LaserSoft Imaging

I can shed some light in the area of batch scanning.

I use SF Nikon Ai v6.0.2 to batch scan negatives to 48-bit HDR tiffs. The current version of Ai occasionally has problems with the frame offset but otherwise it's ok. Using Job Manager with SF Nikon Ai though is useless for batch processing because;

(a) the default frame position and dimensions as specified by SF are incorrect for the Nikon 4000ED.

(b) there is *no* way to select a saved Frame Set dimension during import of the images into the Job Manager queue, or to copy and paste frame positions/dimensions between images in the queue. Lasersoft says that it's not possible.

(c) there is *no* way to select a saved Frame Set while editing images in the Job Manager.

(d) there is *no* way to copy and paste frame positions/dimensions between images in the Job Manager queue.

The result is that you have to preview every frame (ouch!) in the Job Manager queue and draw/position the frame manually. It couldn't get any slower than this.


Consequently, I scan my negatives to HDR format using SF Nikon Ai's Batch Mode. After each roll of film has been scanned, I view the images using VLT to check that the frame offsets were ok. If any film borders are visible within the images, I rescan the affected images.

I then process the HDR images using SF HDR. Unfortunately, there is no Batch Mode processing in HDR. You cannot simply select a folder of images and tell SF HDR to batch process them. You *must* use Job Manager. With the SF HDR 6.0.2r13, there is a bug which causes the output dimensions of the selected Setting to be applied incorrectly to the imported images usually producing negative output dimensions. You must therefore edit each and every image in the queue to correct the output dimensions.


The SF Nikon Ai batch scan to SF HDR process routine works but it could be far more efficient. Lasersoft should add the ability to copy/paste frame positions/dimensions in the Job Manager. They should also add the ability to select saved Frame Sets while editing in the Job Manager, and they should allow Batch Mode processing of folders of images with SF HDR. Only then would they have a truly efficient solution for batch scanning of larger numbers of negative images.

I have been requesting these functions for almost a year without any apparent response. I seem to be the only person using SF to batch scan rolls and rolls (or strips as the case may be) of negative film with the goal of building an archive of both the 48-bit HDR tiff images and the corresponding 24-bit screen size jpg images.

(If VLT could catalog the screen size images and link them to the HDR tiffs on the local hard disk or on removable discs/disks for easy access when re-scanning is required, that would definitely be a nice function.)

regards
Gregory

Gregory C wrote:I seem to be the only person using SF to batch scan rolls and rolls (or strips as the case may be) of negative film with the goal of building an archive of both the 48-bit HDR tiff images and the corresponding 24-bit screen size jpg images.

...well make that two of us!

Thank you for your reply.

If, as I heard from the USA sales department, DCPro will soon support 48bit NEF (Nikon's own raw scan format), wouldn't this solve the problem? One would batch scan in Nikon Scan set to NEF and process in DCPro using Job Manager.

Otherwise, we seem to be left with the problems you have identified (I am a new user and do not have first hand knowledge). Do I understand you correctly to say the basic issue for large quantities of negative scans is 1) frame offset in AI 2) frame offset in HDR?

Your workaround is

1) Batch mode with AI (vs. JM).

This does not strike me as a significant issue, because I do not see JM to be any faster than batch mode. True? You go on to discuss VLT review of each roll/strip to look at film borders. Are you saying that even with a prescan of the 1st image in a strip, you will get frame offset problems in that strip? If so, that stinks!

2) HDR using JM for post scan. Again, on the surface this is fine. My first pass post scan processing would be for slideshow and e-mail, etc use of lower resolution copies. I would need to rotate and crop (if offset is off) and do auto correction and NegaFix. I assume these things work well with HDR JM. True? I would then likely manage/name/sort etc this database of lower resolution copies in Photoshop Album (or whatever is the best management software).

At this point I would have my raw 48bit archive AND my working copy, both in the appropriate form and with appropriate management for intended use.

