Scanning Geekery!

[PMN]

Hi all!

A while ago when Rob and I acquired our drum scanner there was mention of a comparison thread between it and a conventional scanner. As I have nothing better to do on this wet and dreary Sunday afternoon than sit in my office sulking feeling ill with the cold from hell (and because I'm a massive geek who enjoys playing with scanners) I thought I'd do a comparison thread!

I've tried to make this as fair a comparison as possible, none of the images have had any sharpening or anything other than basic standard correction when scanned and they were all resized using the same process. The only difference is the drum scans are 2000dpi and the V500 ones are 2400dpi, this is to keep resolutions native and avoid interpolating and other processing but hopefully it will still give a good idea of what the differences look like.

First up, one of Rob's images shot on Provia. Rob reckoned it could do with being 1/2 a stop brighter and having scanned it I now agree, nonetheless the drum scanner's been able to pull quite a lot more detail out than the V500.

The full size image straight out of the V500:

And straight out of the drum scanner:

A 100% crop from the V500:

And 100% from the Howtek:

 

[PMN]

Next up, another of Rob's from New York shot on Acros. I adore this film, it scans spectacularly well and the contrast is sublime. A real joy to work with!

Full size V500:

Full size Howtek:

The differences don't really show too much on something like Acros if you're resizing down to web kind of sizes, but when you dive right in and look closer the differences are nothing short of staggering. This is the first time I've actually seen side by side 100% comparisons between the two and I almost spat my tea over the drum scanner. (Don't worry Rob, I didn't actually...)

100% from the V500:

And 100% from the Howtek:

 

[PMN]

Another Acros shot from Rob. To be honest I'm really quite amazed he managed to take so many sharp images, I credit the Fuji 690…

V500:

Howtek:

100% from the V500:

100% from the Howtek:

 

[PMN]

The last one for now, a frame of Agfa Vista 200 shot on the Olympus OM20 + 50mm f/1.8.

Full size V500:

Full size Howtek:

100% V500:

100% Howtek:

I've been spending a huge amount of time in between work getting to grips with this scanner and I can honestly say it's changed the way I look at film. I was pretty happy with the results I was getting with the V500 but when I started finally getting good results with the drum scanner it made me realise just how much detail there actually is in film, and also just how beautiful it is when you start extracting that detail. A 2000dpi scan of a 35mm frame comes out at around 5-6 megapixels, that doesn't sound a lot compared to what digital cameras output these days but when you consider it's 5-6 megapixels of pure detail you start to realise that's actually more than enough for most things. The Howtek has a max resolution of 4000dpi and I honestly can't imagine I'll ever need to use it, most things could even probably be scanned at less than 2000, especially for web use.

I'll post more examples if anyone's interested in seeing them, for now I'm gonna get back to my geekery! Well, via the kettle for another pint of tea...

 

[freecom2]

Interesting. If anything, it shows how impressive the V500 is - considering the much smaller size, lower cost, no faffing about with mounting on a drum, and ability to scan several frames at once.

 

[PMN]

Interesting. If anything, it shows how impressive the V500 is - considering the much smaller size, lower cost, no faffing about with mounting on a drum, and ability to scan several frames at once.

For web use and anything smaller than around 1200 pixels I really don't think there's much at all in it, at least for negative film. Even at small sizes the drum scanner can pull out much more useful contrast detail than the V500 ever will on slide film but for neg film and smaller edits the V500 performs very well!

 

[desf]

Interesting. If anything, it shows how impressive the V500 is - considering the much smaller size, lower cost, no faffing about with mounting on a drum, and ability to scan several frames at once.

Agreed. I thought, wow those are good while scrolling down. Then seeing the drum scanner pics you can see the improvement but not until I saw the crops did appreciate just how much better the drum scanner is.

The V500 is a great product for the cash.

Great stuff.

 

[Steve Smith]

shot on Acros. I adore this film, it scans spectacularly well and the contrast is sublime. A real joy to work with!

I agree. I have just started using it (again) in 120. Even on my Epson 4870, it scans very well.

However, your drum scans show how much detail there is in the film which is usually lost in scanning and most methods of printing.


Steve.

