Being a Silverfast user, and given the fact I’m scanning old, seemingly perpetually dusty and oftentimes damaged negatives, I’ve been using the TIFF output from Silverfast after applying dust removal. I would prefer to have DNG as the file type for my master archival files as I tend to convert my camera RAWs to DNG upon import into Lightroom so that would keep things uniform. When I import the TIFF into Lightroom I am unable to automatically convert to DNG as it’s not a camera RAW, but I have found that I can export the TIFF to DNG and then use NLP to convert that DNG as if it was a Silverfast DNG and get very good results. Furthermore, when I export the Silverfast TIFF to DNG the file size goes from ~120mb to ~85mb, another bonus.
My question is whether I’m losing data/pixels/editability by converting this TIFF file to a DNG in this manner. I read somewhere that by converting a TIFF to a DNG you make a “linear DNG” which is not inherently the same as a RAW file and you therefore lose the original editability of the file… I’m not knowledgeable enough about this to know what this means or if that is what I’m doing, but I do know that my end goal is to have files as close to RAW that remain as editable as possible for long term archival purposes. If I can keep things all DNG and save about 30% of disk space in the process, even better.
Never mind about this… I’ve decided to just use the 64bit HDRi Raw scans from silver fast as my archival format. I guess more data the better right?
The only reason I previously even wanted to do the above was because I’m sort of afraid Silverfast might not be a thing in the future, in which case I may have a lot of files I wouldn’t be able to use the infrared dust removal for. But I figure, Silverfast probably isn’t going anywhere anytime soon, and even if they did there should be at least some software option to use those infrared channels.
Down side I guess is now my RAWS are >230mb each… anyway around this?
On the subject of longevity:
From what I understand about the DNG format is that it is based on the TIFF format internally for storing the pixeldata. For demosaiced images, rougthly: DNG = TIFF + metadata. Inside the TIFF standard there is an option for 4 channel 16-bit data, e.g. 3 for RGB and 1 for dust removal. If you are worried about the longevity of either file formats than don’t, both DNG and TIFF will be around for a long time and if not the chances are very high that there will be a way to convert.
On the subject of file size:
Since DNG and TIFF can store pixeldata in identical ways, there is no reason that your files are vastly different in filesize other than there must be a significant difference in how the data is stored. Let’s look at some examples:
(1 byte = 8 bit)
one pixel stored as 1x16bit = 2 bytes (i.e. 1Mpx image = 2Mb)
one pixel stored as 4x16bit = 8 bytes (i.e. 1Mpx image = 8Mb)
one pixel stored as 4x32bit = 16 bytes (i.e. 1Mpx image = 16Mb)
etc.
I’m guessing your files are so large because either silverfast is storing many pixels (that may have the same colors as their neighbors) or it is storing each pixel in a large type format (many bits per pixel) or both. Whenever you go from an unreasonably large file size to a smaller one you WILL loose information. However, the amount of loss in quality can be very small compared to the amount of reduction in filesize, especially when going from huge files to large files. It is up to you if the filesize reduction comes at a quality price you are willing to pay. Have a look at this nice paper if you want to understand this more quantitatively.
Linear DNG files contain demosaiced images. You can get these a) in Adobe DNG Converter if you check the respective custom settings or b) by a raw developer that does so like e.g. DxO PhotoLab.
The DNG file I exported using Lightroom had a size of 26 MB vs. 133 MB when exported by DxO PhotoLab as a linear DNG. While Lightroom’s DNG file preserves the R/G/B/G mosaic, the linear DNG only contains RGB pixels. A linear DNG can’t (afaik) be reversed to get you the original raw file back unless you embed it in the DNG.
Exporting a TIFF to DNG can create a DNG file that is smaller than the original TIFF. As far as I found in a few tests I made I while ago, NLP will convert from both TIFF and DNG files without issues. Again, your mileage may vary and I’d propose to make a few tests with a set of negatives that are everything between over- and underexposed in order to see what happens under what conditions. You can also run both formats in parallel for a while and decide later which one to keep.