Finally, for those images I want to print (to higher standards) or otherwise perform custom processing, I would go back to the raw file and work from that. In my case, this will be less often. I would locate the raw file by finding the copy in PS Album, getting the original name/roll and quickly locating the original raw scan.

I think you are saying there is a problem in HDR that will make the above workflow harder than I have described. Where does it fall down?

Sorry for all of the questions...it?s just you seem to know something about this and I am trying to avoid scanning 5,000 negatives more than once

Thanks

--Caleb

First, a little explanation of my terminology.

Frame offset
The position of the film frame within the scanning area. If we're lucky, the film frame will normally be centred within the scanning area. Sometimes though, the frame will be off-centre, occasionally with part of the frame actually outside of the scanning area. SF has a Frame Offset tool which we can use to tell the scanner where the beginning or end of the frame is. Using this tool allows the scanner to reposition the film accordingly. With SF, this tool affects the position of every frame in the current strip.

Until recently, the Frame Offset tool was very unreliable. After a little investigation, Lasersoft found a bug in the Nikon-provided MAID libraries (ie, not Lasersoft's fault). The Frame Offset tool in the current versions of SF Ai works much more reliably and is actually useful.

On the down side, it can be difficult to reset the frame positioning after ejecting one strip of film and loading another. I have found that I usually have to: (a) load the film in the scanner, (b) open and refresh the Overview in SF, (c) prescan the last frame, (d) refresh the Overview again and perhaps (e) prescan the first frame. Only then will the frame positions be reasonably accurate and only then will the thumbnails in the Overview reflect the actual positioning of the frames when scanned. I always regret forgetting to tell the film processing people to not cut the film into strips. I own a roll adapter. Scanning a single roll of film is much quicker than loading and scanning 9 or 10 strips of 4-frame film!

I have heard that the Job Manager can be used to automate the scanning of multiple images across multiple strips of film, allowing the user to eject and load film during the job. I have not been able to do this though so I have no idea if it actually works.


Frame position/dimensions
Within the scanning area (ie, the preview), we can select an area to scan to disk. The dotted line which depicts the area we have selected is called a Frame. We can save the frame's x,y position and dimensions via the Frame-Set popup menu under the General tab of the main control window. We can also save multiple frames in a single Frame-Set (which might explain why SF does not allow us to select Frame Sets when editing images in the Job Manager where each image in the queue can only contain a single frame).

Under the Frame tab in the main control panel, we can save exposure settings with the Settings popup menu. These settings do not include the frame's position and dimensions though. Nor do they include the frame's name or the output dimensions. The Setting remembers the Output Scale parameter; not the Output Dimensions. There is currently no way to specify absolute output dimensions for images of various original sizes. NikonScan is much more versatile in this area.

Question for Lasersoft
When I change the Picture Settings Exposure or White Balance values, the Settings popup menu changes from the current saved Setting name to "Save" indicating that the Settings also remember the Picture Settings parameters. If I set exposure to -0.5 and save the Setting, and then use this setting when importing images to the Job Manager, the exposure setting is not reset to the saved value of -0.5 when I load the image for editing. What's happening? Are the Picture Settings controls meant to be an indirect interface for other controls within SF such as the Global Color Correction control? Which controls are affected by the Picture Settings controls and in what way?

If, as I heard from the USA sales department, DCPro will soon support 48bit NEF (Nikon's own raw scan format), wouldn't this solve the problem?

The main problem with using SF Ai to batch scan rolls/strips of negative film is Frame Offset. Frame Offset though is controlled by the Nikon scanner's firmware and by the Nikon MAID libraries when using both NikonScan and SF. The frame offset problems could therefore exist in both programs. Any differences consequently would be in the way the user selects the images to scan. With both programs, you need to make your selection via the thumbnail view (Thumbnails in NikonScan. Overview in SF). With both programs, it is not necessary to refresh the thumbnails before making the selection although it is preferred to improve your chances of getting accurate frame positioning and avoid what I call "offset drift".

