I agree. Also with dslr you can wet mount the film which does change the grain rendition. If you happen to use a Sony a7R4 or 5 you also have pixel shift at your disposal to bypass the sensor filter. I've also seen people use enlarger lenses, which perform very well. The APO variants boost micro contrast. As with everything there are many options. The DSLR scanning workflow can also be automated quite a bit, although not sure there are commercial solutions (probably more in the DIY space).

@@jackskalski3699 I've tried Schneider Kreuznach Componon-S from 50mm to 150mm and I honestly can't recommend them. The contrast is great but the resolution leaves a lot to be desired. IMHO a Olympus OM 50mm F1.8 on extension tubes gives better resolution; if worst contrast depending on the exact model. I love the Componons for wet printing just not scanning. Maybe I'm doing something wrong so any advice is appreciated. I used them on a set of Pentax M42 bellows and like I said I wasn't thrilled. I haven't tried fluid mounting for camera scanning, just for 4x5 on the v700. I'm going to have to try that! The grain alisising is my main issue. Focus stacking helps a bit but not a huge amount. I add a tiny bit of grain to my scans to make them look more natural. So hopefully fluid mounting helps! Thanks!

Much appreciated!

You can also easily do such inversions via Photoshop (even in batch if you prefer). Using levels you can set your white and black points and see how far your exposures are off 😅

Do you mean dust removal on the negative or dust removal on the digital file? You can buy negative cleaner at the same place you buy your film chemistry, I think it’s just isopropyl alcohol but I always get a branded one in case it has something special to protect the film. Cleaning the dust off the digital file is just a case of zooming to 100% in photoshop and panning around with the spot healing brush tool.

@@JonathanNotley In terms of work flow, removing dust and scratches on the negs can be very time consuming, so if you're in a hurry auto dust removal must be the way to go.

Use a ColorChecker Passport would mitigate the interpolation of colors. It really does affect the colors of my negatives.

@@shang-hsienyang1284 How do colorcheckers mitigate the interpolation? That's baked into the engineering due to how 1 pixel is actually 4 pixels RGGB interpolated into 1 colour right?

@@professionalpotato4764 it's kind of difficult to explain everything in a comment. You can find many tutorials online on how the colorchecker passport can create a bespoke Lightroom profile for your camera+lens combo under a certain light temperature.

You can use negative lab pro with a vuescan DNG or a silverfast .tif but you can also hammer a nail in with a tin opener

@@JonathanNotley Terrible way to take constructive criticism. it's a difference that can be eliminated and is repeatedly asked in your comments.

I use my Sony A7rIV for this with the Sony 90mm Macro lens and the quality of my 120 film is insane. Many people have asked if it was done on a large format camera due to keeping the high resolution during the scanning process. One tip in reducing the vibrations is using the pc-usb tethered method since it allows to mainly control setting while not having to touch the camera or stand. This great since you can place the setup in a dark space and sit in a well lit location and be able to make any need changes. This also doable since I have an automatic film advance/ negative holder for 35mm and 120 that I made from 3d printing. Inspired by the negative lab pro holders but improved upon and made on a machine that cost less than a full set.

So strange why people shoot film( analog ) only to use a digital camera to take a picture of the negative and edit that image like any other digital image.. Why don't shoot digital which is absolutely neutral and true to live in reproduction and you can edit it in post to look analog or look however you want it..

The PIE XAS resolves, on average, 4300ppi. That is equivalent to the old scanners you mentioned. I have personally tested the 5400ii, A7R4, and XAS. The max I saw out of any frame via scanning was 24mpx of data. The only scenario I got more was to darkroom print and then flatbed scan, which achieved approximately 40mpx with tmax 100. The only exception was Adox cms 20 ii. That film, when shot with the 40mm or 105mm Sigma Art, out-resolves the XAS, and provided data transfer rates between 40 and 50 megapixels when scanned with the A7R4. I give a range because at those resolution levels it impacts data transfer, so chart testing isn't sufficient for a resolution description. Dark room printing the usaf chart got me close to 80mpx or so though:) There was a point where you could see printing larger yielded no additional data. The biggest issue for the cms test was matching the glass to the film. I chose the 40mm and 105mm Art because Roger Cicala's testing found those to be the highest resolving normal and portrait lenses in existence. I did all this in pursuit of deciding if I really needed medium format for the sizes I print, and also, what are the limits of analog.

@@jw48335 I follow your ;points and generally agree. One issue remaining is that PIE units do not resolve anything close to the manufacturer's ratings, a common problem with all of these currently produced Chinese models. The best PIE model is rated at 8200, but lab tests show that it's maximum capacity to 3250. Additionally, its's software/hardware produce a 210 meg file at that resolution, while only yield 42 meg of picture file, the rest being duplication, noise and the by-products of bad software. Frankly, the same resolution and a much smaller file can be had from a 30 year old Minolta dual Scan IV.

