I finally got my hands on the 100 Makro Planar ZE to shoot with my 5DII. I am really enjoying working with it so far. I feel like when I look at the LCD while shooting the images have a nice pop to them, but when I get them into LR3 they lose a little something. I have the canon 200 1.8L and those images don’t really need processing to see the special properties; they really jump off the screen right away.
What kind of processing are you guys using to get that zeiss look (I don’t want to say 3D) that I see in the zeiss thread? I am only using LR3 (may upgrade to LR4 soon) but could do some photoshop.
Also if there are any particular apertures, or shooting techniques to maximize the pop, would love to know that too!
Could you maybe outline what the special properties are of the 200/1.8 which you feel come automatically, and give an example, and say what the special properties are of the 100MP which you feel you don't get? Is it just the 3D. "Just" is the wrong way to describe this, by the way; many people don't get the 3D look in spite of trying hard. It takes a lot of factors coming together right, and the Zeiss lenses only make it easier, they don't guarantee anything.
Thanks for this link. I would have a hard time describing what I am missing. Perhaps it is the 200 1.8L's unique "Eye of Sauron" effect that makes images from it jump out at me so much. It could be the sharpness and color too which are much better than any of my L series primes (they seem pretty flat by comparison as far as raw files go). I should try some side by side tests I suppose. I will try to find some images that are good examples.
sanproduction, before I started to use step sharpening method almost every photo I posted to internet was "well this looked good on my screen, but this websize image doesn't show the nature of this lens etc. etc.". Sharpening methods used in Lightroom and Aperture are based on principle "scale to final size and then sharpen". This method is only able to show the detail, which still has enough contrast (=edges and very rough textures) on the downscaled image. When I still shoot with Canon this didn't bother me much, but shooting with Leica and Zeiss and using this "scale to final size and then sharpen" methodology takes so much away from the images. These simple methods also tend to make DOF appear larger.
What comes to Zeiss rendering style big part of it are textures - if you want to preserve textures (or appearance of them) in your websize images I don't know what else to recommend than step sharpening - maybe "post bigger pictures" if you insist on using Lightroom. Step sharpening can be found from thread Carsten linked.
I have automated my own post processing to happen with ImageMagick script. Earlier I used PhotoShop but full Unix based script requires less actual work from me. Results are similar with both of these tools. I also read (haven't tried yet) that new version of GIMP finally supports 16-bit images, so soon it will be possible to do this stuff with it as well (I doubt they have yet all pieces in place e.g. have all filters to work in 16-bit, support to ICC profiles etc.).
Looking at the images you posted - comments from postprocessing (resize and sharpening) point of view:
#1 would have some texture/shape on the photographer, even such small size, if proper technique for resize/sharpen would have been used
#2 the ball shaped object completely lacks texture, I doubt it was that clean and smooth as it appears, lack of texture most probably caused by LR postprocessing
#3 just edges, not textures, damn same what sharpening method
#4 the "body" of the plant you focused should have some texture, now it appears as smooth "stick" without the texture this species has in the "body" in real life (sorry about terms like "body" and "stick", I don't know much about biology terminology in english)
#5 wow LR post processing really sucks, the road doesn't look any different in focus plane
#6 just edges, good results can be achieved with whatever resizing and sharpening technique (if this would be slightly larger and there would be enough pixels in needless to show the needle surface texture then naturally stepsharpening would be better)
#7 really weak detail in the leaves, due to LR post processing
#8 wow again, LR again manages to skrew up the concrete texture does look like it's made from plastic
Thanks for the great advice Samuli. I am not sure what step sharpening is, but I will look it up in the thread. I didn't make it all the way through, I was distracted by a project I got into.
Step sharpening is a way of preparing forum photos by resizing and sharpening alternately, until you reach the desired size. It retains much more of the micro-contrast and sharpness of the original.
Sorry to hijack this thread, but could Carsten and Samuli do me a huge favour and let me know if you think the images hosted on my website could benefit from step sharpening? I currently export a 4mb file from Lightroom and upload it to my website (cargo collective) which downsizes a 1024px wide image on my page. I used to think this was ok but after getting an iPad 3 I've started to become more concerned about what is actuall considered sharp and not just crunchy.
The only thing that has put me off before was keeping processing time to a minimise pp time. If it could be better I guess I could just take to ps and stepsharpen a 1024 for direct uploading.
@Robert: Nice images on your website.
Doing step sharpening isn't not difficult, but what will take most time is tuning the sharpening strengths to something that doesn't make your images too "crunchy" looking (which can also make the bokeh more harsh looking). It all depends on the type of image how much sharpening it can take.
Most of the stuff on your page looks great, although a few shots do like crunchy, like the third shot on the "April" page, the 2nd and 6th shot on the "Late Snow" page, and the first shot in "Hyde Park". I only checked the first two pages, as I am a bit short on time tonight.
Apart from that, a few shots suffer from stair-stepping, such as the "Newark" shots. These might also benefit from a gentler step-sharpening process, making sure that you never downsize by exact ratios, like 1/2 or 1/4.
