HDRi iSRD not affective at non-original resolutions?

[Gregory C]

hi.

I'm scanning my film at the highest hardware resolution possible for my Nikon Coolscan 4000ED scanner. I'm saving to HDRi. I'm then opening the files in SF HDR Studio.

There is considerable mold on my negatives (20 years old, and Hong Kong is rather humid for film) so iSRD is important. I've encountered a problem though which I find confusing.

If I set my Frame resolution to that of the original (4000dpi) at 100%, the iSRD works as well as expected. If I set my Frame resolution to 100dpi (screen resolution) to export at 2000x1400 pixels, the iSRD is almost entirely ineffective; i.e., the output file appears as if iSRD was not even turned on. I have been careful to set the output resolution and scale before playing with the iSRD settings.

Here are screenshots of the iSRD screen for the two output settings:

4000dpi, 100%, using iSRD Detection of full strength 20

HDRi iSRD 20 @ 4000dpi.jpg (234.18 KiB) Viewed 2709 times

100dpi, 2000x1400pixels, using iSRD Detection of full strength 20

HDRi iSRD 20 @ 100dpi.jpg (192.06 KiB) Viewed 2709 times

100dpi, 2000x1400pixels, using iSRD Detection of 16

HDRi iSRD 16 @ 100dpi.jpg (140.94 KiB) Viewed 2709 times

Please notice that the iSRD misses much of the mold area (the white spots and spider patterns) when output is set 100dpi.

Is this a limitation of saving to HDRi? That the files have to be saved not at full resolution, but at the planned output resolution for iSRD to be able to use the saved Infrared affectively?

sincerely,
Gregory

[Gregory C]

judging by the recommended procedure of setting output settings before configuring SRD, and by what I'm seeing here, I'd assume that you're scaling the Infrared layer before matching it against the image. Unfortunately, the scaling does not accurately match the image and after scaling, the Infrared layer can become unusable.

I would suggest that in some future version of SF HDR, that you process the Infrared version of iSRD before scaling to get maximum use and accuracy of the Infrared layer. The current process of scaling first, matching later would make sense with the software version of SRD but not with the Infrared version.

In addition, I suspect that matching the Infrared against the raw data would be more productive than matching it against the Negafix reversed image. I have no idea though if you're already doing that.

Since a correction of this magnitude in the software is not going to happen soon, we as users will have to work around it; i.e., use iSRD at original scale and resolution to export a cleaner 48-bit HDR image first, and then run SF HDR again on the clean HDR image to export to the scale and dimensions we desire, possibly with software SRD where appropriate (sometimes the software SRD is more effective than the Infrared SRD; kudos to the engineers). Fortunately, it's a simple work-around.

I'm archiving everything. My original idea was to change the resolution to 100dpi for display use because leaving it at 4000dpi automatically displays the images at 1"x1.5" when viewed in Preview. However, changing my process to archive everything at the original 4000dpi won't be a problem either :s)

cheers,
Gregory

[SveinPetter]

Hello.
Started scanning one week ago with Reflecta Scanner RPS-7200 Profesjonell and SilverFast Archive Suite.
Scanning negatives so far. A few day ago, I scanned 36 negatives from Kodak, and after some negatives the ISRD started to get offset.
A line(scratch) is marked as red, not on line but offset. In some scans very little, others more.
When processing, the red area is "fixed" but the scratch is not fixed.

[LSI_Ketelhohn]

Dear customer,

please make sure your scanner is placed on a stable vibration free surface.
Make sure your SilverFast and drivers are the most current ones.
Also the infrared channel has a lower resolution than the RGB channel.
An offset is more likely to occur at higher resolutions.

kind regards,
Arne Ketelhohn.