@@randallstewart175 I understand what you are saying, but the XAS and XES are rated at 10000 dpi, and out of that, they resolve 4300 and 4100 dpi (respectively). My testing and others have found that to be the case using usaf 1951 targets. The XAS has autofocus so it squeaks out a bit more than the XES. That equates to ~24mpx of real-world data. It compares *very* favorably with camera scans performed with my EOS R. Silverfast iSRD has no practical impact on image quality 90% of the time. I have done a *mountain* of testing on this stuff. I have even done scans using an A7R4 with pixel shift of an Adox CMS 20 frame shot with a Sigma Art 105 :) It does crack me up the people spending all that cash on old scanners.

You are plagiarizing my post from another thread on KZfaq or from reddit, LOL. I can cite the lp/mm resolution for the 105mm Sigma from memory, along with the Sigma Art 40mm. I used 12 ISO for the CMS 20. It's nice people are referencing the posts with my results, but don't pass them off as your own.

@@jw48335 Prior to your post in reply top mine above, I've never heard of you nor am I aware of your posting of any research or data. I thought reddit was devoted to silly stories of bad sex. In fact, the more favorable resolution performance you claim of the scanner is about 25% greater than published reports from a German site that has credibility. I chose not to remark on that point because there is no serious standard for these resolution tests, so no way to determine objective accuracy. I'm sure you believe in the validity of your methods, as other do for their own. IMO, the Primefilm models are a piece of Chinese junk, mainly for their dedicated software, but today they seem to be the least worst of what is available new. You are free to disagree.

Watch around 4:30, the way it works is basically just using lightrooms auto WB dropped off a proprietary curves adjustment applied to invert the negative image and allow LR to work. There’s edge buffer built in (basically a crop tool applied then unapplied by NLP) which I think works at 5% by default. Thinking of NLP as a series of macros for the LR engine rather than its own engine makes it more intuitive.

The scanner is $489 at B&H (I must point out that the reflecta RPS 10M is €900, crazy) and the macro setup can be achieved in different ways. I guess if I was to suggest a macro setup for a similar price I’d recommend a film era macro lens with an adapter for a digital camera body you already own. Since manual focus is best it doesn’t really matter if you get something without AF so long as the optics are good. You can pick up one of the copy box things for around €100 which makes a big difference to not only quality but also convenience.

@@JonathanNotley Thanks so much Jonathan. I scan on my three-in-one printer and it is actually not that bad. Improvement is always welcome though. Meanwhile I found out that for my Olympus camera a pretty good macro lens is available from €499 and that might get me into macro photography as well. Cheers!

I got the light source and film holder here: www.skier.com.tw/web/shop/shop_in.jsp?pd_id=PD1599466967388 My copy is the v2 and it looks like he's now selling a v3

@@JonathanNotley how is that film holder? Im planning in buy one. Thanks! 😊🙏

It’s good, I think the v3 is a little overpriced though (mine is a v2). Look at the negative supply co ones too and make a choice

@@JonathanNotley thanks for answering!! i think i will take ir becouse it has all you need for scanning. thanks!

ISO 100, aperture 7.1

It’s called a skier sunray copy box ii

I don't understand, how do you use a mirror to ensure the film is leveled?

@@pakhimli2585 When you centre the camera lens to align with itself through the mirror, it’s perfectly perpendicular to the mirror.

Yeah but now I’m going back under a rock for 5 months, it’s the way this channel works

Using NLP and Silverfast together in the same image pipeline is like using a horse to tow your car. They’re both methods of transport/(neg inverters). If you like Nate’s curve presets you can just make a Lightroom preset for silver fast jpgs/tiffs

No. Emulsion towards your lens.

Emulsion towards lens. You can tell by the way it reflects light or by facing the mirrored numbers on the film strip towards the lens.

Huh? The downsampling retains extra detail from the negatives. You’ll see an enormous difference between a 4k dpi scan down sampled to 3k long-side and a 3k long-side scan. A nice proof of this is an ability to tell the difference between scans on KZfaq, which resamples for video resolution.

@@JonathanNotley In the video, you scanned at 300dpi - not 4k dpi. Did I misunderstand something?

@@MD-en3zm I scanned at 5000dpi on the scanner and whatever the native dpi is on the digital camera body is, minus a bit of wiggle-room for the crop I applied. It's an unfortunate quirk of film scanner branding that the scanner will produce 'maximum detail' at a DPI below maximum for most manufacturers (Epson are one of the worst for this, plustek too!) I think that Minolta are more honest and Nikon can generally be trusted. I think a 4000dpi scan from a Nikon coolscan would look better than a 5000 dpi scan from this Primefilm XAs, which would definitely look better than a 10,000 dpi scan from my Epson V850.

