Luminous Landscape Forum
Equipment & Techniques => Cameras, Lenses and Shooting gear => Topic started by: trevarthan on August 01, 2014, 03:08:51 pm
-
I've been using my 24mm PC-e on a Nikon D810 with a 105mm Polarizer on the Lee Foundation Kit front ring adapter (usually with two slots installed on the lee kit, but no filters in them). I've shot two scenes at f16 using the DOF scale on the lens to keep everything in "reasonable focus" with a 15 to 30 second exposure recently, in RAW, and edited using Lightroom CC. Both images were too soft for microstock.
This waterfall photo, in particular, was abnormally soft, I thought (I actually took this with the third stack of filter holders installed in the Lee Kit - would that make a difference?):
(https://farm6.staticflickr.com/5556/14763716672_303180a0c5_c.jpg) (https://flic.kr/p/ouBUBb)Benton Falls (https://flic.kr/p/ouBUBb) by Trevarthan (https://www.flickr.com/people//), on Flickr
The guy at Shutterstock was kind enough to send me a 100% crop so I can be lazy and not make one myself:
(http://i61.tinypic.com/bg9sv9.jpg)
(I can just barely see some blur there - it makes my brain hurt a bit for some reason)
I recently submitted that file for the third time, this time downsampling from full res to 24mp, hoping that would be adequate.
Is f16 just too soft for stock? Or am I doing something wrong? Should I be using f8 and focus stacking instead? What's the best way to get the best result in the majority of circumstances?
-
Is f16 just too soft for stock? Or am I doing something wrong? Should I be using f8 and focus stacking instead? What's the best way to get the best result in the majority of circumstances?
The D800/800E/810, will start losing microcontrast with f5.6 or narrower apertures. A lot of that can be restored with proper Capture sharpening. F/16 is pushing it, and is much more difficult than f/11, because some detail can no longer be recovered in most cases.
Other than that, a bit of deconvolution sharpening will drastically improve the image, without producing offensive (to stock agencies) halos. See attached example.
Cheers,
Bart
-
The D800/800E/810, will start losing microcontrast with f5.6 or narrower apertures. A lot of that can be restored with proper Capture sharpening. F/16 is pushing it, and is much more difficult than f/11, because some detail can no longer be recovered in most cases.
Other than that, a bit of deconvolution sharpening will drastically improve the image, without producing offensive (to stock agencies) halos. See attached example.
Cheers,
Bart
That looks quite good. How did you do that?
-
That looks quite good. How did you do that?
FocusMagic to restore resolution from diffraction blur, and a Blend-if layer at 80% reduced opacity, both to mitigate the risk of clipping the sharpened high/low lights. There is more that could be done, but that would require the 16-bit/channel source image, rather than a JPEG as input.
Cheers,
Bart
-
FocusMagic to restore resolution from diffraction blur, and a Blend-if layer at 80% reduced opacity, both to mitigate the risk of clipping the sharpened high/low lights. There is more that could be done, but that would require the 16-bit/channel source image, rather than a JPEG as input.
Cheers,
Bart
This FocusMagic? http://www.focusmagic.com/
-
This FocusMagic? http://www.focusmagic.com/
Yes!
-
Yes!
Ok. Looks fairly inexpensive, so I'll consider it. I'd like to focus on how to get better images in camera though. Could I have used tilt on the PC-e and maybe f8? Or maybe focus stacking? The PC-e doesn't have a DOF Scale mark for f11, just f16 and f8, so I'm not sure how I could use f11 reliably.
-
This is a very easy case for DoF stacking.
I would think that 5 images at f6.3 will have enough overlap.
Helicon focus is a bit expensive, but V6 works real well and would crunch a perfect image in less than one minute.
Cheers,
Bernard
-
How many degrees did you tilted your lens? ::)
-
The PC-e doesn't have a DOF Scale mark for f11, just f16 and f8, so I'm not sure how I could use f11 reliably.
Do not rely on the DOF scale mark on the lens. The criteria used for those scales is a 8"x10" print viewed at a distance of 10" ~ 12". That is far (very) from 100% in a 36MP image.