 

[steveo_mcg]

Not sure I'd use a drum scanner for b&w I think I'd like to wet print anything I'd like to do properly but the slides, jings!

 

[excalibur2]

For web use and anything smaller than around 1200 pixels I really don't think there's much at all in it, at least for negative film. Even at small sizes the drum scanner can pull out much more useful contrast detail than the V500 ever will on slide film but for neg film and smaller edits the V500 performs very well!

Well done thank you...for me I would like to see how a Asda scan (Fuji Frontier) 1800 X1280 scan compares against the drum scanner at this scan...for detail only, as it would either prove that I'm talking a load of codswallop OR I'm right in saying "it matches a V750\v500 set at best, for detail"............maybe you have some other spare time. :shrug:

 

[PMN]

Well done thank you...for me I would like to see how a Asda scan (Fuji Frontier) 1800 X1280 scan compares against the drum scanner at this scan...for detail only, as it would either prove that I'm talking a load of codswallop OR I'm right in saying "it matches a V750\v500 set at best, for detail"............maybe you have some other spare time. :shrug:

I don't really follow! Are you asking me to scan one of your frames so you can compare it to your own Asda scan?

 

[excalibur2]

I don't really follow! Are you asking me to scan one of your frames so you can compare it to your own Asda scan?

erm well yes, and I would gladly post a strip of negs to choose one, but only it would be of some interest to you, and you think the members would be interested (maybe a poll)..and maybe at the same time you can wave the film flag to show how good a 35mm neg can be with crops, as the negs should be of very good quality on Reala 100 ISO film and a flek lens...

 

[joxby]

That's the problem with scanner tests, everyone needs to scan the same frame if you want to compare everything with everything..

 

[excalibur2]

That's the problem with scanner tests, everyone needs to scan the same frame if you want to compare everything with everything..

But all my negs are scanned by Asda so it doesn't matter what neg is scanned by other scanners, but it would be comparing JPGs though but that should show the difference in detail.

 

[PMN]

I still don't quite follow! Are you wanting me to scan it on both the V500 and the drum scanner so you can compare them both to your Asda scan?

 

[excalibur2]

I still don't quite follow! Are you wanting me to scan it on both the V500 and the drum scanner so you can compare them both to your Asda scan?

Well no unless you want to, as you have already compared a drum scan to a V500..it would be Asda scan ver a drum scan at 1800 X 1280 just to see what detail a Fuji frontier can get off the neg. and I can also scan with a V750 if there is some interest among the members. But as you know with low scans on 35mm you can't do large crops, but what I've found setting the V750 to say 4,800 dpi the shot looks better because of less pixel breakup but the detail is the same as the low Asda scan. :shrug:

 

[PMN]

I'm up for doing that as a comparison, I'm not quite sure how we'll post the results at 1800 for comparison with the forum's max allowable size of 1024 but I'm sure we'll find a way. The only thing is I'd need the frame cutting down so it's just an individual frame on it's own, if you're okay doing that then I'll fire you a PM with my address and I should be able to do it towards the end of the week.

 

[joxby]

But all my negs are scanned by Asda so it doesn't matter what neg is scanned by other scanners, but it would be comparing JPGs though but that should show the difference in detail.

Well I wasn't just thinking about what Brian wants tbh.

We had a scanner comparison thread but its only really useful if everyone scans the same neg.

Anyway, your ASDA scan is not gonna cut the custard, and its not just about scan res, the majority of peeps who take their films to asda for a scan do not have the appreciation for scans/pictures/prints that we do, most of them are happy with something they can recognise, asda will cater to that market, not TPF's crash and burn critique.
There is no doubt that the Noritsu equipment they use is very capable of top quality scans, but you don't get top quality unless the GP demand it and are willing to pay for it.

 

[freecom2]

Additionally, the post-processing sharpening is going to skew the results of any comparison.

 

[PMN]

Additionally, the post-processing sharpening is going to skew the results of any comparison.

Yes it is and this has been my point about high street scans for years, they're usually sharpened to high hell and back but I'm willing I'm willing to do a comparison just for fun!

