How to you guys have the energy to keep this going?
Sent from my iPhone using Tapatalk
How to you guys have the energy to keep this going?
Sent from my iPhone using Tapatalk
We never give a bone
But right now this is quite funny because there is smashing evidence that low performance of the Epson is a urban legend coined by people not knowing how to focus it and having little edition skills, reiterating it: smahing evidence https://www.largeformatphotography.i...=1#post1557267
Unfortunately, I hadn't realized Alan's comparison was a bit unfair-- I knew the Epson looked softer than my scans (and I have the v800, not the v850), so I looked at it again in more detail.
The Howtek image is at 106%, and the Epson image is at 159% zoom. I know why-- to show the same crop at the same size. Unfortunately, 150% zoom will always look softer than 100%, unless you're using some very smart software. It explains why the Epson sharpened up so much easier, as well. So what this really shows is that 4000 PPI is sharper than 2400 PPI, which I don't think anyone's going to argue with.
If I claimed that 4x5" has more resolution than 6x7cm, I don't think anyone would find that surprising, or claim that 6x7 is a terrible format.
Downsampling the Howtek scan to 2400 DPI, and comparing them at 100% would be more balanced, but should still give a noticeable edge to the Howtek. Shouldn't leave the Epson looking quite so dismal, however.
The Epson is also scanned "flatter" in terms of overall contrast-- some people prefer this, because it makes it easier to edit in lightroom. Others would rather have the scanner do all the work. Personally, I make sure there's no clipping at either end, but otherwise leave the histogram alone for the most part. I did, once I loaded it into Affinity, assign the "Epson Gray Gamma 1.8" profile, and that seemed to give a more natural tonal range on my test image.
Not at all grat...
the dynamic range of the negative surpasses the monitor capability, so a pro edition requires to take all the histogram and compressing what necessary in the curve edition, in Pro edition you always start with a 16/bits channel dull and flat image, if not you don't keep your choices open.
Please see this: https://www.largeformatphotography.i...=1#post1557499
If you want to know, use the histogram equalization feature in Ps with both images, you'll find matching results, this demonstrates that the images have the same potential to get exactly the same tonality. This is about editing to your taste, not about accepting a canned automatic result.
Also I don't agree to that.
First we are pixel-peeping at x20 and at x40 in the monitor, if we are to enlarge to that then the image quality in the negative is what it fails. Observe the image at x10 which is a very large factor and you won't find a difference. Most important is using "bicubic, ideal for reductions" in the image downsizing to edtion size, etc..
Regarding the Epson image capability to be sharpened this is because it has not been well optimized internally in the machine. Specially in the past, when drums were manufactured, computers had only a fraction of today's RAM, CPU speed and had no SSD or M.2, so a Pro machine had to deliver a very well optimized image to not overload the entire processing chain, so Pro machines do that very well, with the Epson you have to scan 6400 is wanting 2900 effective, fortunately modern computers can easily deal with that, but in the past those Epson inefficient files were a heavy burden.
Anyway, if having a doubt ask for the original 16bit/ch files and compare.
The truth is that we should be pixel-peeping at x20 or x40 to see if a difference is there, but practical difference to craft a sound 1m print is zero, even if inspected with the nose on it.
Well done. You gave a technical reason for what I stated.
I never said it was bad, merely that it was a common choice when scanning.
That was implied, but apparently I should have been explicit: The flatter range from the Epson scan is easier to edit in lightroom/PS/affinity. The Howtek scan is more "ready to go" and will require (and tolerate) less editing.If you want to know, use the histogram equalization feature in Ps with both images, you'll find matching results, this demonstrates that the images have the same potential to get exactly the same tonality. This is about editing to your taste, not about accepting a canned automatic result.
The truth is that pixel peeping at 100% or 200% or 2000%, is largely a waste of time-- The real answer is, how easy is the image to tune to the preferences of the photographer, and does it produce an acceptable end product for the photographer's purpose? As you say, the practical difference to printing is zero, so why get excited about it?The truth is that we should be pixel-peeping at x20 or x40 to see if a difference is there, but practical difference to craft a sound 1m print is zero, even if inspected with the nose on it.
Can I, using <method X>, produce an image which is acceptable for sharing on the web? For display on a 4K (or 8k) projection system? For printing, to be hung on a wall?
Anything beyond that is semantics and what-iffing. For what I do, as an enthusiastic amateur with a day job, the Epson is slightly overkill. If I were displaying my photographs in an art gallery, or selling them, it might not be sufficient-- although a number of professional photographers do sell prints made from Epson scans.
They also spend a great deal more time working on the images from the scan to the edit, to the proofing, to the printing-- watching a professional making curve edits to match the gamut of a particular brand and SKU of paper prior to printing is enlightening-- although, I suppose, really, they're endarkening the paper, not enlightening it.
Goes with the territory, 'fraid to say. No matter what gamma etc you use when you scan it, the file off an Epson with the histogram maxed is noticeably flatter than a scan off a high end scanner, even when the latter has had the histogram maxed and a curve added to open up what will become highlights after inversion - this is for scanning negatives for manual inversion in Photoshop. The low contrast you see on Epson scans (without improvement in ability to see deeper into denser parts of the neg) is largely down to the shortcomings of the Epson's optical system.
False, the lenses in the Epson optical system resolve 17,000 effective pixels in the scan row in a single pass, point another flatbed scanner able of that.
Also riduculous, the Epson has top prestige for the inversions it delivers, see min 22:00 here: https://www.youtube.com/watch?v=Q9d8BukUgzI&t=1684s
Also a ridiculous and absurd statement, you edit the Epson curve and you get the tonality you want, here the Epson edited for a less flat image than the Howtek:
Take the Howtek scan and the Epson scan, edit curves and if you are not a total dummy you will match the tonalities with a few clicks. In this thread you have the Howtek scan made by Peter Finge and the Epson one, first try to match the images on your own, you'll find it easy, if you are not able I'll show you the curve editions that are to match the images, I'll teach you how to do it.
Even at micro level tonality is matched, beyond resolving power:
Attachment 204938
How good is that Epson !!! totally matching a Howtek drum even when pixel-peeping at x20 !!!!
I don't know, maybe you need to calibrate your monitor. Certainly "if you are not a total dummy you will match the tonalities with a few clicks" is an interesting statement accompanied by such an obvious difference in different parts of the image (also clear in the crop).
Thanks Alan for providing the comparison. As a flatbed user, I always wonder what a drum scan would do for me. In my case, I usually print my 4x5 stuff as 16x20 on inkjet, and scan resolution of 2400ppi seems quite adequate. I don’t feel I am missing much that I would be able to see.
Bookmarks