I would say that the size difference between your portrait and landscape shots is too large. The landscape shots are a good size, but the portrait ones are very tall. Maybe stick to the same maximum dimension for each, or at least a slightly less dominant vertical size?
You have a lot of terrific shots, but there is also a fair amount of stuff that just isn't nearly as good, which lowers the overall standard of the site. Some is not focused right, in a way which detracts from the content, some has composition which is not so exciting, and so on. Often it looks like it might be older work. A few shots would benefit heavily from some purple defringing. If the site is partly just for friends, not for portfolio purposes, this is understandable, although you could separate the two functions, if you wanted. Some galleries are one shot, others are many. Generally, I think some curating would yield a much more enjoyable site of a higher standard, in case you wanted to take the time, and felt stronger about this than just keeping certain memories alive.
R.Young wrote:
Sorry to hijack this thread, but could Carsten and Samuli do me a huge favour and let me know if you think the images hosted on my website could benefit from step sharpening? I currently export a 4mb file from Lightroom and upload it to my website (cargo collective) which downsizes a 1024px wide image on my page.
Robert, I spent 2 hours in GIMP (few months back) to get post processing script to work in it. I failed, I would say you have very little change of getting it to work in 8-bit mode with any software. Based on that your have 4MB file it's either few hundred pixel TIFF or then JPG (=only available as 8-bit). So quality can be saved if your website script first changes 8-bit file to 16-bit file, but doing multiple operations in 8-bit file just tends to take the spark off the images.
Your website images, from technical presentation point of view, do seem better than average images in web. Majority of images we see in web are ridiculous size e.g. 600px in longer edge - I use sometimes that size thumbnails in Apple Aperture when browsing photos in fullscreen! In addition majority of images we see in web are raped by various photo sites resizing algorithms (some are less bad, e.g. if you upload really big image to Flickr it tends to come out halfdecent to largest downscaled sizes). Some of your not-so-perfect post processing is saved by having images large enough to not loose the smaller textures. "+1" to Carsten's points, except too large vertical images - I will never complain about too large images, even the vertical images didn't fit to 2560x1600 display in Mac (no F11 fullscreen mode), however I might not like them as much if I would have some tiny 1366x768 laptop to watch the photos...
Some of your images seem familiar - are they only on your website? I have slight feeling that I have did see them when I did search Scotland (planning to visit at some point this summer) images from Flickr but my memory doesn't always work flawlessly and I might mix up to something else (or I did search also from google in addition to Flickr).
R.Young wrote:
The only thing that has put me off before was keeping processing time to a minimise pp time. If it could be better I guess I could just take to ps and stepsharpen a 1024 for direct uploading.
You can create droplet from PhotoShop action (and depending your operating system you can "monitor" directory with script or program, which triggers the PhotoShop droplet) so it won't add any time, depending your computer speed 5-20s/image, which can be passive time, which you can use for making some coffee, smoking cicarette or read newspaper whatever else than sitting in front of computer (or if you wish you can sit in computer watch it to work if you prefer so...)
I wish I would have my website at home or somehow in same network; it's easy to automate post processing with ImageMagick. Transferring 120+MB 16-bit TIFFs to website would consume the network quota, due to which I'm having virtual Linux box running in my Mac with VMWare Fusion, which is automatically processing TIFFs placed to specific directory (creates multiple sizes and multiple different sharpening). Then I just delete all not needed images via Apple Finder and then execute SSL FTP transfer script (in the virtual Linux box) for remaining JPGs to specified directory in my website. For doing post processing to 10 images I need active computer time:
- 1s to start the script for resize/sharpening script (could schedule it to crontab if I would like to avoid this step)
- 100s (10x10s) to delete not needed sharpening levels and sizes
- 5s to start the transfer script
==> 106s for 10 images is not that bad
If I'm 100% sure what sharpening/resizing I need then active time drops to 6s for 10 images; most of the times I know these, but for example if I have shoot with "hard" lenses (2/100, 2/50, 2/25, 2/21 and some extend 2/28 and 2/35 thou not as distinctive as previously mentioned) and "smooth" lenses (1,4/85, 1,4/50 and 1,4/35) in same shoot I may prefer different sharpening levels to be created and then manually select the ones I want to be published to web. Even more so if I shoot with Zeiss and non-Zeiss lenses on same shoot. Naturally subjects can also have effect what script I use for sharpening, but usually I don't have for example architectural images with hard light and nature images in same shoot, so their TIFFs are not processed at same time.
carstenw wrote:
Step sharpening is a way of preparing forum photos by resizing and sharpening alternately, until you reach the desired size. It retains much more of the micro-contrast and sharpness of the original.
I have not seen much evidence for that statement, if any at all, provided one uses a proper tool for the single-step approach.
Thanks for your feedback guys, really appreciated. The fact that you pointed to shots mainly taken on film and developed at a high street chemist is a relief. They came back on disk plenty crunchy enough... And the Hyde park shots are taken on my gfs camera (a 450D) so weren't the best files to start with. I'll keep to the same method for the time being but I always look out for the crunchies and the stair stepping, usually a slightly larger file upload gets rid of that.
