Drum Scan Forensics

[JW Dewdney]

Hi all. Well - today I just received close to $1K worth of drum scans from a very well-known service bureau, supposedly somewhat renowned for their quality scans. Out of the interest of protecting them, in case this is a simple mistake or misunderstanding - I'm not going to name names. At any rate - I was just going over the scans, and was a little bit appalled at what APPEARS to be poor quality.

So- the POINT of this post is to determine what factors one can use to determine whether or not a scan is a QUALITY scan. The criteria I used to determine these were 'poor' scans were the following;

1. Colour Balance - surprised to find it pretty far off!!
2. Blue channel is terrifically noisy in some of the scans. It goes BEYOND grain - it's actually 'chunky'..!!
3. Histogram is spiky as hell - previous drum scans had super dense, rich histograms, with seemingly good bit depth.
4. They simply don't look that great - to me. I've made C-prints off the same film - and the result was superior.

Is it possible that I'm in the wrong in terms of my assessment? Why? Any suggestions? I think I would agree with the criteria cited in the scanner shootout results posted on this site (you know the one, right?). Can anyone make a good suggestion as to what good evaluative criteria might be otherwise?

Thanks much for your time reading this!
J.

[Frank Petronio]

The local service I used recently lost their best person and I spent the afternoon "training" the replacement. They probably had some staff problem too, that is the most likely scenario.

By all means complain!

[JW Dewdney]

from the scans - it looks like the PMT for the YELLOW channel is dying or something. But I don't think that actually HAPPENS with PMT scanners. Just CCDs... wouldn't you think that the fact that the blue channel is extra noisy would be good evidence that a CCD flatbed was actually used (well -that's my theory thus far)...? Good point frank - but regardless - I'm just trying to figure out if, in fact, my judgement is on the mark first...

[George Kara]

I strongly suggest buying a good used scanner and doing it yourself. I purchased a scitex eversmart pro for around 3k including software.

I process everything using silverfast hdr and purchased the silverfast book which is effective for color management.

I can get it right pretty well every time and it really isnt that difficult to do.

[JW Dewdney]

I strongly suggest buying a good used scanner and doing it yourself.

Well- I use a heidelberg scanner with silverfast personally. it's fine for most stuff. I think I can pull a better result than pretty much any semi-pro scanner (esp. from a neg). But I'm doing some 20x24 prints for my portfolio - and I don't want to mess around with this stuff. I need the ABSOLUTE best quality. The lab uses a crosfield - so that SHOULD be plenty to do the job. But anyway...

[Doug Dolde]

Call them up and complain ! The squeaky wheel gets the grease !

[JW Dewdney]

oh - of course I will, Doug. I have no problem with that.. I'm just trying to figure out what the source of the problem was first - so that I'll be in a better position to deal with it when I go talk to them tomorrow.

[Gordon Moat]

The lab I use for scanning, and a service bureau I sometimes use, would both do the scans over at no additional charge if I indicated to them a problem with the scans they did for me. Noise in only one channel sounds like either a problem with one PMT, or getting the settings very wrong on the scanner. I would simply ask them to do the scans over, while explaining what you felt was wrong.

Ciao!

Gordon Moat
A G Studio (http://www.allgstudio.com)

[George Kara]

I assume you dont mean you own a tango. I suggest you take a look at the excellent scanner review comparison on the LF page. The crossfield seems like a decent scanner in comparison but doesnt outperform the other pro-level products.

There is a great kodak/creo/scitex on e*** right now with an asking price of around 1350. This is one of the best scanners on the market.

Scanning is like managing your own money. Nobody will do a better job than you.

[Kirk Gittings]

JW, Just curious:


But I'm doing some 20x24 prints for my portfolio

Is this a commercial portfolio? Why are you doing portfolio prints so large? I don't use a portfolio any more (haven't needed one since websites), but when I did I wanted it to fit in a Fed Ex box.