Some people seem to think that NikonScan allows the user to save true raw data at the highest bit depth allowed by the scanner, while SF loses data when saving the raw output. I'm not an expert and have absolutely no idea whether or not this is true. I also have no knowledge of how to test this so I'll leave this argument to other more knowledgable people in this forum.

All other things being equal, if you only wish to scan the images to disk in raw format, you might be better off using NikonScan because it doesn't cost you anything. I already own SF Ai (Nikon) so I'll continue to use it. Two important things to note:

(a) SF installs a version of the Nikon MAID libraries where the Frame Offset tool bug has been fixed. NikonScan still installs the old buggy Nikon MAID libraries.

(b) If SF HDR cannot open the NEF files, you will need to use the NikonView software to open and edit the NEF files. NikonView is far from the power and flexibility of SF HDR. SF HDR can also open raw 48-bit HDR images very quickly using embedded previews in the raw files that SF Ai creates when saving the raw images to disk.

Question for Lasersoft: Is this feature only possible when the HDR images are saved in tiff format? jpeg HDR images seem to open slower than tiff HDR images.

Do I understand you correctly to say the basic issue for large quantities of negative scans is 1) frame offset in SF Ai?

Frame offset can be a problem with SF Ai when scanning the original film but it can also be a problem with NikonScan. The Nikon MAID libraries use the same code.

2) frame offset in SF HDR?

The problem in SF HDR is frame positioning and size; not "offset". However, to be fair, it's only a problem if you want to select less than the whole image because the Job Manager's default is to select the whole image.

I mentioned before that the saved Settings do not save absolute output dimensions. This though is also not a big problem with Job Manager because you can fortunately edit the output dimensions of one image and then copy/paste those dimensions to every other image in the queue quite easily.

It would therefore seem that I was a little harse in my previous message but my requests that Frame-Sets can be selected, and frame positions/dimensions copied/pasted within the Job Manager remain valid.

1) Batch Mode with Ai (vs. JM).
This does not strike me as a significant issue, because I do not see JM to be any faster than batch mode. True?

Batch Mode in Ai is much faster and more efficient than JM.

You go on to discuss VLT review of each roll/strip to look at film borders. Are you saying that even with a prescan of the 1st image in a strip, you will get frame offset problems in that strip? If so, that stinks!

Unfortunately, it's true. As I explained above though, it's not a Lasersoft fault. It's a Nikon fault. Performing the procedure I outlined above (Overview, prescan, overview, prescan) can help a lot though. As an extra hint, I try to remember to turn off Auto-Exposure and Auto-Focus before performing prescans. This dramatically cuts the amount of time required to perform a prescan; possibly by as much as 50%.

2) HDR using JM for post scan. Again, on the surface this is fine. My first pass post scan processing would be for slideshow and e-mail, etc use of lower resolution copies. I would need to rotate and crop (if offset is off) and do auto correction and NegaFix. I assume these things work well with HDR JM. True?

True. All of these things work very well within JM. Apart from Negafix and cropping (ie, frame position/size), all of these settings and others can be copied and pasted between images in the JM queue too; a big time saver, obviating the need to preview every image in the queue that needs adjustment.

I would then likely manage/name/sort etc this database of lower resolution copies in Photoshop Album (or whatever is the best management software).

Ahh. You're on Windows. I've heard good things about Photoshop Album. It will never be available for the Mac but we have iPhoto and the newly released VLT version of SF HDR to work with. We also have countless shareware offerings and then there's also the venerable iView MediaPro which I have yet to test.

At this point I would have my raw 48bit archive AND my working copy, both in the appropriate form and with appropriate management for intended use.

The holy grail for those of us trying to build an archive and scan every strip of film only once. I'm going to be saving my HDR images to DVD-RW. What are you planning to save your raw images to?