-
I agree with the previous post.
It doesn't seem like you're using tilt. The dof scale on the lens is calculated for a film standard. If someone is viewing the file at 100%, it would be better to calculate dof with a smaller circle of confusion. Instead of a hyperfocal distance of 4' for 24mm, for a d800 it could be closer to 8'. But using live view with magnification is probably the best way to make sure everything is sharp enough.
-
The D800/800E/810, will start losing microcontrast with f5.6 or narrower apertures. A lot of that can be restored with proper Capture sharpening. F/16 is pushing it, and is much more difficult than f/11, because some detail can no longer be recovered in most cases.
Other than that, a bit of deconvolution sharpening will drastically improve the image, without producing offensive (to stock agencies) halos. See attached example.
Cheers,
Bart
But Baaaart...., ...you've ended the "subject"! ;)
-
This is a very easy case for DoF stacking.
I would think that 5 images at f6.3 will have enough overlap.
Helicon focus is a bit expensive, but V6 works real well and would crunch a perfect image in less than one minute.
Cheers,
Bernard
So.... this waterfall image would work well for focus stacking because the light is pretty consistent, but would sunsets work? Wouldn't the light change too quickly?
According to this, I don't need Helicon. I just need Photoshop, which I already have: http://digital-photography-school.com/getting-landscapes-sharp-focus-stacking/
-
So.... this waterfall image would work well for focus stacking because the light is pretty consistent, but would sunsets work? Wouldn't the light change too quickly?
According to this, I don't need Helicon. I just need Photoshop, which I already have: http://digital-photography-school.com/getting-landscapes-sharp-focus-stacking/
You need to keep the exposures reasonably short, but it is often not a problem because
1. You use an aperture like f8,
2. Raising the ISO up to 400 ISO is often manageable from an image quality standpoint,
3. You can capture the sky in a single image (several if you combine this with a stitch),
Cheers,
Bernard
-
How many degrees did you tilted your lens? ::)
I didn't, because I'm not very good with it. I'm useless with it outside a studio tabletop environment or simple perspective corrections. Takes forever to get the focus plane right. I thought I could get away with f16, but it looks like that won't work if I want the best quality images. I'm starting to think I just need to master the tilt shift lenses I have for now. Maybe that would solve my focus problems in the majority of landscape scenes by allowing me to use f8. I was brushing up on my movement theory this evening. Might be time to just practice until I'm proficient.
-
I believe that the split screen zoom capability of the D810 in live view in portrait mode makes it really easy to use with T/S lenses.
I need to try that with my 24mm.
Cheers,
Bernard
-
Does the 810 not have focus peaking? That's the best way to see the plane of sharpness move across the image as you tilt.
Not used photoshop to stack for landscapes but in the studio I found helicon focus did a far better job than photoshop and with less errors especially for complex images.
-
Does the 810 not have focus peaking? That's the best way to see the plane of sharpness move across the image as you tilt.
No, it doesn't unfortunately.
Cheers,
Bernard
-
Focus peaking is available if you spend the $500 for the Sony 5 inch EVF - as my eyesight deteriorates,
it is a godsend on my D800E.
-
Focus peaking is available if you spend the $500 for the Sony 5 inch EVF - as my eyesight deteriorates,
it is a godsend on my D800E.
Do you have a link to that? I'm intrigued.
I believe that the split screen zoom capability of the D810 in live view in portrait mode makes it really easy to use with T/S lenses.
I need to try that with my 24mm.
Cheers,
Bernard
I've been playing with my 24mm PC-e and the d810 live view today, using my dining room table as a subject. Yeah, split screen only works for portrait orientation because the split is locked on the horizontal axis. Stupid limitation, IMO. I hope they fix that in a firmware update.
I really, REALLY miss live view zoom with wheel functionality from my D3s. Pushing the zoom in and out buttons repeatedly is way slower. Only a matter of time before they're using electrostatic touch pads like the ipad and double tap zoom.
I found the exposure meter/preview in live view (OK button). That helps a lot.