[bob carnie]

Phone the owner of the lab, he/she will/ should be concerned. 1k of scans is a very good order for any service company and if you are not happy you are the client.
I would be extremely suprised if they did not give you immediate attention and work to solve the issues.


Well- I use a heidelberg scanner with silverfast personally. it's fine for most stuff. I think I can pull a better result than pretty much any semi-pro scanner (esp. from a neg). But I'm doing some 20x24 prints for my portfolio - and I don't want to mess around with this stuff. I need the ABSOLUTE best quality. The lab uses a crosfield - so that SHOULD be plenty to do the job. But anyway...

[JW Dewdney]

I assume you dont mean you own a tango.

No - it's a umax derivative.. a 1450 or something (i forget). Personally - I'd pit a crosfield or a big Linotype-Hell against pretty much ANY modern scanner. There's no substitute for photomultiplier tubes attached to a good power supply in my book.

[JW Dewdney]

JW, Just curious:
Is this a commercial portfolio? Why are you doing portfolio prints so large? I don't use a portfolio any more (haven't needed one since websites), but when I did I wanted it to fit in a Fed Ex box.

Yes, Kirk. My arch. stuff. I'm trying to work a different angle and show them a level of detail they've never seen before - kind of push the 'film advantage'. I normally prefer to walk in with work myself - although this size is sort of unwieldy. It remains to be seen how I'm going to 'handle' pieces this big.

[JW Dewdney]

Phone the owner of the lab, he/she will/ should be concerned. 1k of scans is a very good order for any service company and if you are not happy you are the client.
I would be extremely suprised if they did not give you immediate attention and work to solve the issues.

Oh no, bob. I'm SURE they'll re-do. The scans just seemed really odd to me. There are also some steep cliffs on the black end of the histogram - that would seem to me to be an odd hardware limitation that really shouldn't be there on a PMT scanner.

The scans look pretty good once you clean them up - it's just that I found it quite odd to have to do alot of colour correction to get there (except for the shadows - which are a bit blocked). You'd think they'd target the right range of tones first - and then scan it - seems you should be able to do this on a PMT scanner.

[David_Senesac]

Particularly with drum scans one is at the mercy of the competence of the operator. Over the last decade I've paid for well over a hundred drum scans from several labs. Even within the same lab the quality has occasionally varied considerably. A few scans I've gotten back from well respected labs that usually do great work have been simply mediocre. Probably the most common issues are color balance and foreign matter like dust that requires excessive amount of cloning out work. Notable and importantly, I've never had a drum scan come back that wasn't sharp. Generally for large format any of the various drum scan models are more than capable of excellent output for typical 40x50 inch max print intentions given a good operator As for the operator issue, one can imagine even the most respected labs have a number of people that do scanning and that during times like vacations, before Christmas, and when workloads are high, it is likely some work might be done by those of less experience who don't exactly work like the regular operators.

With the foreign matter/dust issue the problem might be dusty scanning room air, inadequate dusting, and especially dirty reused oil. I dust my transparencies before sending them in so any stuff on the scan file is mainly the result of the lab process. When using a new lab, I always bother to discuss what they do regarding dusting transparencies as I am not of fan of brushes. I'd rather prefer labs use high compressed air like I do. For color and image quality, of course the operator can gang mount several transparencies on a drum at the same time. Thus if each is scanned at the same time, the raw photomultiplier outputs will be the same for all slides instead of being optimized dependent on the actual image. Suspect that is why I have seen dark images come back too dark as though the operator never considered boosting the sensitivity a bit to compensate for the dynamic range. With 8-bit scans I've tended to ask for, that of course is more an issue to later post processing work though has never been that big a deal to ask for a re-scan. I tend to make labs aware I expect decent color balance and neutral color. If one does not emphasis that, the operator may try and boost the colors to what many customers prefer. For instance with fall color and sunset shots, extra saturation, reds, and purples. The better labs like WCI make clear the resulting drum scan file will be far from print ready, usually somewhat dark, color neutral, and flat so customers can do those adjustments in Photoshop themselves. ...David