Fuji Frontier SP3000 vs Mirrorless scanning

[David Mackovic]

Hi

Im having a bit of a dilema. I run a small film develop/scan operation in Czech republic and im in process of upgrading my current scanning setup ( epson + mirrorless scaning ) for a better mirrorless only setup.
I have A7R4 and i have tryied many lenses, general conclusion is that this is as good as if not better than Hasselblad x1 i have used. So nothing to go on about there, But what pains me is the color workflow which even with NLP is pain. Few days ago i got to text to a gentleman in other country and possibility of buying SP3000 Frontier for a reasonable price. But I wanted to ask someone more experienced than me if this is more feasible way to doing scanning "for masses" these days. My turnaround is around 30-50 films a month and Im aware that SP3000 will take long to reccoup its costs but you just cannot beat those colors and the profi film scanner lab workflow. OR do I just stay with what i already have and suffer ?

I hope it makes sense.
Thanks for advice.
David

 

[brbo]

If you are a member of Noritsu Scanner Users you've probably witness this story, but maybe not as you would probably not be asking this question. So, here it goes...

There was a guy, owner of a one-man-band lab working out of his garage. He had a Noritsu mini lab and a Noritsu LS-600 scanner (APS and 35mm only). As he was the smartest person on the planet, he was 101% percent sure that 120 format will finaly die within days and literally ridiculed anybody who asked any question about Noritsu medium format scanner (HS-1800). Months/years passed and 120 still did not die. What's more, people kept asking about medium format film development and scanning.

Since only a very stupid person would own a minilab 120 scanner, he got a digital camera, lens, copy stand and NLP and was AMAZED by the results. It was the best thing. The sharpness, the speed, the price... In his eyes, anybody still runing lab scanners was now even more stupid.

So, he was set and started taking in 120 film.

He lasted 1 (one) day and then placed a WTB ad for Noritsu HS-1800.


If you average turnaround is 2 films per day and SP-3000 is a lot of money for you, just suffer. Don't even try a lab scanner. It's better to not know how painfull camera scanning really is compared to lab scanners.

 

[David Mackovic]

If you are a member of Noritsu Scanner Users you've probably witness this story, but maybe not as you would probably not be asking this question. So, here it goes...

There was a guy, owner of a one-man-band lab working out of his garage. He had a Noritsu mini lab and a Noritsu LS-600 scanner (APS and 35mm only). As he was the smartest person on the planet, he was 101% percent sure that 120 format will finaly die within days and literally ridiculed anybody who asked any question about Noritsu medium format scanner (HS-1800). Months/years passed and 120 still did not die. What's more, people kept asking about medium format film development and scanning.

Since only a very stupid person would own a minilab 120 scanner, he got a digital camera, lens, copy stand and NLP and was AMAZED by the results. It was the best thing. The sharpness, the speed, the price... In his eyes, anybody still runing lab scanners was now even more stupid.

So, he was set and started taking in 120 film.

He lasted 1 (one) day and then placed a WTB ad for Noritsu HS-1800.


If you average turnaround is 2 films per day and SP-3000 is a lot of money for you, just suffer. Don't even try a lab scanner. It's better to not know how painfull camera scanning really is compared to lab scanners.

So you think its not worth to invest into that Frontier (5k€ price) and stick with mirrorless ? Im not the smartest i guess (trying to get any execuse to buy that sp3000 ) because i really dont enjoy the current process …

thanks
David

 

[foc]

I have owned and operated a few Fuji scanners when I has a minilab a few years ago.

The SP3000 is a very good scanner and with the automatic 35mm carrier very fast.

A lot depends on what carriers you are getting with the machine (auto 135, manual 120, 6x4.5, 6x6, 6x7,6x9) and what software and windows version it is running on. Also does it come with led light unit or bulb?

Scanning 120 will take a little time whereas 35mm can be semi automatic (you preview 6 thumbnails on screen and you can over ride the auto corrections if needed. To be honest all that is ever needed. now and then, was a density correction)

With simple good lab housekeeping, this scanner will produce the goods, quickly and with less stress to you.