Also found the accelerometer level in live view. That thing is awesome. I think that eliminates the need for a flash shoe level.
Wish zebra stripes from the video mode live view worked on still frame live view. Another silly limitation I hope they fix in a firmware update.
I also swapped my 24mm PC-e with a screwdriver to shift and tilt in the same plane. I think that will help for landscapes. Luckily I didn't strip any screws. Tried the same thing on my 85mm PC-e and stripped two screws. I'll have to either drill them out or send them to Nikon for service. Too much lock tite. Lame. This needs to be a button feature, not a service call.
My gear head for my tripod hasn't arrived yet, so I had to measure the camera angle with a protractor. I think I finally got a successful tilt focus at 55 degrees tripod tilt and 3 degrees lens tilt. I'm trying to wrap my head around how the heck I'm going to use this to photograph landscapes in the field. Practice, I guess. I need a lot of practice. Also, is there a chart or something? Or an app for iphone to preview tilt DOF?
-
My testing shows that diffraction on my d800e, which I believe have the same sensor, starts to get visible at around f/9. From there you have up to about f/13 where you can recover most of it with deconvolution, but at f/16 some detail is lost.
Unless you want to do focus stacking, I recommend setting the sharpen radius in camera raw to about 1.1, and the strength/amount to about 40-50. Keep the detail slider low maybe 10. Maybe pull the clarity slider up to 10-20.
Then do another round of sharpening in PS using the smart sharpen filter set to "lens blur", with "more accurate" checked. Use a radius around 0.9-1.2 and adjust the amount to taste. The smart sharpen filter uses deconvolution so it might be able to recover some more detail.
I highly recommend helicon focus, it´s very precise, and if you get the helicon remote app you can use a tablet as a remote control for your camera, which then does the stacking for you, and you get reproducible and predictable results.
I would not go below f8 when focus stacking, as you risk have problems with uneven sharpness. You don´t lose detail at f8-13, and if it´s a little soft you fix that with the procedure described above.
-
Also, is there a chart or something?
Hi,
Yes there is a chart (http://www.northlight-images.co.uk/article_pages/using_tilt.html) for the plane of focus, the DoF is relatively vertical (more perpendicular to the plane of focus).
Cheers,
Bart
-
There is also an app called simple DOF (https://itunes.apple.com/us/app/simple-dof-calculator/id301222730?mt=8) that lets you specify the CoC that the calculations are based on. I use 0.20 to be on the safe side. Lens markings are usually based on 0.30 which is too big for the d800 sensor.
-
There is also an app called simple DOF (https://itunes.apple.com/us/app/simple-dof-calculator/id301222730?mt=8) that lets you specify the CoC that the calculations are based on. I use 0.20 to be on the safe side. Lens markings are usually based on 0.30 which is too big for the d800 sensor.
Hi,
Indeed, 0.030 mm (!) is way too large for the CoC. My DOF output quality tool puts it all in perspective, for the specific focusing distances (perpendicular to the focus plane). Of course, the D810 is selectable from the predefined list of brands/models since it was announced.
Cheers,
Barr
-
I've been playing with my 24mm PC-e and the d810 live view today, using my dining room table as a subject. Yeah, split screen only works for portrait orientation because the split is locked on the horizontal axis. Stupid limitation, IMO. I hope they fix that in a firmware update.
I really, REALLY miss live view zoom with wheel functionality from my D3s. Pushing the zoom in and out buttons repeatedly is way slower.
Are you aware that it is possible to customize the behavior of the central button the rear joystick so that one activations jumps to either 100% or 200% zoom in live view?
Cheers,
Bernard
-
Are you aware that it is possible to customize the behavior of the central button the rear joystick so that one activations jumps to either 100% or 200% zoom in live view?
Cheers,
Bernard
Well I am now! You just saved me nine clicks each way, man! Thanks!
-
Hi,
Yes there is a chart (http://www.northlight-images.co.uk/article_pages/using_tilt.html) for the plane of focus, the DoF is relatively vertical (more perpendicular to the plane of focus).