I would locate the raw file by finding the copy in PS Album, getting the original name/roll and quickly locating the original raw scan.

Every time I try to conceive a naming system for my film and the corresponding images, I get a headache. It occurs to me that possibly the best naming method would be to use "shoot date #serial number" (eg, "20030428 #01", "20030428 #02", etc) and rely on keywords within the cataloging software to locate specific photographs.

I think you are saying there is a problem in HDR that will make the above workflow harder than I have described.

I was probably too harsh on the software in my previous message. All things considered, the whole process would probably be relatively simple as long as you do not need to draw selection frames within Job Manager images.

Overall, SilverFast is an excellent piece of software and probably the best scanning software currently available. There are just a few places where it could use a little improvement.

regards
Gregory

Gregory:

Thank you again for taking the time/effort to write such a full/clear reply!

This problem of frame offset is a real pain!! I tested a 4 frame strip and in BOTH AI AND Nikon Scan. Images 1,3,4 were fine. Image 2 was offset to the point where I saw part of the adjacent frame!!!!! Boo Hiss!!!! Interestingly, v**s**n seemed to get ALL frames fine!! If I cannot solve this issue, I think it will be a non-starter for me in using either SF AI or Nikon Scan. I am hoping to scan WITHOUT ANY prescan/overview. This is sounding less and less realistic, unless v**s**n proves to be sufficiently reliable?Now on a subsequent batch scan of the same strip in v**s**n, image 1 is offset?double boo hiss! Clearly this offset problem is significant.

With both programs, you need to make your selection via the thumbnail view (Thumbnails in NikonScan. Overview in SF).

In NikonScan you can select frames without even doing a thumbnail. I was hoping to do this.

SF installs a version of the Nikon MAID libraries where the Frame Offset tool bug has been fixed. NikonScan still installs the old buggy Nikon MAID libraries

Is there any talk of Nikon addressing/fixing this issue?

[quote/]If SF HDR cannot open the NEF files, you will need to use the NikonView software to open and edit the NEF files. NikonView is far from the power and flexibility of SF HDR. SF HDR can also open raw 48-bit HDR images very quickly using embedded previews in the raw files that SF Ai creates when saving the raw images to disk. [/quote]

SF new version of DCPro will handle 48bit NEF and do all post processing that HDR does?.

President_LSI wrote:<b>HiRePP, SilverFast DCPro and NEF-Format</b>

Yes you can add HiRepp to raw scans from other manufacturers by passing files through SilverFast HDR or DCPro.

Yes, SilverFast DCPro will also support the Nikon NEF format.

Yes, post-pocessing with DCPro will be the same as with HDR.

SF The holy grail for those of us trying to build an archive and scan every strip of film only once. I'm going to be saving my HDR images to DVD-RW. What are you planning to save your raw images to?

I have LOTS of hard drive space, several terabytes in RAID arrays. For my raw scans, I will start using this and backup or move to DVD. Hard drives are about $1/gig now, or less.

Every time I try to conceive a naming system for my film and the corresponding images, I get a headache. It occurs to me that possibly the best naming method would be to use "shoot date #serial number" (e.g., "20030428 #01", "20030428 #02", etc) and rely on keywords within the cataloging software to locate specific photographs.

I keep each raw roll in a separate folder (within a single directory). I will manage/label/sort/find my lower resolution images in Photoshop Album. These will have the same name or folder name as the raw file (but in a separate directory). PS Album will show me the file and its folder name. I will then find the raw file/folder with a quick find command on explorer?.at least that?s the concept.


All of this said, my primary concerns are:
1) Figuring out if there is a benefit to NEF vs. v**s**n vs. AI for the archive quality of the raw file.
2) Figuring out if there are compatibility issues if I were to use non-SF for scanning raw, and then use SF HDR/DCPro for processing and bulk processing., i.e. if I conclude I should use v**s**n or NEF to scan, will I get a mis-match of levels/settings/etc when I later use those raw scans in HDR/DCPro. If there is a mis-match from this workflow, it may outweigh the possible benefits of the other scan software.
3) Figuring out how to deal with this offset issue, to achieve an efficient process in AS or Nikon Scan.

?the search continues!

Thanks again!

--Caleb

All of this said, my primary concerns are:
1) Figuring out if there is a benefit to NEF vs. v**s**n vs. Ai for the archive quality of the raw file.

I believe NEF contains scanner data.
v**s**n's raw files are saved as ordinary tiffs although v**s**n may offer compressed tiff as a choice of saved formats. SilverFast does not.
SilverFast can save the raw files in either tiff or jpeg. You wouldn't be interested in the jpeg format but other storage space conscious people might be. SilverFast's HDR format contains the preview that allows large files to be loaded and previewed quickly in SilverFast HDR. Lasersoft's demonstration of the difference in speed when opening either an HDR file or a tiff file uses a very large file; 200+ MB. Most of my 35mm negative uncompressed tiffs are around 125MB. The speed difference in opening these as HDR or as plain tiff might not be so significant.

You'll have to think over the pros and cons yourself and make a decision. I sometimes worry about the future compatibility and accessibility of the various formats though. tiff is definitely going to be around for a while. I don't know about NEF.

2) Figuring out if there are compatibility issues if I were to use non-SF for scanning raw, and then use SF HDR/DCPro for processing and bulk processing., i.e. if I conclude I should use v**s**n or NEF to scan, will I get a mis-match of levels/settings/etc when I later use those raw scans in HDR/DCPro. If there is a mis-match from this workflow, it may outweigh the possible benefits of the other scan software.

Lasersoft will have to answer this query. I have no relevant knowledge.

3) Figuring out how to deal with this offset issue, to achieve an efficient process in AS or Nikon Scan.

My immediate answer would be to get a scanner which is not a Nikon. Between this issue and the incredibly non-durable nature of their APS adapter (the plastic cogs within my adapter has been stripped during normal use at least 3 times in the past year), I wonder if Nikon really cares about the quality of their scanners. Of course, if you already own the Nikon scanner, then you (and I) will just have to make the most of it and work around the offset problem.

regards
Gregory

Gregory, you wrote the following:

On the down side, it can be difficult to reset the frame positioning after ejecting one strip of film and loading another. I have found that I usually have to: (a) load the film in the scanner, (b) open and refresh the Overview in SF, (c) prescan the last frame, (d) refresh the Overview again and perhaps (e) prescan the first frame. Only then will the frame positions be reasonably accurate and only then will the thumbnails in the Overview reflect the actual positioning of the frames when scanned. I always regret forgetting to tell the film processing people to not cut the film into strips. I own a roll adapter. Scanning a single roll of film is much quicker than loading and scanning 9 or 10 strips of 4-frame film!

I'm working with roll-adapter for my 35-mm film.
I did a, b, c. - everything is ok.
What do you mean under refresh the Overview again (d)?
Do you mean to press green play button to reload thumbnails?
When I do this, I always get the same result as after b. No difference at all.
Thanks

when working with film strips, the offset is atrociously unreliable. I sometimes throw in the extra Refresh Thumbnails in an attempt to cure the offset problem. it usually doesn't work.

with full rolls of film, the film offset seems to be marginally better, and I won't do two thumbnail refreshes if everything seems ok.

are you having any offset problems? that is, when you load your film, is every frame positioned accurately automatically?

I have the problem with the entire film.
The first image on my film is ok, but the next one is already shifted to the right. Some times I even see the rest of the prevous image on the next one. On 99% of my images I see a black strip on the left side of my image while missing a littke on my right side.

that's what a lot of us are seeing with the Nikon scanners. not nice is it?

I've tried scads of different techniques to solve this problem but with no success. add to that the random nature of the Film Offset tool and scanning film with a Nikon scanner can cause some serious frustration.