Best of luck with it.

 

[David Mackovic]

I have owned and operated a few Fuji scanners when I has a minilab a few years ago.

The SP3000 is a very good scanner and with the automatic 35mm carrier very fast.

A lot depends on what carriers you are getting with the machine (auto 135, manual 120, 6x4.5, 6x6, 6x7,6x9) and what software and windows version it is running on. Also does it come with led light unit or bulb?

Scanning 120 will take a little time whereas 35mm can be semi automatic (you preview 6 thumbnails on screen and you can over ride the auto corrections if needed. To be honest all that is ever needed. now and then, was a density correction)

With simple good lab housekeeping, this scanner will produce the goods, quickly and with less stress to you.

Best of luck with it.

135f and 120 6x7 + calibration carriers included
win machine + soft
led source

 

[Steven Lee]

@David Mackovic I would look at Filmomat Smart Convert before exploring hardware options. I have played with it for a while and it appears to deliver much better no-effort default conversions than NLP. Even its UI is mimicking lab scanners.

 

[_T_]

I know as a consumer that if my lab switched from mirrorless scanning to a frontier I would have to find a new lab. But please take that with a grain of salt because I generally won’t trust a lab to do my scanning and have been doing 99.9% of my own scans for over a decade.

 

[David Mackovic]

@David Mackovic I would look at Filmomat Smart Convert before exploring hardware options. I have played with it for a while and it appears to deliver much better no-effort default conversions than NLP. Even its UI is mimicking lab scanners.

I have tried that and see no difference vs NLP. Yes controls are easier but lot less control in my opinion. But thanks for advice

I know as a consumer that if my lab switched from mirrorless scanning to a frontier I would have to find a new lab. But please take that with a grain of salt because I generally won’t trust a lab to do my scanning and have been doing 99.9% of my own scans for over a decade.

Interesting, because I was always told that, at least my customers, were more interested in colors rather than sharpness and quality that comes from mirrorless scanning. And from that i have assumed that mirrorless is just an less expensive albeit better, alternative to lab scanner.

Thanks for replys
D

 

[brbo]

I know as a consumer that if my lab switched from mirrorless scanning to a frontier I would have to find a new lab. But please take that with a grain of salt because I generally won’t trust a lab to do my scanning and have been doing 99.9% of my own scans for over a decade.

Why? I mean, your lab uses the "superior" method, yet you never use their scans and you've been camera scanning your film for over a decade now?

 

[JWMster]

This reply is to address camera scan vs. scanner on my own experience, and the possible application of that to commercial.

Simply as a user, I found the old Nikon LS8000 I had slow and replaced it with DSLR and Mirrorless scanning for a period. That experience was fine for personal quick scans, but I wasn't happy even with the B&W's. I could get them quickly and that was nice, but DUST remained a real bugaboo. I had all the rest of the gear to produce good stuff, but for the most part, I found a lot of the added gear to produce really good, steady quality camera scans as expensive as a really good scanner, and if you don't have the space to have a permanent setup.... I was actually making more work for myself than a dedicated desktop scanner requires. And the closer I looked at it, the more I concluded that plenty of folks out there were producing higher quality output with lower quality scanners than I had available, so the key was likely know-how rather than gear. I switched back to an Epson for the flexibility to do everything up to 4X5 and the Silverfast 9 with HDR and automation is pretty remarkable. Stabilizing the platform also helps simplify my process, and that's a plus. That said, I wonder that without camera scanning whether Silverfast would have felt compelled to upgrade the capabilities produce more from existing gear... but it has, and that's a real positive.

So on the basis of my own experience, I'd caveat ambitions on the basis of whether the user for whatever gear you choose is deep and broad enough to help you with the new films and new chemistry coming out to continue producing the quality that keeps photographers shooting film. If it's too narrow, you might find you're having to do too much figuring out on your own for these adaptions, and while that's always the case to some extent, I can conceive how that could feel frustrating and/or limiting. I hear the Harmon is helping to develop scanning standards for their new color film with certain scanners is great. But user experience is key, too, and with many scanners now orphaned.... this could become a narrowing option.