Cheers,
Bart
I read this, and I'm really really excited to see the math behind the tilt effect (well, charts and diagrams anyway - they suggest the math at least). This really clears a lot up for me and I think with a tape measure I can now quickly set my tilt focus, which is GREAT news.
However, it looks like this chart is for Canon lenses, as it has 90mm. Is there one for my 85mm Nikon?
-
Is there one for my 85mm Nikon?
Hi,
It's easy to calculate and make a table for your lens and for the distances relevant to your type of subjects.
Tilt angle in degrees = arcsin( (focal length / 1000) / Jdistance ), with focal length in millimetres, and J distance in meters.
So for an 85mm it becomes: Tilt angle = arcsin( 0.085 / Jdistance )
Cheers,
Bart
-
Hi,
It's easy to calculate and make a table for your lens and for the distances relevant to your type of subjects.
Tilt angle in degrees = arcsin( (focal length / 1000) / Jdistance ), with focal length in millimetres, and J distance in meters.
So for an 85mm it becomes: Tilt angle = arcsin( 0.085 / Jdistance )
Cheers,
Bart
Thank you. Given that information, I took an hour and a half this morning and created this sweet Google Spreadsheet of Nikon Tilt in Degrees for Distance: https://docs.google.com/spreadsheets/d/1kIL55Bet0PQnXhW5gHYf2BPyV0vG-XlLEL9E8ujUyFY/edit?usp=sharing
I've also attached an image version.
-
So how many actual filters did you have in front of the lens? A polarizer and 1 ND? f8 would surely be better and tilting your lens about 1° down from the top would be about perfect. You don't have any real serious DOF needs in your shot since the FG isn't right in your face.
I've had the Canon TS-E 17mm since it was introduced, and last month I got the 24mm II. Luckily they have easy user rotation in the field, so I don't need a Phillips screwdriver to swap directions!
Here's a good video to use your live view for nailing focus. https://www.joshuacripps.com/2014/06/focus-depth-of-field-landscape-photos/ And here's an article I wrote about the dreaded diffraction! https://www.facebook.com/notes/john-maclean-photography/pay-it-forward-photo-tip-1-dont-shoot-your-lens-at-f22/10151671520967312
-
I've been using the following method for a couple of years (I learned on Naturescapes.net from Royce Howland):
Tilt/Shift Without Tables - by Royce Howland
1. Using LV dialed up to 10X, adjust focus to nail infinity focus first, with the lens centered out (no tilt, no shift). Get your baseline exposure and focus first.
2. Shift to compose (if required).
3. Tilt for DOF. Use LV at 10X to determine how much tilt is needed to bring the FG into focus.
4. Repeat procedure until no further adjustment are required.
I threw out my tables shortly after learning this method. Seems to work just fine.
Glenn
-
http://www.luminous-landscape.com/tutorials/focusing-ts.shtml
-
1. Using LV dialed up to 10X, adjust focus to nail infinity focus first, with the lens centered out (no tilt, no shift). Get your baseline exposure and focus first.
Hi Glenn,
I still do not think that initial focusing on the infinity point is the most efficient starting point. It depends on lens design (http://www.luminous-landscape.com/forum/index.php?topic=79076.msg637493#msg637493), but e.g. for the T/S-E 24mm II, I can achieve results faster by starting at a focus distance that's a bit shorter, and with the T/S-E 90mm a focus distance closer to 1/3rd of the distance is even a better starting point to get there faster.
Starting at half the distance may be a more useful rule of thumb, if one doesn't know the lens more intimately already (like my 90mm example).
Cheers,
Bart
-
Too bad we don't have in DSLR lenses assymetrical tilt as on the Sinar/Ebony, that removes mostly the need to iterate to achieve focus.
Cheers,
Bernard
-
I'm not very good at links and such, but if you go to b&h and type in Sony 5 inch screen, that should do it.
They are about $500.
-
I don't mind carrying a tape measure. I've found that 90% of the time the tape measure and chart method gets my near focus point fairly sharp and all I have to do is focus to bring the far point in focus. I'm loving this method. Takes all the frustration out of the process. I'm making sharp images with tilt in a tenth of the time. I make them as quickly as I can perform a measurement and turn the focus ring. I don't know why anyone would do it any other way.