@@MD-en3zm Also I think what confused you was that 'preset' box where I can set my output dimensions in Silverfast. The dpi number box is for the digital file the software outputs, it's the slider which moves from green to red that is my acquisition dpi, which you can see I set to 5000

@@JonathanNotley That may have been what confused me - I thought you were scanning at 300dpi, which would be a rather useless comparison. Thanks for the clarification!

Why use two raw conversions when you can use one? I mention the whole roll automation thing around 21:00, it’s the one of the main advantages of the scanner

​@@JonathanNotley probably in your question is the answer. Greetings

@@JonathanNotley The point of it is that you do not do two conversions. You simply create a linear RAW scan first using either VueScan or Silverfast (it still looks like a negative) and then process that RAW file via NLP. That way you can make sure that the actual color conversion is done by the same software. The conversion of colors is not standardized.. both NLP and Silverfast use their respective algorythms for that. Regarding detail: I should note that you applied sharpening during your scans. The LAB sharpen option was on in NLP... not sure if you sharpened the scan in Silverfast too but it's just better to establish an even ground first. You can still sharpen both later to see how both scans are treated by it. Atleast you will be sure that you are sharpening via the same software and by the same amount. Another point regarding the noise; since you are using a photo taken on your digital cam it's being demosaiced by the software (in this case Adobe's LightRoom/CameraRaw). There are different demosaicing algorythms and they are basically converting the raw pixel/color data into the photo. A good example of how different the demosaicing can work is to look at Adobe's horrible demosaicing of Fuji Xtrans sensor photos (very well known issue) and that you get a very different result using something like RawTherapee (free and allows to select/try different demosaicing algorythms) or Capture One. I'm not 100% sure with your setup/gear but afaik a certain amount of color noise supression is also applied in that process (before you even use noise reduction in post). And with that it can also lead to actual color noise of the film stock to be removed, too. All this is to say that digital photography uses a completely different type of sensors and produce the final RAW image it's being digitaly treated/interpreted. So the amount of noise and how it is represented or rather supressed is different. Most cameras like yours use BAYER cmos sensors with its mosaic array. The ones used in dedicated scanners are CCD inline sensors. They produce better color depth/separation and beside all that don't need the kind of demosaicing. That was a lot of text but I thought I might be able to provide some additional info.

The closer you look the more the details come apparent - that’s the point of comparison. You can make anything look the same if you print it the size of a postage stamp and view it from 15ft.

@@JonathanNotley if you need to blow it up 6x maybe you need more 6x more megapixels in your scanner/ camera?

Ya, i agree. I dont think that is color noise. Color film grain is colored, not monochromatic like the lightroom "Grain" effect seems to think it is.

1 - there’s a resolution limit to 35mm film, even if I scanned it on a phase one a lot of the pixels would just be added bulk. Normalising the files to 3000x2000 let me strip out the fluff and compare relative detail. 2- Aren’t they just presets for the sharpening tab in LR? Maybe leaving them all at 0 would have been fairer

@@JonathanNotley I agree but, 3000x2000 just feels a little low to me. In my experience scanning with my camera I have gotten much higher usable resolution than 6MPx (it might be enough for fuji c200 though, but Gold showed lots of details that would get lost in a 3000x2000 resize). About the sharpening, yes, it just messes up with LR bars but I don’t really like the way it does with Lab

@@JonathanNotley I would have used the real-world optical resolution of the XAS, which is 4100 to 4300 ppi, depending on autofocus. My best results with usaf chart testing and t-max 100 came by scanning at 5k and down sampling to 4300. That also yields a nice 24mpx image. My results were consistent with other testers. I find that to be a convenient size for 12x18" prints. Running those files through 2x enlargement in Gigapixel AI gets you to a 16x24" print out of the XAS, and my eyes can't distinguish that print from the output of an A7R4:) This assumes top end glass etc. On a related note, you get ~2700 ppi out of an optimized V7X0 /V8X0 scan. It would be unlikely you would see a difference between that flatbed and the XAS if all you're working with are 6mpx resolution. Cheers.

Color noise... It's simply the noise level of the scanner electronics. He did not used Multi Exposure. That would solve problem.

Yes the DSLR kit is more expensive but I was working on the idea that people interested in the comparison had at least 2/3rds of the camera scanning kit to hand. I guess you could buy a previous-gen pro body and a MF macro lens for far far less.. I heard of people selling d810s for £250

Oh yes there are thousands of ways you can go for dslr/ml scanning which could be less expensive. #frankenscan

Dust removal only works with color films. If you shoot b&w, it's no help. Also, both methods exaggerate b&w grain (“grain aliasing”) but the film scanner seems to exaggerate it more in my limited experience.

@@jlwilliams correct, only with color film (except for Ilford XP2), and it’s a tremendous time and effort saver. And if film scanner is “exaggerating” grain (you imply it’s making it worse, I don’t agree), it’s probably because it has a higher overall practical resolution in this task.

TURN UP YOUR SPEAKERS THEN (CAPS LOCK FOR ADDED VOLUME)