I also agree on your comments Carsten, I've been really struggling with my photography recently. I often take my camera out on my lunch breaks in central London and never actually take the camera out of my bag. I either feel like there is nothing to shoot, or feel embarrassed with there being so many people on their lunch who think I'm some tragic tourist photographing a gate... So yeah I think a bit of my photography has been a bit forced. I need to find a better outlet for my creativity whilst in the city. I'm actually debating taking the site down as it was initially for showing friends and family what I was up to, but my family never look at it and I don't feel comfortable putting photos of my friends and family on an open website anyway. That leaves me with landscapes and the odd Bmx photo which there really is no audience for
I've also noticed with the 100mp particularly, opening the same shot in canon dpp and in LR4 and in LR4 it seems to have nowhere near as much fine detail/sharpness as in dpp at standard raw sharpening.
Can I suggest that you find a university/college campus that has a photography dept and get a public subscription to their library. I do this and its like drowning in chocolate (the web is like day-old coffee). I just found a picture yesterday by Harry Callahan that I'd never seen before and had to bring the book and a hundred others home and stare at the pages in wonder.
Imagine if I'd never seen 'The photojournalist' properly printed http://en.wikipedia.org/wiki/Andreas_Feininger and that if I was hit by a bus it wouldn't have been one of the images that flashed past my eyes
....and then I get all itchy and twitchy and desperate to take more photos.
Toothwalker wrote:
I have not seen much evidence for that statement, if any at all, provided one uses a proper tool for the single-step approach.
I tried various things for years before finally giving up the single-step approach and doing step sharpening. Which tools are you referring to, and do you have an example?
R.Young wrote:
I also agree on your comments Carsten, I've been really struggling with my photography recently. I often take my camera out on my lunch breaks in central London and never actually take the camera out of my bag. I either feel like there is nothing to shoot, or feel embarrassed with there being so many people on their lunch who think I'm some tragic tourist photographing a gate... So yeah I think a bit of my photography has been a bit forced. I need to find a better outlet for my creativity whilst in the city. I'm actually debating taking the site down as it was initially for showing friends and family what I was up to, but my family never look at it and I don't feel comfortable putting photos of my friends and family on an open website anyway. That leaves me with landscapes and the odd Bmx photo which there really is no audience for...Show more →
I feel the same way when I walk around Berlin's busier parts with my D3, and my solution has been to keep the Nikon for more project-oriented photography, and add an Olympus E-PL3 with 12/25/45 kit for casual, more fun photography. This has really freed my mind up a bit, and I am having more fun with casual photography again now. I don't know if it would work for you, but there are lots of small affordable cameras with good enough IQ for casual work. The Fujifilm X10 is another, which a friend of mine uses.
'I have not seen much evidence for that statement, if any at all, provided one uses a proper tool for the single-step approach.'
This is the consensus almost everywhere, web fora, professional Photoshop guidebooks, developers, etc. Not here though, obviously. Each to his/her own, but claims to superior retention of fine detail may come across as gilding the lily.
'Sharpness' should not draw attention to itself, rather the image should be valued for its content and be sufficiently sharp in so doing. An opinion I picked up somewhere, and seems fair enough. Valuing sharpness for its own sake is like drinking sweet sherry by the flagon, or enjoying cartoonish saturation, or any other excess. Balance in all things.
Robert, your work looks fine to me, so there is another data point for you.
Toothwalker wrote:
I have not seen much evidence for that statement, if any at all, provided one uses a proper tool for the single-step approach.
One advantage that I have experienced in some experiments is that step sharpening helps to reduce aliasing/stair steps.
Of course, the intermediate sharpen operations do accentuate the acutance in textures and other fine details, but that doesn't mean that the downsized image is closer to the original -- they are just more visible at reduced size, which is probably the main goal of the method.
carstenw wrote:
I tried various things for years before finally giving up the single-step approach and doing step sharpening. Which tools are you referring to, and do you have an example?
Regardless of the number of steps involved in the downsizing, one needs to find a balance between sharpness and artifacts such as aliasing and ringing. Lanczos3 is generally well received in this regard, but there exist numerous other kernels that are all variations on the theme. If one method yields a more 'crisp' looking image than another, chances are that there are also more artifacts that are mistaken for original detail. To me, the fact that some of the multistep gurus use different scripts for different images is an indication that they steer for 'what looks good' rather than a faithful rendering. That is fine with me, but one should be careful with claims about original detail.
There were examples in a thread last year. Hmm, I can't seem to find it.
I think I remember the thread you mean, but in the end, what I am looking for is detail in the downsized image which looks like the detail in the large image. If it looks like that and has few artifacts, it is great. I don't get this from single-step solutions, but I do get this from multi-step solutions. While I do not know that the detail I get is somehow "original", I don't see any particular reason to doubt that it is, since it is repeatable with the same process and many different images with different properties.
