, yes, but 6.5 stops is just nuts.
The zones are partially outdated, since digital has a very hard stop in zone 10. You could get away with putting target material up in zone XI and even XII on a negative - which you often do/did with scenes with very large contrasts - as long as you planned for some dodging / burning in the development.
In digital this is just not physically possible. Once the values in the raw file reaches 15,000 or so (different for different cameras) you get a full blank. Nothing. Flatspot with miscoloration in the entry/exit from the blown area.
But from a photographical PoV, the practical use for anything actually requiring more than -3Ev is severely limited. The type of lenses we use limit the usable range of the projected image, and the deepest shadow detail will be very diluted by stray light.
I've seen demonstrations of 18Ev capable surveillance systems, and even though some very competent lenses were used the shadows just got "cleaner" not "more detailed". Contrast was very, very low in the lowest stops. But it's still usable for areas like office entrance surveillance, in-car warning systems and so on.
Apr 25, 2012 at 03:43 PM
p.8 #17D800 spanked the 5DIII in every DXO category!
carstenw wrote:
If you are exposing for photography, there is no need to push by 6.5 stops. Put your zones where they belong and you will never need to know the behaviour demonstrated above. This is just extreme. 3 stops, yes, but 6.5 stops is just nuts.
If you ever shoot ISO 6400 you are "underexposing" by the same 6 stops.
If you are exposing for photography, there is no need to push by 6.5 stops. Put your zones where they belong and you will never need to know the behaviour demonstrated above. This is just extreme. 3 stops, yes, but 6.5 stops is just nuts.
If you ever shoot ISO 6400 you are "underexposing" by the same 6 stops.
Apr 25, 2012 at 04:27 PM
p.8 #18D800 spanked the 5DIII in every DXO category!
AhamB wrote:
^Heavy blocky posterization in the water. Looks like the tonal gradations have been destroyed by jpeg compression.
Yes, I see it now on large screen. This was because I didn't reduce down to 1024 pixels on a side (did 2000 instead) and then had to save on too low of a jpeg quality level to keep it below the 450k upload limit.
@ Tariq, I didn't have LR 4 installed yet last night, so I thought I would quickly throw up this pic but I realized I needed to redo the step reduction sharpening alg I use since starting with higher rez D800 shot.
@Fly Yes, I noticed when I took shots of the Stanford church with this rented ZF 21 that it is decentered and soft on the right side which corresponds to the top of this picture where the log is going across. That is why that moss is blurred.
I focused on the mossy exposed tree root near the center of the pic just above the water. It is nice and sharp.
I also know that it is the ZF 21 being soft on the right side and not the camera because my 24/1.4G was equally sharp on both sides and had no problems.
Will post some better shots tonight now that I have LR 4 installed. Luckily my battery's ser # doesn't have the F in the 9th digit which correspond to the bad/recall batch of Nikon batteries.
^Heavy blocky posterization in the water. Looks like the tonal gradations have been destroyed by jpeg compression.
Yes, I see it now on large screen. This was because I didn't reduce down to 1024 pixels on a side (did 2000 instead) and then had to save on too low of a jpeg quality level to keep it below the 450k upload limit.
@ Tariq, I didn't have LR 4 installed yet last night, so I thought I would quickly throw up this pic but I realized I needed to redo the step reduction sharpening alg I use since starting with higher rez D800 shot.
@Fly Yes, I noticed when I took shots of the Stanford church with this rented ZF 21 that it is decentered and soft on the right side which corresponds to the top of this picture where the log is going across. That is why that moss is blurred.
I focused on the mossy exposed tree root near the center of the pic just above the water. It is nice and sharp.
I also know that it is the ZF 21 being soft on the right side and not the camera because my 24/1.4G was equally sharp on both sides and had no problems.
Will post some better shots tonight now that I have LR 4 installed. Luckily my battery's ser # doesn't have the F in the 9th digit which correspond to the bad/recall batch of Nikon batteries.
Apr 25, 2012 at 06:36 PM
p.8 #19D800 spanked the 5DIII in every DXO category!
snapsy wrote:
If you ever shoot ISO 6400 you are "underexposing" by the same 6 stops.
Really? I thought most DSLRs (esp. in the bracket of the D800) have analog gain up to ISO 1600. ISO 6400 would make for only 2 stops underexposure then.
If you ever shoot ISO 6400 you are "underexposing" by the same 6 stops.
Really? I thought most DSLRs (esp. in the bracket of the D800) have analog gain up to ISO 1600. ISO 6400 would make for only 2 stops underexposure then.
Apr 25, 2012 at 09:52 PM
p.8 #20D800 spanked the 5DIII in every DXO category!
AhamB wrote:
Really? I thought most DSLRs (esp. in the bracket of the D800) have analog gain up to ISO 1600. ISO 6400 would make for only 2 stops underexposure then.
You're still adding gain of some kind to make up the difference, whether it be analog or digital. The choice is up to you whether you want to use the camera's analog/digital gain, or in the case of EXMOR cameras with linear DR, you may want to just shoot at base ISO and boost exposure in the raw converter. Depending on the raw converter, you may actually improve your DR by boosting in the converter.
snapsy wrote
If you ever shoot ISO 6400 you are "underexposing" by the same 6 stops.You're still adding gain of some kind to make up the difference, whether it be analog or digital. The choice is up to you whether you want to use the camera's analog/digital gain, or in the case of EXMOR cameras with linear DR, you may want to just shoot at base ISO and boost exposure in the raw converter. Depending on the raw converter, you may actually improve your DR by boosting in the converter.
Apr 26, 2012 at 12:12 AM