 

[ChrisR]

This is fascinating, Paul. The Acros and Vista shots show how much more detail can be extracted with the drum scanner; as you say that's not particularly important if you're viewing on a computer or via the web. I guess it becomes important at some point when printing; could you guess at what sort of size that extra detail becomes an issue? (This is complicated I guess, as we tend to view larger images from further away.)

But what particularly struck me was the first shot with Provia. I can't work out the logic of why colour should shift so much in transparencies scanned in different ways, but it confirms my own experience. The drum scanned shot is fabulous!

 

[excalibur2]

Yes it is and this has been my point about high street scans for years, they're usually sharpened to high hell and back but I'm willing I'm willing to do a comparison just for fun!

Ok thanks and I'm sure you can email a jpg file as an attachment .....but folks you are going down the wrong path...it's not about the quality of the Asda scan as I can correct this by reducing contrast, brightness, spots etc in Photoshop..but it's about the detail a Fuji Frontier can pick up off the neg which you can't get using any program after the scan......if it's not there its not there. :shrug:
One thing though, the Asda scan could well clip the highlights and shadows (like you can do with the V750\V500) and that might show up if they do.
So when Paul does the drum scan I can show the difference from my Asda scans, side by side, with some crops to the limit until you get severe pixel breakup, also I can send him my Asda Jpgs if he is interested and wants to play with them at his end.

 

[ChrisR]

?!

 

[joxby]

You are still stuck with the Asda scan as a reference point.
The point is you didn't do it, you've no idea whether its an optimum scan for the neg or the same settings as auntie Mabel's rose photos which was the roll before yours.
You don't know if Asda blew the sky or failed to record details in the blacks or both, all you have is the neg.

 

[PMN]

Brian, I'm not interested playing with scans in PS to try make them look right, I'm interested in what comes straight out of the scanner. Nothing more, nothing less!

 

[s162216]

But what particularly struck me was the first shot with Provia. I can't work out the logic of why colour should shift so much in transparencies scanned in different ways, but it confirms my own experience. The drum scanned shot is fabulous!

Scanners are like monitors, get 10 different models and you'll get 10 different representations of the same image/slide by default, unless you calibrate them all to the same standard where they will be pretty close to each other.

 

[abbandon]

Interesting. If anything, it shows how impressive the V500 is - considering the much smaller size, lower cost, no faffing about with mounting on a drum, and ability to scan several frames at once.

There are drum scanners and then there are drum scanners but just about all have the ability to batch scan.

Back in the day imaging bureaus ran these machines 24 /7 with three shifts the most productive drum scanners general being those with larger drums even with the small drum on a D4000 you could mount effectively a whole 35mm roll and scan the lot.

 

[robhooley167]

I got exactly the same colour shift when I scanned it on my V500 about a year ago, using Vuescan rather than the Epsonscan that Paul used on his V500.


Sunset At Kimmeridge by Rob Hooley, on Flickr

 

[excalibur2]

You are still stuck with the Asda scan as a reference point.
The point is you didn't do it, you've no idea whether its an optimum scan for the neg or the same settings as auntie Mabel's rose photos which was the roll before yours.
You don't know if Asda blew the sky or failed to record details in the blacks or both, all you have is the neg.

Well the machine is set by the engineers for all supermarkets using the Fuji Frontier and am not sure what the operator has to do other than feed the neg into the machine....if they can dev a neg, scan to a CD, do a separate Index, neatly put all negs in sleeves in correct order, in 25mins they can't be doing much..and would say a lab can't make it pay if they scan every individual frame, of 36 exp, to perfection at business costs of £40/per hour, as most people would go elsewhere.
Anyway Paul's drum scan should show the difference and if he wanted to, could select his own negs scanned by a well known lab and compare them. :shrug:

 

[excalibur2]

Brian, I'm not interested playing with scans in PS to try make them look right, I'm interested in what comes straight out of the scanner. Nothing more, nothing less!

That's what I'm saying...it's just comparing the detail from the neg, but it's been side tracked by people saying the quality from Asda is carp.

 

[joxby]

Its not sidetracked, you have unwavering tunnel vision.

 

[freecom2]

There are drum scanners and then there are drum scanners but just about all have the ability to batch scan.

