Scanning film: Flextight X5 vs Sony A7R IV

[Steven Lee]

Ladies,

I sold my Coolscan many moons ago. Back then I was primarily shooting Provia and Velvia and was quite happy to leave film behind when I could finally afford a Canon's full-frame DSLR. When I got back into film a few years ago, I discovered that Nikon got out of the scanner market. I experimented with Epson V850 and Plustek scanners, but couldn't deal with the slow workflow and scanners being dust magnets (I have been shooting only B&W) drove me to camera scanning.

I built a dedicated scanning rig based on an enlarger stand and the permanently mounted 60 megapixel Sony A7R IV with a Sigma macro lens. In terms of resolution I was quite happy. Very few negatives actually contained the level of detail which exceeded my scanning limitations. Moreover, I figured I could just order fancy scans (drums!) for eventual masterpieces to hang on my walls.

Recently I've started to shoot color again. Inverting color has been quite a challenge. I was never happy with the automatic conversions offered by various plug-ins (NLP, Negmaster, etc), so I've been inverting manually. Occasionally I would ask myself whether I was leaving something on the table. On photo forums people would often say that you need a purpose-built scanner, with a tri-linear CCD instead of a color filter area sensor used by cameras.

So I took a few recent rolls of Portra 400, Ektar and Fuji 400H Pro and ordered raw scans using the "premium" service at my local lab. They use Hasselblad/Imacon Flextight X5. They delivered the files in 16-bit TIFF and FFF formats, uninverted, so I was able to apply my usual/manual color inverting routine.

I inverted a bunch of X5 images trying to quickly please my own taste. Then I found my old camera scans to see how they compared. Below is one of such comparisons. I won't even tell you which one is which, but I will say that I can tweak either image to look identical to its counterpart. The difference in colors you see is explained by me inverting them separately, being in a different mood, a few months apart.

My conclusions are:

• The workflow of camera is scanning is incredible, compared to scanners.
• X5 has an edge on resolution, but as I said that difference has almost zero practical value.
• In terms of color, I can clearly see that I can get the same color with the Sony as I get from the X5.
• Price-wise, my entire scanning rig was about $5.5K compared to $10-15K for the X5.

Same Ektar negative, converted manually from X5 and A7R IV raw files:

 

[Steven Lee]

Below isPortra 160. Again, any image can be easily tweaked to match the other. I was a bit sloppy on black point setting on the red channel on the left.

 

[madNbad]

It’s already difficult to find service for a high end scanner but if you want a better digital camera, that’s pretty easy. Scans look good, I tried the Sigma Art 70 macro but prefer the Sony FE90. It focuses faster and holds focus better. Keep at it!

 

[faberryman]

Steven, what do you do with your scans?

 

[brbo]

I like the ones on the right better.

 

[Helge]

I take it that you are shooting. 6x6 and 645?
Did you use a tripod?
What lens and what aperture?
It’s really difficulty for us to follow your reasoning without pixel level crops.
Would it be possible to provide these?
Perhaps a small square from a detailed spot in the middle and the edge of the frame?

 

[runswithsizzers]

Thanks for posting your results. I am in the process of doing some similar comparisons, although with much more modestly priced equipment.
I would be curious to know more details about your manual color inversion process.

 

[Patrick Robert James]

The Bayer sensor is the downside of camera scanning color. The subtle colors can be compromised. I think the character of film is altered a bit. I've never used a pixel shift camera though which would theoretically get around that. I'd be curious what people's experiences with them were.

My conclusion over the years is camera scanning is worth it for proofing or small enlargements if you have a setup that makes it quick to do. I also think it is worth it if you want to eek out every ounce of resolution with stitching. That is such a pain to do though it rarely is worth it. I still use my Nikon scanner because it is convenient and does the whole roll. The scanner takes more time but it is less of my time. For medium and large format I just use a flatbed. I still think drum scanning is king for medium/large format.

I'm surprised no one (or maybe someone has) come up with an automated stitching system similar to a cinema rig. A Sony like the A7RIV/V with it's pixel shift would be pretty nice if the lens had enough resolution. Or better yet, a medium format camera like the Fuji. Of course at a certain point you might as well get a drum scanner.