TIFF is an old and proven file format that also includes metadata and more while DNG is a more recent format that has been adopted by a few camera manufacturers but not by the market leaders.
I see you’ve decided to handle this differently, but in case anyone else is pondering, here’s an obscure fact (buried in the DNG spec): the DNG format is simply the TIFF IV format with some metadata fields added on for manufacturer-specific tags. This means that in terms of image quality, neither of these has any inherent advantage over the other.
Just wondering how this works. If I’m correct, you cant use 64Bit HDRi Raw files in Lightroom/NLP. Does silverfast allow you to export a readable version? And does this version have iSRD dust removal applied?
First off, Thank you for everyone to contributing answers to my questions. I appreciate the technical details.
64bit HDRi Raw files work in both Lightroom and NLP, in the sense that the files show up normally and can be manipulated normally. They can be process by NLP just like a 48bit silverfast raw. However, the extra 16bit infrared layer can’t be used by LR or NLP, you have to use the Silverfast HDR Studio software to do this. My initial conundrum was based on my idea that I wanted to keep my “master” files as the Tiff/DNG files AFTER noise reduction was applied. My reason was because I was afraid that the Silverfast software itself might not be around in the future, leaving me with these 64bit HDRi files that I can use as above in LR/NLP but would not be able to reapply dust reduction. However, after thinking on it I decided that most likely, Silverfast will be around for quite some time, and even if they disappear there should be some software somewhere or some work around that would be able to take advantage of that extra infrared channel… And to have will always be better than to not have it.
In the meantime, this means that my workflow is as follows:
I scan my negatives into 64bit HDRi DNGs with Silverfast 8 Ai. These are saved forever as my Master Files.
I convert those files into 48bit Tiffs with HDR Studio by Silverfast, WITH dust reduction (iSRD) applied and exported at Gamma 2.2
I then process the dust reduced Tiffs with LR/NLP. I export into JPEG just for web sharing and to store in my Apple photos library.
I then delete the 48bit Tiffs… Keeping these would be nice since they would exist in my Lightroom catalog with all of my edits. However, keeping these plus the 64bit raws just seems too redundant for me, and deciding on the two, I’d rather keep the big 64bit files around as masters.
Not a perfect solution but for now it’s the best I can come up with.
Hi… I’m kinda new to silverfast but am using Silverfast 9 ai studio to scan my old slides for archival… I have my settings at hdri raw and output dng… I changed the gamma to 1.0 in the preferences… I now find that if I do a single scan I get a dng file but when I do a batch scan the resulting files are tiff… Any ideas? Thanks in advance…
What type of scanner are you using? If it is something like an epson flatbed, you would usually have to adjust all the frames with the settings you want, then batch scan. If you are using something like the Prime Film XA, like I have, and it is auto advancing the roll, I believe the first settings are applied to all but I cannot remember since I found that function cumbersome because the alignment would always mess up at some point. Hope that helps.
Hi… Thanks so much for mail… I’m using the Epson perfection V850 pro flat bed scanner with the scan settings as per the below screenshot… I’ve changed the gamma to 1.0 and have disabled the CMS… I’m doing a pre scan and then using “find frames” to select all 12 slides in the holder… If I scan just one slide at a time it creates a DNG file but if I do a batch scan I get all TIFF’s… I just can’t figure it out… Would appreciate any help because I have 3,000 slides I need to archive and I can’t do it one by one… Otherwise is it ok to just go ahead and save everything as a TIFF file?
Thanks again.
once you hit “find frames” you then have to click each frame individually and set the settings again. You could also just do one frame with the settings you want and then hit “duplicate frame” x 12 and drag it around the other frames.
So in conclusion, you have to set each frame individually but then once you do that you can still batch scan and it will then scan all 12 for you at once… and if you are using holders, then once you set up those frames you shouldn’t have to touch the settings much again since they’ll be aligned and set appropriately otherwise.
