Page 4 of 4 FirstFirst ... 234
Results 31 to 37 of 37

Thread: Starting my focus calibration of V850 using resolution target

  1. #31
    Steven Ruttenberg's Avatar
    Join Date
    Sep 2017
    Location
    Mesa, AZ
    Posts
    1,080

    Re: Starting my focus calibration of V850 using resolution target

    I made some adjustments, but I had to use the built in Tmax100 profile and set the CI to .8 to get 3.05 from scanner. Yes, it was set to 16bitin and 16 bit out. But, it seems bettet scanning as 48bitrgb and output 48bit rgb. I was referring to line pairs per mm and using a rough estimate I remember from some reading I did a while ago. If it is 200 line pairs per mm, that is incredible. Once I get my hi res target, I can see exactly what scanner is capable of. So far, it seems a bit easier at least to use the scanner profile for tmax than to adjust manually, the downside is you don't get a linear output I think. But I will check this out I as I continue to learn this scanner which does seem quite amazing for 1000 bucks.

  2. #32

    Join Date
    Jul 2016
    Posts
    2,659

    Re: Starting my focus calibration of V850 using resolution target

    Quote Originally Posted by faberryman View Post
    I believe the effective resolution in dpi is just a mathematical calculation based on lpm. Most testing sites use the test target.
    IMHO the testing sites and reviewers use the "effective dpi" jergon because scanners are sold with dpi capability as a main commercial factor, by rating "effective dpi" vs "hardware dpi" consumers may understand better what they were buying.

    Here there is a mathematic calculation for that, used by a known testing site:

    https://www.filmscanner.info/en/Aufloesung.html
    https://web.archive.org/web/20180425...ufloesung.html

  3. #33

    Join Date
    Jul 2016
    Posts
    2,659

    Re: Starting my focus calibration of V850 using resolution target

    Quote Originally Posted by Steven Ruttenberg View Post
    If it is 200 line pairs per mm, that is incredible.
    Steven, let me comment on the TMX datasheet because it has a loose wording. I had been confused with it for some time.

    TMX sports (datasheet, page 8) 200 "lines per mm" at high contrast, and 63 lines per mm at low contrast, while TMY sports 50 at low contrast, "ratings according to a method similar to the one described in ISO 6328".

    I think that they made a mistake in the wording because lines/mm should be line pairs.

    The MTF graph that it's included is shown in cycles/mm, this is a serious wording.

    The MTF graph is the serious way to show resolving power, saying what lp/mm we have at contrast extintion it's easier and also useful, but a MTF graph is what a technician wants to see.

    Here you have the TMX graph:

    Click image for larger version. 

Name:	_MTF.jpg 
Views:	9 
Size:	52.3 KB 
ID:	184681

    It shows that it retains 40% of Modulation Transfer ("contrast") at some 130 cycles/mmm (line pairs/mm = lp/mm = lppmm), beware because the Horizontal scale is not linear, it's logarithmic, it looks 150 but it's not linear.

    So if we have 40% at some 130 LinePairs/mm then it's possible that contrast extintion is at 200 lp/mm for high contrast, because that performance cannot be for low contrast, they only show the graph that's nicer, but common pictorial situations would be ruled by a less nice graph (a lower contrast MTF, say 1:10 instead 1:1000).

    But this performance is only at very high contrast that's relatively uncommon in pictorial photography, you may have that in a silhouette with a very low flare (MC + front compendium shade).

    In that situation (high contrast) in a contrasty edge we have some blur because the big grains in the negative building some level of density ...and on it we have a nitid line made with the fraction of very small crystals that are only exposed with insane amounts of light.

    In the TMX case this is an strong effect because of "linearity instead shoulder", showing that extremly low ISO crystals (very small) are still sensitive to exposure values in the extreme highlights. This is what digital sensors are not doing as nice.

    For this reason film resolving power vary depending on contrast, and that variations depends on grain formulation, this is the relative amounts of crystals of different size and kind, I say "kind" because films may have even different kinds of emulsions overlaped, or different emulsions mixed, each with different natures and sensitizations.

    (We have expose crystals to light that are developed into silver grains...)