Color conversion is the main stumbling block with camera scans. I do it myself. Just looking even at Steven's small images on screen I can see why he does too. He obviously has a fine eye for color.

 

[Les Sarile]

What was rhe X5 scanning dpi - 2000, 4000or 8000dpi?

 

[Helge]

The Bayer sensor is the downside of camera scanning color. The subtle colors can be compromised. I think the character of film is altered a bit. I've never used a pixel shift camera though which would theoretically get around that. I'd be curious what people's experiences with them were.

My conclusion over the years is camera scanning is worth it for proofing or small enlargements if you have a setup that makes it quick to do. I also think it is worth it if you want to eek out every ounce of resolution with stitching. That is such a pain to do though it rarely is worth it. I still use my Nikon scanner because it is convenient and does the whole roll. The scanner takes more time but it is less of my time. For medium and large format I just use a flatbed. I still think drum scanning is king for medium/large format.

I'm surprised no one (or maybe someone has) come up with an automated stitching system similar to a cinema rig. A Sony like the A7RIV/V with it's pixel shift would be pretty nice if the lens had enough resolution. Or better yet, a medium format camera like the Fuji. Of course at a certain point you might as well get a drum scanner.

Color conversion is the main stumbling block with camera scans. I do it myself. Just looking even at Steven's small images on screen I can see why he does too. He obviously has a fine eye for color.

I’m very surprised anyone haven’t stepped up to the plate and made a cheap, small, super high quality scanner based on a the ideas of camera scanning yet.

It’s what the market I clamoring for.
We need it very soon, if we are not to loose all the people who jumped aboard film in the last half decade.

Plustek proves it can be done with production costs in check. They just need to ditch the 1998 tech of linear CCD and 135 exclusivity.

It’s faaaar more important than new cameras, film holders, the trillionth “new” B&W film, or whatever else people are faffing about with.

Software for stitching and reversal is probably the biggest expense.
It should just be written in the highest level, interpreted, most user friendly language available for easy development and maintenance.

No one cares if it takes 1 millisecond or thirty seconds to convert a photo, as long as it can be done in batches.

 

[Steven Lee]

@Les Sarile All scans were done at the maximum resolution. For the X5 it's a different number based on the negative size, you can look up the specs on Google. The samples above are 6x6 and 35mm.

@Helge I'm traveling this weekend and do not have access to full resolution files. Just wondering why you need the 100% patches? I was trying to compare color here, not resolution. Otherwise I would have picked different negatives, exposed on a tripod, at optimal aperture, with good detail in focus present in the scene, etc. And still... I wouldn't want the focus of the conversation to flip to resolution, which happens all the time! For example, the X5 scans all have aggressive sharpening applied, which the lab claims they couldn't disable. Subjectively, I think it makes grain look unnatural, while I know others may prefer it. Anyway, I may get back to this thread when I return home.

 

[Helge]

@Les Sarile All scans were done at the maximum resolution. For the X5 it's a different number based on the negative size. The samples above are 6x6 and 35mm.

@Helge I'm traveling this weekend and do not have access to full resolution files. Just wondering why you need the 100% patches? I was trying to compare color here, not resolution. Otherwise I would have picked different negatives, exposed on a tripod, at optimal aperture, with good detail in focus present in the scene, etc. And still... I wouldn't want the focus of the conversation to flip to resolution, which happens all the time! For example, the X5 scans all have aggressive sharpening applied, which the lab claims they couldn't disable. Subjectively, I think it makes grain look unnatural, while I know others may prefer it. Anyway, I may get back to this thread when I return home.

Well you did mention resolution yourself.
And that is in general an interesting difference between the Flextight and the camera.
Both the actual resolution and the artifacts accompanying it.

 

[NortheastPhotographic]

I was really into camera scanning for a couple years, and had a very nice Negative Supply set up. I have a very heavy duty Bessler copy stand, mirrorless with pixel shift, Sigma 70 ART...I put a lot into it.