Good luck!

 

[brbo]

Fact is that going for ANY minilab scanner is risky even though their productivity (and a lot of image quality aspects of their output) is unmatched by anything out there even three decades later. Even the most popular ones like Fuji Frontiers, Noritsus or Pakons have problems with servicing and replacement parts.

It's not an easy decision for a "lab" with not much throughput, a pretty straightforward for a lab with a lot of traffic - you simply need those scanners.

 

[_T_]

Why? I mean, your lab uses the "superior" method, yet you never use their scans and you've been camera scanning your film for over a decade now?

Apologies. I was unclear. I meant that if I did rely on lab scans I would not stay with a lab that produces their scans using a frontier. And my caveat was there to make clear that I don’t rely on lab scans so I am not really the market for the service.

 

[_T_]

Interesting, because I was always told that, at least my customers, were more interested in colors rather than sharpness and quality that comes from mirrorless scanning. And from that i have assumed that mirrorless is just an less expensive albeit better, alternative to lab scanner.

Thanks for replys
D

The quality of the colors in a scan are primarily a factor of the skill of the operator. I’ve seen a wide range of outcomes from both systems, but I do think the types of controls afforded by the high volume all in one machines can make it easier to make bigger alterations to the image.

Or maybe it’s because the kinds of labs that use these machines are under the time constraints of being higher volume labs, but the scans I’ve seen have a greater tendency to have the contrast cranked or the black point clipped or some other flaw that makes them difficult to correct.

But as I said I am not the target audience for lab scans. I want the maximum control over my image and that’s why I generally do my own scans. So what I need from a lab scan could be very different from the majority of your customers.

 

[wiltw]

But what pains me is the color workflow which even with NLP is pain.

The above statement I believe to be the one of the significant considerations to the 'scanner' vs 'camera' debate in conversion of film image to digital file.
In the case of color transparency, the camera is a fine solution that I personally consider to be largely equivalent to each other, with the scanner resolution vs. camera resolution being the variable in the final consideration. Let's think this one through...

• If the scanner operates at effectively 4800 dot per inch scan resolution, the 24 x 36mm image area (transparency not in slide mask) translates to 4535 x 6803 pixels, or 30.85 MPixels..
• If the scanner operates at effectively 9600 dot per inch scan resolution, the 24 x 36mm image area (transparency not in slide mask) translates to 9070 x 13606 pixels, or 123.4 MPixels.

So the 32 MPixel camera is better than the 4800 dpi scanner, but worse than the 9600 dpi scanner, in terms of image resolution.

If we switch over to the color neg, the OP comment about color workflow is a consideration, but I think what can be a far greater pain is the accuracy of the color conversion in the neg-to-positive image conversion, something for which NLP is often suggested for ease of conversion.

I took a photograph of a color neg (originally shot in 1981), and for which I have an actual print from that year to compare digital conversion.

1. I took a photo of the same neg with my Canon 7DII camera and daylight-balanced illumination source, then tried the inversion software contained within Corel Paintshop Pro....the process illustrated
2. I used a Canon scanner with the supplied software for neg-to-positive conversion during the scan process and file creation.

the final outcome...

Someone on POTN later posted their effort to convert my first image (above) to positivie, and the result was indeed better than what I achieved above. Unfortunately the demise of POTN prevents me from posting a link to the NLP conversion that was done.

What using the Canon 8800F scanner with supplied EX Navigator software's neg-to-positive scan mode does...

...was superior to even the NLP conversion that someone posted!

 

[David Mackovic]

The above statement I believe to be the one of the significant considerations to the 'scanner' vs 'camera' debate in conversion of film image to digital file.
In the case of color transparency, the camera is a fine solution that I personally consider to be largely equivalent to each other, with the scanner resolution vs. camera resolution being the variable in the final consideration. Let's think this one through...