  4. #34

    Join Date
    May 2017
    Posts
    168

    Re: Starting my focus calibration of V850 using resolution target

    Quote Originally Posted by Steven Ruttenberg View Post
    I made some adjustments, but I had to use the built in Tmax100 profile and set the CI to .8 to get 3.05 from scanner. Yes, it was set to 16bitin and 16 bit out. But, it seems bettet scanning as 48bitrgb and output 48bit rgb. I was referring to line pairs per mm and using a rough estimate I remember from some reading I did a while ago. If it is 200 line pairs per mm, that is incredible. Once I get my hi res target, I can see exactly what scanner is capable of. So far, it seems a bit easier at least to use the scanner profile for tmax than to adjust manually, the downside is you don't get a linear output I think. But I will check this out I as I continue to learn this scanner which does seem quite amazing for 1000 bucks.
    You not going to able to make any meaningful Dmax tests if you use the negative scanning part of vuescan. Not least because you do have the formula that vuescan used to create the image. The Dmax for these units is quite high, somewhere between 3 and 4, however the amount of noise is not insignificant. (Ken Lee did raw scans in a previous thread of stouffer wedge). i.e. It usually has more than enough range for a negative.

  5. #35

    Join Date
    Jan 2016
    Posts
    29

    Re: Starting my focus calibration of V850 using resolution target

    Quote Originally Posted by Pere Casals View Post
    It's matter of diminishing returns, with the V850, from 3200 you get nothing more in the vertical axis, but you can get a 20% more in the horizontal axis. Beyond 4800 you get near nothing else.

    But working with an oversampled image allows a better conservation of the image quality over the edition (if computer performance vs file size allows it...). For example rotation, sharpening and downsizing algorithms may take advantage of it.

    As Frank Berryman noted a problem can be the additional scanning time, more than file size.
    Thxs to both you and FB.

  6. #36
    Steven Ruttenberg's Avatar
    Join Date
    Sep 2017
    Location
    Mesa, AZ
    Posts
    1,080

    Re: Starting my focus calibration of V850 using resolution target

    Quote Originally Posted by Ted Baker View Post
    You not going to able to make any meaningful Dmax tests if you use the negative scanning part of vuescan. Not least because you do have the formula that vuescan used to create the image. The Dmax for these units is quite high, somewhere between 3 and 4, however the amount of noise is not insignificant. (Ken Lee did raw scans in a previous thread of stouffer wedge). i.e. It usually has more than enough range for a negative.
    I did get it to show slightly better than 3.05, but had to use the built in Tmax100 set to a CI of.8. I was abke ri know it is 3.05 because that is how dense my step cjart is and I could image that final step wirh some head room to spare.

    When doing everything manual since I want a linear scan output, it seems to have trouble no matter the settings. This concerns me because maybe I am not getting as much out of negative as I should be. I am checking my ptocess to see if doing something incorrect with vuescan.

    I am going to try a positive scan, but that beats the purpose of bw neg scanning and doesn't provide the required output.

  7. #37
    Steven Ruttenberg's Avatar
    Join Date
    Sep 2017
    Location
    Mesa, AZ
    Posts
    1,080

    Re: Starting my focus calibration of V850 using resolution target

    Okay,
    I think I have this figured out.

    1. When you make the scan of the step wedge, it will not look very pretty, Here I made zero adjustments to enhance the image before I saved it.
    2. I tried three different ways of scanning and saving
    a. Color slide, 48bitrgb output to 48rgb
    i. Brought into photoshop and desaturated
    ii. Used Colorperfect to convert the image (of note, this required the least amount of adjustment in Colorperfect. Had to tweak the "film" gamma setting a bit to a lower value (about .75 to .9, don't remember exactly, but it did need some adjustment) then minor adjustments on the other sliders
    iii. Minor curves adjustment in PS

    b. B/W negative, 48bitrgb, output 16bit gray
    i. Brought into Photoshop then into Colorperfect (no adjustments made in Photoshop before conversion)
    ii. The "film" gamma in Colorpefect required a huge change (default is 1, had to change to .380), made minor adjustments on other sliders
    iii. Minor curves adjustment in PS