Long story short, I now scan with an Eversmart Supreme II. Camera scanning can indeed be very useful in certain applications but when you need to knock it out of the park, Scitex or Drums are the way to go. Speed is relative, especially as you go up in formats you end up doing a lot of adjustments for weird light bleed and distortion. I think my camera scanning rig may have produced better 35mm scans when using pixel shift.

 

[Steven Lee]

@NortheastPhotographic hehe these very same negatives are in the mail on their way to Michael to be scanned on the iQ3. @Helge I will post 100% patches in PNG format later if you're curious.

 

[Les Sarile]

@Les Sarile All scans were done at the maximum resolution. For the X5 it's a different number based on the negative size, you can look up the specs on Google. The samples above are 6x6 and 35mm.

I could but there are many orher unknown variables so I ask, what resolution was used.

 

[albireo]

Long story short, I now scan with an Eversmart Supreme II.

Hi Northeast, I'd love to read your impressions/a review of your Eversmart set up and even see some samples if you have the chance. Thanks!

 

[Helge]

@NortheastPhotographic hehe these very same negatives are in the mail on their way to Michael to be scanned on the iQ3. @Helge I will post 100% patches in PNG format later if you're curious.

Yes please. Would be very much appreciated.

 

[Eric Rose]

I use a Nikon D800e which eliminates at least the anti-aliasing filter. Lens is a Nikkor 60mm macro. My scans are VERY sharp.

 

[NortheastPhotographic]

Hi Northeast, I'd love to read your impressions/a review of your Eversmart set up and even see some samples if you have the chance. Thanks!

Sometimes you can see samples on our Instagram. My impressions of the Creo are that I love it first of all. It makes scanning easy. The 5600ppi is IMHO a little low for some 35mm films, that's where those Imacons really sing. But for all other purposes I would prefer the Creo. The upper spring loaded glass holds the film perfectly flat. The top glass is ANR, and for the base glass the fluid mount glass is slightly etched while the dry is the clearest optical glass I've ever seen. However, after some testing I determined that the fluid glass is fine to use regardless of dry or wet mounting.

For 120 - large format, the 5600ppi goes from being great to overkill pretty fast. I've learned to appreciate lower PPI scans as assuming everything must be a mural just adds workflow headaches. You can scan an 8x10 sheet at 1000ppi and get a large print, and very little dusting. Sure I could get a 5000ppi scan of that with multi sampling of quadrants and stitching (software has a 3.5gb file size limitation), but it is a complete waste of time and effort. Once you exceed the needs of your largest printer you're already doing more than you need to. So it's been a bit of a lesson in workflow planning.

The scans themselves are incredibly detailed, low noise, and have no geometric or flaring issue. They're right in line with what you get from drum scanners. You can be extremely productive with this scanner and the quality is top notch, that's what I love about it.

 

[Steven Lee]

@Helge I have full-sized scans (16-bit TIFFs) of the same negatives made with the Imacon, Creo iQ3, and Sony A7R IV with Sigma 105mm. The files are massive, approaching 1GB total for 3 scans of the same negative. Color balance, file dimensions, and sharpening are not matched. Flextight scans aren't even inverted. Scanner TIFFs have ungodly amount of dust.

I do not have the time to clean them up and make them directly comparable, but I can find a way to post links to full-sized raw data tomorrow. If someone has a better idea, LMK.

 

[Helge]

@Helge I have full-sized scans (16-bit TIFFs) of the same negatives made with the Imacon, Creo iQ3, and Sony A7R IV with Sigma 105mm. The files are massive, approaching 1GB total for 3 scans of the same negative. Color balance, file dimensions, and sharpening are not matched. Flextight scans aren't even inverted. Scanner TIFFs have ungodly amount of dust.

I do not have the time to clean them up and make them directly comparable, but I can find a way to post links to full-sized raw data tomorrow. If someone has a better idea, LMK.

You don’t have to post full size. Just one or two crops of the same place on the frame from the different methods. Or comparable places on diffferent frames.

 

[Steven Lee]

@Helge https://www.photrio.com/forum/threads/comparing-imacon-x5-creo-iq3-and-sony-a7r-iv-scans.200990/