• If the scanner operates at effectively 4800 dot per inch scan resolution, the 24 x 36mm image area (transparency not in slide mask) translates to 4535 x 6803 pixels, or 30.85 MPixels..
• If the scanner operates at effectively 9600 dot per inch scan resolution, the 24 x 36mm image area (transparency not in slide mask) translates to 9070 x 13606 pixels, or 123.4 MPixels.

So the 32 MPixel camera is better than the 4800 dpi scanner, but worse than the 9600 dpi scanner, in terms of image resolution.

If we switch over to the color neg, the OP comment about color workflow is a consideration, but I think what can be a far greater pain is the accuracy of the color conversion in the neg-to-positive image conversion, something for which NLP is often suggested for ease of conversion.

I took a photograph of a color neg (originally shot in 1981), and for which I have an actual print from that year to compare digital conversion.

1. I took a photo of the same neg with my Canon 7DII camera and daylight-balanced illumination source, then tried the inversion software contained within Corel Paintshop Pro....the process illustrated
2. I used a Canon scanner with the supplied software for neg-to-positive conversion during the scan process and file creation.

the final outcome...

Someone on POTN later posted their effort to convert my first image (above) to positivie, and the result was indeed better than what I achieved above. Unfortunately the demise of POTN prevents me from posting a link to the NLP conversion that was done.

What using the Canon 8800F scanner with supplied EX Navigator software's neg-to-positive scan mode does...

...was superior to even the NLP conversion that someone posted!

Thanks for lenghty reply.
I know that my issue ( color workflow) is not related to camera per say. You are correct that it is mainly about conversion. At this point why i was considering the lab scanner in the first place was the built in color workflow. Ideal state would be sharp scan from camera + colors from lab scanner. Cannot have both right ?

Thanks
D

 

[brbo]

Ideal state would be sharp scan from camera + colors from lab scanner. Cannot have both right ?

Do your customers pick your lab because you provide the sharpest scans (I guess you really mean scans with the highest resolution)?

 

[Steven Lee]

Ideal state would be sharp scan from camera + colors from lab scanner. Cannot have both right ?

I would twist this a bit and suggest the trio of [scanning time, resolution, colors]. You can have only two. BTW, I was surprised to hear about your experience with SmartConvert. To me the difference vs NLP is night and day, and it is also better than auto-everything Noritsu scans from my local labs. If I wanted to open a competitor I wouldn't hesitate to pair SmartConvert to a mirrorless camera for high volume scanning. Perhaps its output varies based on the combination of a light source and a camera?

 

[David Mackovic]

Do your customers pick your lab because you provide the sharpest scans (I guess you really mean scans with the highest resolution)?

I think so, there are many labs out there but they are not doing big sharp files. And also, im very local so my customers still have other ways to go but they are more stuck with me because im local to them haha. Im just trying to bring all out best results for them thats all. Thats why i have started this small operation in first place. Good clean dev + scan. People reaching to me because their films got scratched and unevenly developed in autolabs.

I would twist this a bit and suggest the trio of [scanning time, resolution, colors]. You can have only two. BTW, I was surprised to hear about your experience with SmartConvert. To me the difference vs NLP is night and day, and it is also better than auto-everything Noritsu scans from my local labs. If I wanted to open a competitor I wouldn't hesitate to pair SmartConvert to a mirrorless camera for high volume scanning. Perhaps its output varies based on the combination of a light source and a camera?

Yes I was surprised also. I had both NLP and SmartConvert side by side and from one exact scan i got two conversions which were both almost identical. But yes, with those keys the operation is much much smoother.

Thanks
D

 

[Prest_400]

I would twist this a bit and suggest the trio of [scanning time, resolution, colors]. You can have only two. BTW, I was surprised to hear about your experience with SmartConvert. To me the difference vs NLP is night and day, and it is also better than auto-everything Noritsu scans from my local labs. If I wanted to open a competitor I wouldn't hesitate to pair SmartConvert to a mirrorless camera for high volume scanning. Perhaps its output varies based on the combination of a light source and a camera?

Interesting. Jumping in the thread as I am interested on the workflows for personal use and 120 (645 to 6x9). For B&W funnily I just don't scan aside of a quick "contact" of a whole roll in sleeve, or just workprints, doing mostly darkroom. But color I have gotten the bug to shoot more during the upcoming couple years and plan to take a long trip aka perhaps having 50 rolls EOY to process and scan. Neg, slide and B&W as optional.

I have postponed camera scan ad nauseam because the 1 roll per couple months average is fine sent to a lab. Previously I used a UK lab that did TIFFs out of a Noritsu, now found one with a Frontier that also outputted TIFFs; both without paying crazy amounts on the scan side... Because almost all labs will surcharge the extra resolution and Tiff. Jpegs are fine but most often I have seen compression artifacts, and in Medium Format one wants bigger everything and not a puny tiny scan. Then, Frontier indeed gives quite some interesting (incl. some pleasing coloration) color, while Noritsu sometimes is "warm, rusty blown". So it's all around the place as we have discussed.

(Own rambling) I did test with my EM5 and aside of the dust, it did much nicer work on a 35mm slide than the Epson flatbed. Have a repro stand in the community darkroom and also an Epson Flatbed. That might be replaced to a V700 which could do a much better job for medium format and can be a no cost source; just set it to scan while I print in the darkroom too.
I'm thinking to just get a second hand m43 camera with HiRes mode so it 1-Allows for full color information (pixel shift vs Bayer) and 2-Higher output resolution. That choice is merely because I have already used the system, otherwise Sony I have seen good work and results. Including NLP/filmomat

 

[brbo]

I'm thinking to just get a second hand m43 camera with HiRes mode so it 1-Allows for full color information (pixel shift vs Bayer) and 2-Higher output resolution.

That "pixel-shift does away with Bayer" is a myth.

(I have E-M5II with pixel-shift)

 

[Prest_400]

That "pixel-shift does away with Bayer" is a myth.

(I have E-M5II with pixel-shift)

I'm about the laziest digitalist and recall a listing for an EM5II on sale where the seller mentioned that he was not satisfied with the color depth of this camera. Do regret not picking one for about 300€ pre pandemic as now the bodies are as well as old but still more expensive.

But still Color wise, Bayer and color space are also something to count. I recall reading around that Frontiers do 8 bit, though I don't know how the CCD's capture.

 

[brbo]

E-M5II is 12bit. I picked mine at the start of Covid for 300 EUR with a 30/3.5 macro and honestly wouldn't pay that price today let alone more...

Don't know about Frontiers, but Noritsus are 12bit.

 

[Steven Lee]

That "pixel-shift does away with Bayer" is a myth.

What do you mean? In my case I just couldn't get it to work due to my inability to stabilize the camera due to constant micro-jitter of the house I live in.

 

[brbo]

What do you mean? In my case I just couldn't get it to work due to my inability to stabilize the camera due to constant micro-jitter of the house I live in.

People say that pixel-shift negates the drawbacks of Bayer sensor. I’ve noticed no such benefit, there’s just as much colour crosstalk when using pixel-shift as without.

My findings are limited to the only digital camera I have (Olympus E-M5 II). It would be nice if somebody would demonstrate those benefits (like how pixel shift raw files do not need to go through debayering) on any other camera since this could steer people interested in camera scanning to equipment that is indeed better suited for the task…

 

[Steven Lee]

@brbo Heh, maybe you have the same issue as I have! But other folks had success, like Jim Kasson from dpreview. He has other blog posts where he compares its effect on film scanning as well. IIRC his advice was not to expect noticeable resolution bump, but more realistic grain appearance. The basic principles of pixel shift are solid IMO, given a stable platform I see no reason for it not to work.