    c. B/W negative, 16bit gray, 16bit gray
    i. Brought into Photoshop then into Colorperfect (no adjustments made in Photoshop before conversion)
    ii. The "film" gamma in Colorpefect required a huge change (default is 1, had to change to approximately .380), made minor adjustments on other sliders
    iii. Minor curves adjustment in PS

    b. B/W negative, 48bitrgb, output 48bit gray
    i. Brought into Photoshop then into Colorperfect (no adjustments made in Photoshop before conversion)
    ii. The "film" gamma in Colorpefect required a huge change (default is 1, had to change to .380), made minor adjustments on other sliders
    iii. Minor curves adjustment in PS

    I get similar results for each setting, but setting to 48bit rgb color slide and saving as 48 bit rgb and then desaturating in PS prior to going into Colorperfect appeared to provide the cleanest file, that required the least amount of adjustments. This suggests, (I say suggests, not requires) that there may be benefit to scanning bw negatives as a color transparency and then desaturating before converions to bw when using Colorperfect. There are various other options one could use for sure, but the initial conversion from negative to positive is the most important step as we all know. The advantage I see of using color transparency for scan setting is that it is scanning and recording the negative exactly as we see it, no intermediate steps to scan and create the negative.

    However, whatever is ones preferred method, if it works, then stick with it.

    My next step will be to apply this same 4 settings on an actual image and see what i get.

    Note, when using the color transparency option, I did not assign any preset film recipes to it. So, it should be a linear raw tif file. It certainly behaves that way however, I do not have a means to actually check the saved file to see if it is truly a linear scan, but it did not seem to matter and it does give the cleanest file.

    As you can see from the images, the scanner does have better than 3.05D (which is the densest the step wedge I have goes. There is a bit more that could be shown if I had a larger step wedge range. Next time.

    For now, I think I understand my scanner and the basics it needs to be used with. I also seem to have similar PS curve settings for the entire image that appear similar to the "film" curves. Whether or not this means anything I do not know. In the end, as long as you are producing the images with the details you want, it doesn't matter how you get there as long as you get there

    https://www.flickr.com/photos/157376...57698054054940

    One item I noticed in these target scans, is a reflection I think, this might have been from scanning during the day. I have decided that when I scan from now on to cover the scanner with a towel to block any stray light either from a window or from the overhead lights in my office. It doesn't put out any heat so to speak so it is not an issue. Plus it will keep the reflections away due to stray light. Not going to redo the targets though, the stray light did not affect the scan results and I do stand by them.

    Another item I noticed, and I think Ken Lee tried to explain to me, but it didn't click right away is that whether or not you are scanning with input as slide or bw (I use slide and 48 bit as you get all three channels and can choose the one you want in PS if you want or just convert all channels to 16 bit gray. Doesn't seem to affect the linearity of the file.), if you set the bit depth to 16 bit gray, you can either let Vuescan choose which color channel to use or you can choose it yourself. I would make prescans for each channel and small output files and see which gives best result then make final scan using that color channel.

    Another item I am learning now, is to make several smallish scans (again, I use 48 bit in and out as color slide) and vary the red, green and blue analog gains. It seems to make some difference, although I could be wrong and then choose the setting you like best. Right now, I vary all three equally.
    Last edited by Steven Ruttenberg; 24-Nov-2018 at 20:34.

Similar Threads

  1. Does anyone know of a 4x5 resolution target?
    By Steven Ruttenberg in forum Digital Processing
    Replies: 14
    Last Post: 15-Sep-2018, 11:18
  2. DIY resolution target
    By SURF in forum LF DIY (Do It Yourself)
    Replies: 13
    Last Post: 19-Nov-2012, 05:43
  3. Reflective Target Calibration for Scanners
    By Jay Decker in forum Digital Processing
    Replies: 7
    Last Post: 21-Mar-2011, 13:12
  4. Resolution Target - Air Force PDF File?
    By Richard Rankin in forum Resources
    Replies: 4
    Last Post: 27-Dec-2000, 22:20

Bookmarks

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •