APPENDIX A: FINAL REPORT OF OPERATING TECHNICIAN
UIUC/KODAK DIGITAL IMAGING PROJECT
FEBRUARY 15, 1996
SUMMARY OF WORK ORDERS
The 100 slides provided by Architecture Professor Richard Betts for the first imaging order were a challenge and proved to be good preparation for learning the capabilities of the Kodak equipment. It was quickly proven that the major limiting factor on the quality of the digitally captured image was the condition of the original. A number of the slides were ten to twenty years old, and suffered from wear, dirt and dust ingrained on the emulsion, and general deterioration. A second problem was that a number of the images were black and white, but the resulting color slide gave them a yellowish tint, which the older software Kodak provided for the PIW could not correct. This is something that will have to be done using Photoshop or a similar computer application that can render them in grayscale. In addition there were two slides, one with a decidedly reddish tint and the other greenish, which the software could adjust to a more monochromatic state. There were also problems with the lighting conditions under which a number of the slides were taken, which had to be adjusted according to their being too light or too dark. A final, technical problem was that two of the slides were mounted backwards, so that the image was reversed. These were re-scanned after the slide was turned around.
The resulting digital images recorded on the Photo CD were as good as could be expected with the limitations of the software and the condition of the slides. Dr. Betts was pleased with the results, especially since we were able to provide more personal attention to adjusting the quality of each scanned image than a commercial firm would have done.
There were far fewer problems with the 114 slides from the Motley Collection provided by the Rare Book and Special Collections Library. Several images were either too faint or dark, which was due primarily to the condition of the original drawing. This could be corrected with the density adjustment feature on the PIW. The only other problem was that in a number of cases the frame of the slide obscured the target that was placed below the object during the filming process. In all of these cases the image of the object was not affected, and was reproduced in its entirety, with no interference from the slide frame.
The scanned images were almost without exception of high quality in terms of color, contrast, and sharpness, with very little touching up work needing to be done from a computer photo application.
The 186 images generated from the collections of the Krannert Art Museum were of even higher quality, thanks to the physical condition of the slides provided, and the professional conditions under which they were taken. Although two slides were slightly dark, the only factor affecting the fine quality of the scanned images was a glare on the objects caused by the lighting used during the filming. This only affected a small portion of each image, and can be corrected using Photoshop.
The fourth order was for 174 slides of toxic plants from the Veterinary Medicine Library. They required only minor corrections for density on three images that were overexposed. It was recommended that changes to the quality of the images not be made so that the color of the plants was not altered. The scanning work went very quickly, with less than six hours total needed to complete the digitizing part of the process.
Project Number five was for 132 slides of the African Art Gallery in the Krannert Art Museum. The slides were divided into eleven sets of twelve slides, taken at eleven different sites in the gallery and showing a 360 degree view from each one. An adjustment had to be made while scanning because of the yellowish tint to each slide caused by the lighting in the gallery. The PIW was set so that blue was increased at the expense of the yellow, to produce a more natural color. The contrast was not changed, so the original exposure was maintained. The result is that while the white walls of the gallery are properly illuminated with the reflected light, the objects in the center of the room appear darker due to the lack of reflective surfaces and the dark wood floor. Any objects seen from a distance are underexposed.
In viewing the index prints it does not appear that the compensation for the yellowish tint was enough to overcome the distortion caused by the artificial lighting. The PIW seems to be able to produce excellent scanned images with their true color more readily using natural light than it does with artificial. The major reason for the problem in judging how to adjust the images is that the monitor does not produce a resolution high enough to get a true representation of the digital image, which can only be seen by creating a photographic quality print. Because of this, any adjustments have to be estimated, and the slide scanned and print created before the true quality of the image is determined. This can result in some wasted time until the proper setting for creating an accurate scan is found.
The sixth project completed was a set of 962 slides of diseases and disorders of the horse, which was done for the Veterinary Medicine Library. These slides were organized into six units, and a booklet with reproductions of all the images accompanied each set. This allowed a comparison between the scanned image on the monitor and the printed version, so that any adjustments to the image could be made based on this and not sole reliance on viewing the monitor image. There were only a small number of slides that required any adjustments (38 of the 962, or four percent), and most of these were due to the slide being too dark or having a bluish tint. The work on digitizing the images went very rapidly because of the quality of the slides, reaching a top rate of 50 slides per hour, which should be close to the maximum possible with the PIW as it is currently configured, with no slide carousel or other means of feeding through a large number of slides at one time.
One final project involved using 35mm roll film to produce 37 images of historical maps from the Map and Geography Library. The scanning involved using the manual gate because two exposures were filmed of each map, so only one was actually scanned. Extra care had to be taken to insure that each exposure was free of dust, since even small specks appeared as bright white spots on the digital image. The images produced were of excellent quality, and even at a higher level of magnification they appeared sharp, clear, and highly legible.
A problem arose with the printer when attempting to create an index print. As the information was being loaded into it the printer stopped, and gave a system error 251 message and the command to power down and then restart. This was done but the problem persisted. A system shutdown was then tried, and during the bootup process the printer was not recognized by the system.
The problem was corrected by checking the printer diagnostics and doing a reset. An internal test print was run to see if the printer was now functioning properly, and it successfully printed the test. The index print was attempted again, and this time the process did work.
Several more prints were made before the problem resumed. From this point on four or five attempts were made, consisting of powering down and resetting, for each print produced. There was still a problem with the printer buffer not having the capacity to hold the data for the index print. A similar problem occurred that was written up in the Kodak newsletter, but this involved the system having a default setting to automatically make a print, and the printer being manually turned off to prevent this. The files were still created, leading to the hard disk filling up and the system shutting down. This was not the case here, however, since the default was set for no index prints to be made.
A Kodak service representative came to look at the printer, and he made some minor adjustments. After that visit a number of prints were made with no problem, although it did occur on two more occasions when the final set of images was being scanned. Powering down once was all that was required before the index print was successfully printed. This problem with the printer is something that Kodak should look into, because a great deal of time was taken up in trying to get it to work properly.
A second problem developed with the printer during those times that the file for the index print would actually load in to the printer and not cause a shutdown. In these circumstances the error message would say "Ribbon Advance Error - Code 144, Yellow/Clear." An Error Code 144 is a failure to find the yellow or clear ribbon. A look at the ribbon showed the yellow area, but there was no indication as to why the printer could not find the yellow ribbon, or what corrective measures should be taken. This problem occurred suddenly after a number of successful prints.
It was finally decided that the problem involved the ribbon and not a mechanical fault of the printer, so the ribbon was removed and a new one installed. A print was then made successfully, which did show that the printer was fine. The Kodak service representative said that the error message basically meant that the ribbon was out and needed to be replaced. He offered no explanation as to why there is no specific message to inform someone that the ribbon is out. This would seem to be a logical and obvious thing to do. A lot of time would have been saved if the printer had simply stated "Ribbon out, please replace." The other disturbing aspect of this is that the people at Kodak on the help line had no idea what the error message meant, even though it should have come up many times over the last several years whenever a customer had a ribbon had run out and did not know what to do, given the inadequacy of the error message. The people on the help line should have had a reference for it.
A brief problem occurred with the scanner when one slide was being scanned for a makeover. The focus option would not work, and the message given said that the film data could not be found, so the scanner would not focus on the slide. It said to refer to the command log for more information. This was done, and a listing was found for "Film Menu Error (5201): s = I FILM_MENU," which is not listed in the system's section for error messages. Two other later messages were "error on command 'download'" and "error on command 'mainscan.'" It seemed that the film data was still there, but it was no longer being accessed. A call to the Kodak Help line yielded no results, and so I changed the settings for the film type several times to see if the scanner could read any setting. This was done, and the last time it automatically reset to the default, which is the universal setting. When the focus command was tried again it worked, and continued to do so until a periodic system shutdown was done and the system was then rebooted. The focus option once again did not work, and only resumed operation after a preview for a new order was requested. In looking back a total system shutdown was down just before the first occurrence of this problem, which included the turning off of the data manager. Obviously this causes some problem with the scanner accessing the film terms, but at least this problem can be corrected with a minimum of effort.
RATE OF WORK
The first day of work on September 25, 1995 with the Kodak PIW involved familiarizing myself with its operations and beginning the scanning of 100 architecture slides. After some initial trial and error 28 slides were digitized in a two hour period on the first day. The process was slowed by the fact that a number of the images were not of the best quality and some adjustments had to be made to improve their appearance. Although this quality problem was also consistent throughout the remaining 72 slides, they were successfully scanned the second day in a four hour period. Thus the rate of scanning increased from 14 to 18 per hour, with an average rate of approximately 16.
The second set of 114 slides from the Motley Collection in the Rare Book and Special Collections Library was completed in a seven hour period over a three day time span. Although the slides were in much better condition than the previous set, the rate of scanning translates to approximately 16 per hour, which is the same overall rate as the architecture slides. The reason for this is that great care was taken to insure the quality of the resulting images, and a few did have some problems because of the faintness of the original drawing.
For the third set the rate of production increased to 23 per hour, with 186 slides from the Krannert Art Museum scanned in an eight hour period. The quality of almost all of the images was excellent, with the original objects being filmed under professional conditions. The only problem with the images was the glare off some of the objects caused by the lighting, but the Kodak software being used does not allow for this to be corrected, so nothing was done.
The fourth project consisted of 174 slides from the Veterinary Medicine Library. The work went very quickly, with an hourly rate for the first set of 72 at 26 per hour, and the second of 102 at 34 slides per hour, for an overall rate of approximately 30 per hour. The overall quality of the slides and images was very good, and only three required any adjustments at all.
The fifth order was for 132 slides of the African Art Gallery in the Krannert Art Museum. The work went rapidly, at a rate consistent with the previous order of 30 slides per hour. The quality could have been better, but the lighting in the gallery affected the color of the slides and the resultant digital images. The only other problem was with two slides where the photographer's hand can be seen shielding the camera, but it can be cropped out in the final image.
Order number six consisted of 962 slides of horse diseases from the Veterinary Medicine Library. Due to the graphic and visually unpleasant nature of the vast majority of these slides, it was preferable to work at a rapid rate and not spend time closely inspecting the digital image. This was certainly helped by the professional nature of the slides, and the fact that only four percent needed any quality adjustments. The work rate for the first Photo CD was 30 slides per hour, which reflected the time period needed to get a good sense of the slide quality and the amount of work that would be involved. The work rate quickly increased to 40 per hour for the second CD, and then 44 per hour for the third. I then received the assistance of Mary Shultz, who is a student in the Graduate School of Library and Information Science working as a graduate assistant in the Veterinary Medicine Library. I trained her in the use of the PIW, and the rate for her work was 45 per hour for the first two CDs, and 48 per hour for the second two. I then finished the last four CD's in the order at a rate of 50 slides per hour. This is probably close to the maximum rate that can currently be achieved with the system, if care is taken to insure that each image has been properly scanned and any necessary corrections made.
The final order involved 37 images on 35mm roll film of historical maps from the Map and Geography Library, which took 2.5 hours to complete. It would have taken less time except for the need to use the manual gate because two exposures had been filmed for each map, only one of which was used. There were also several instances of images needing to be re-scanned because of bright white marks created by dust on the film that appeared on the first scan. But even with these concerns that slowed the production rate the resulting images were of excellent quality.
In reviewing the work done with the Kodak PIW over the last four months in digitizing a total of 1668 slides and 37 film exposures, I believe that it does produce very good images of a consistent enough nature for any private or commercial users. The quality of the scanned images, with the minor touchup work allowed by the system, is certainly adequate for these purposes. But more is needed if this system is to be adapted for use by libraries and archives for the digital preservation of images. The digital image to be written on a Photo CD should be of extremely high quality and as accurate a reproduction of the original item as possible, so that it can serve as the archival copy, and remain in that state long after the original image is gone. The excellent, detailed images of the historical maps do show the promise of the PIW system for archival preservation, and what can be produced on a more consistent basis if the following adjustments I have listed are made.
There are several areas where the Kodak system can be adjusted in order for it to achieve this level of quality. One is a higher resolution monitor that enables the imaging technician to more accurately judge what the digital image will look like when it is printed out utilizing a high quality printer. The second is the inclusion of Photoshop or a similar application in the workstation, to allow the technician to make any necessary adjustments to the image before it is written on a Photo CD. This allows for the highest quality image being the one digitally archived. A third area of improvement would be the way the scanner captures light in an image. There is little problem with natural light, but any artificial light is reproduced as a bright yellow tint, which obscures the true color of the objects being filmed. The problem with adjusting for this scanning tendency after the fact is that the entire image has to be corrected for this, and not just those portions that really need it. Doing this creates a new distortion of the original image in compensating for the problems caused by the lighting. A fourth area is the need for a carousel or other device that allows for a large number of slides to be scanned at one time, as opposed to the manual, one at a time approach of the current operation. This would make the system much more cost effective in terms of labor expenditures.
In conclusion, I feel that the Kodak imaging system does have the potential for the archival preservation of images, but more adjustments have to be made in the system for the special needs of archives and libraries, which were not a concern when it was originally developed for commercial purposes. The digital preservation of historical items is an area of great opportunity for the imaging industry, and if the concerns listed above are resolved, then the Kodak system can certainly be an important player in this field.
Return to Kodak Digital Imaging Project