(https://farm4.staticflickr.com/3911/14878977831_905ed454ce_z.jpg) (https://flic.kr/p/oENDJx)
Glass Pedestrian Bridge and Hunter Museum at Sunset (https://flic.kr/p/oENDJx) by Trevarthan (https://www.flickr.com/people/26405131@N03/), on Flickr
(https://farm4.staticflickr.com/3884/14859035416_7f29a4c92e_z.jpg) (https://flic.kr/p/oD3rxN)
Glass Pedestrian Bridge and Aquarium at Sunset 24mm (https://flic.kr/p/oD3rxN) by Trevarthan (https://www.flickr.com/people/26405131@N03/), on Flickr
(https://farm6.staticflickr.com/5582/14852770976_d653e14857_z.jpg) (https://flic.kr/p/oCukmd)
Glass Pedestrian Bridge (person merged out) (https://flic.kr/p/oCukmd) by Trevarthan (https://www.flickr.com/people/26405131@N03/), on Flickr
(https://farm4.staticflickr.com/3857/14654313460_f4184af361_z.jpg) (https://flic.kr/p/ojXbQU)
Latest Iteration *Explored* (https://flic.kr/p/ojXbQU) by Trevarthan (https://www.flickr.com/people/26405131@N03/), on Flickr
-
Hi Glenn,
I still do not think that initial focusing on the infinity point is the most efficient starting point. It depends on lens design (http://www.luminous-landscape.com/forum/index.php?topic=79076.msg637493#msg637493), but e.g. for the T/S-E 24mm II, I can achieve results faster by starting at a focus distance that's a bit shorter, and with the T/S-E 90mm a focus distance closer to 1/3rd of the distance is even a better starting point to get there faster.
Starting at half the distance may be a more useful rule of thumb, if one doesn't know the lens more intimately already (like my 90mm example).
Cheers,
Bart
I don't disagree with this at all - in reality, I often focus on a "distant" object that is a hundred metres away - as we know, the difference between infinity and one hundred meters is negligible.
Glenn
-
I like Zerene Stacker for stacking – particularly because of the way that its non-destructive retouching has been designed and implemented.
-
http://www.luminous-landscape.com/tutorials/focusing-ts.shtml
More tables? Not necessary using the method I described (I didn't come up with it, so can't take credit for it).
I'd love to stack landscapes (have Zerene which I use for flowers when the wind isn't blowing), but when the wind is blowing on the water, the waves don't stack too well. :)
Glenn
-
More tables? Not necessary using the method I described (I didn't come up with it, so can't take credit for it).
I'd love to stack landscapes (have Zerene which I use for flowers when the wind isn't blowing), but when the wind is blowing on the water, the waves don't stack too well. :)
Glenn
You list tables in that article, and you don't list a table for the 85mm focal length. So, yeah. More tables. No big deal. It's just math.
-
Sorry, I didn't reference the LULA article on T/S.
That was John MacLean.
When I acquired my 24TSE three years ago, I posted an image on Naturescapes (I had been trying to use tables). I received a reply from Royce Howland (one of the Naturescapes editors), and adopted the method.
No tables, no tape measures, just LV and focusing. It's so dumb-a$$ simple it still amazes me.
Glenn
-
Sorry, I didn't reference the LULA article on T/S.
That was John MacLean.
When I acquired my 24TSE three years ago, I posted an image on Naturescapes (I had been trying to use tables). I received a reply from Royce Howland (one of the Naturescapes editors), and adopted the method.
No tables, no tape measures, just LV and focusing. It's so dumb-a$$ simple it still amazes me.
Glenn
Well, whatever works for the individual, I guess. I think the table method is dumb ass simple. I tried iterative methods for years. Sometimes it worked fine in a few minutes, but often it took a half hour or more for a single shot. With a simple chart and a measurement, I get it right the first time, usually, unless I look in the wrong focal length column.