Back in the day imaging bureaus ran these machines 24 /7 with three shifts the most productive drum scanners general being those with larger drums even with the small drum on a D4000 you could mount effectively a whole 35mm roll and scan the lot.

Several frames at once - with very little faff required for each strip. No mounting fluid, no drums. Drum scanners are superb, no doubting their ability at all, but as a scanner to have in your home office? The V500 certainly holds its own.

That's what I'm saying...it's just comparing the detail from the neg, but it's been side tracked by people saying the quality from Asda is carp.

No, just pointing out that additional PP (that is out of our hands) adds a further complicating layer that makes comparison even more difficult.

 

[robhooley167]

Several frames at once - with very little faff required for each strip. No mounting fluid, no drums. Drum scanners are superb, no doubting their ability at all, but as a scanner to have in your home office? The V500 certainly holds its own.

The V500 certainly holds it's own. that's not being disputed.

Drum scanners aren't completely unfeasible to run in a home office, which is what I was planning to do with it originally.
Mounting doesn't necessarily have to involve fluid or expensive tape and I reckon (however I'm a little out of practice on both of them at the moment) that it's possible to mount a strip of negatives in the drum scanner just as quickly as loading up the V500.

 

[excalibur2]

No, just pointing out that additional PP (that is out of our hands) adds a further complicating layer that makes comparison even more difficult.

Not so as Paul has shown comparing his drum scan to the V500, if he got a perfect scan in one go from his V500 without any further adjustments, that shows he knows what he is doing......then he compared the DETAIL difference in the crops.
An Asda scan whether it's over sharpened or too bright or excessive vivid colours and so on, will still only show a basic detail of things in the shot depending of the ability of the Fuji Frontier scanning optics...of course you would have to turn the brightness down etc (if necessary) to show things, but that applied in the old days of chemical printing and in this old example it also depended on your enlarging lens for details in the print...carp in carp out.

Of course I could be talking a load of codswallop and will admit to this if I'm wrong as I seem to be in a minority of one

 

[Steve]

Thanks for this thread and taking the time to make the comparisons....very interesting.

 

[excalibur2]

Its not sidetracked, you have unwavering tunnel vision.

That's not fair as I've dev my negs and produced B\W prints to 16X20, done my own colour dev and printing for years and own 2480, 4180 and V750 pro scanners..well I did offer the 2480 as a freebie to a member but he never came back to me, so I threw it away. :shrug:

 

[PMN]

But what particularly struck me was the first shot with Provia. I can't work out the logic of why colour should shift so much in transparencies scanned in different ways

Colour correction is a little quicker and easier on the DPL software that drives the drum scanner than it is with Epsonscan so it's a bit easier to get good results 'straight out of the box' on the drum scanner, I must admit though I was rather surprised at the differences between the two straight scans!

but folks you are going down the wrong path...it's not about the quality of the Asda scan as I can correct this by reducing contrast, brightness, spots etc in Photoshop..but it's about the detail a Fuji Frontier can pick up off the neg which you can't get using any program after the scan......if it's not there its not there. :shrug:

No we're not.

The fact you're talking about improving "quality" by reducing contrast, brightness or other bits of fiddling in post-processing means there are problems with the original scans that shouldn't exist, those problems exist because they've been run through a machine with automatic settings rather than being scanned by an actual human being who genuinely cares how the final image looks.

Correcting things like blown highlights on scans doesn't work for one very good reason - you may have started with an analogue image but once scanned you're in the digital realm. Presumably the scans you get back from Asda are 8 bit JPEG? If so then trying to pull back highlights on them is exactly the same as trying to recover blown highlights on a digital camera shooting JPEG - it just can't be done, at least not without looking flat and horrible.

Scanning should really be thought of in exactly the same way as if you were using a digital camera because you're turning the analogue film image into exactly the same type of file (JPEG, TIFF, etc) and you're limited by everything that makes digital such an unforgiving medium. If you shoot digital and screw the exposure there's very little chance you're going to be able to recover it to any high standard and exactly the same applies to scanning film, if you get the 'exposure' wrong on a scan you're either not using enough information in the digital file or you're trying to use too much. It has to be just right to get the best results.

Anyway Paul's drum scan should show the difference and if he wanted to, could select his own negs scanned by a well known lab and compare them. :shrug:

I don't have any of my stuff scanned elsewhere. I was always happy with my results from the V500 (after a few weeks of learning it's ins and outs) and now we have the drum scanner there's even less reason to get stuff sent off to be scanned.

An Asda scan whether it's over sharpened or too bright or excessive vivid colours and so on, will still only show a basic detail of things in the shot depending of the ability of the Fuji Frontier scanning optics...of course you would have to turn the brightness down etc (if necessary) to show things, but that applied in the old days of chemical printing

Again you're looking at it slightly wrong, you're comparing a purely analogue process to a digital one and they're simply not the same thing. Scanned images from film need to be thought of in exactly the same way as those from a digital camera because you're dealing with the same kind of files that have the same technical limitations. Turning the brightness down on a scan in order to compare it to others won't give you a fair comparison, they need to be processed similarly when they're actually scanned so they're all making the most of the information available in whatever the output format is.

To be honest I'm not really sure me drum scanning one of your frames is going to be useful in your comparison because (to me at least) the logic of the comparison is flawed to start with.

 

[TheTimeChamber]

I have to day, the difference in the details is incredible! I am very pleased with the few scans you did for me.

Drum Scan (which I put through lightroom, and forgot to sharpen lolz):


Drum Scanner Ales by Sectionate, on Flickr

Commercial Scan (with god knows what processing)


Beer please by Sectionate, on Flickr

I should have sent up a frame or two from a 35mm roll to compare them against my dedicated 35mm scanner.

Edit: looking at them side by side, the commercial scan has had loads down to it...

 

[PMN]

Nick, I'm really glad you're happy with the results. Thanks for the feedback!

If you want to fire a few 35mm frames up to me then feel free.

 

[excalibur2]

Colour correction is a little quicker and easier on the DPL software that drives the drum scanner than it is with Epsonscan so it's a bit easier to get good results 'straight out of the box' on the drum scanner, I must admit though I was rather surprised at the differences between the two straight scans!




No we're not.

The fact you're talking about improving "quality" by reducing contrast, brightness or other bits of fiddling in post-processing means there are problems with the original scans that shouldn't exist, those problems exist because they've been run through a machine with automatic settings rather than being scanned by an actual human being who genuinely cares how the final image looks.

Correcting things like blown highlights on scans doesn't work for one very good reason - you may have started with an analogue image but once scanned you're in the digital realm. Presumably the scans you get back from Asda are 8 bit JPEG? If so then trying to pull back highlights on them is exactly the same as trying to recover blown highlights on a digital camera shooting JPEG - it just can't be done, at least not without looking flat and horrible.

Scanning should really be thought of in exactly the same way as if you were using a digital camera because you're turning the analogue film image into exactly the same type of file (JPEG, TIFF, etc) and you're limited by everything that makes digital such an unforgiving medium. If you shoot digital and screw the exposure there's very little chance you're going to be able to recover it to any high standard and exactly the same applies to scanning film, if you get the 'exposure' wrong on a scan you're either not using enough information in the digital file or you're trying to use too much. It has to be just right to get the best results.




I don't have any of my stuff scanned elsewhere. I was always happy with my results from the V500 (after a few weeks of learning it's ins and outs) and now we have the drum scanner there's even less reason to get stuff sent off to be scanned.



Again you're looking at it slightly wrong, you're comparing a purely analogue process to a digital one and they're simply not the same thing. Scanned images from film need to be thought of in exactly the same way as those from a digital camera because you're dealing with the same kind of files that have the same technical limitations. Turning the brightness down on a scan in order to compare it to others won't give you a fair comparison, they need to be processed similarly when they're actually scanned so they're all making the most of the information available in whatever the output format is.

To be honest I'm not really sure me drum scanning one of your frames is going to be useful in your comparison because (to me at least) the logic of the comparison is flawed to start with.

Oh well no one can see my point of view, and Paul as no one is interested except me it's too must faffing around and thanks for considering my suggestion\interest...but you could do one thing is to rescan your neg to 9600dpi or the max on your V500 and see if it make any difference in the details of your shot.