Jump to content

  •  

CNers have asked about a donation box for Cloudy Nights over the years, so here you go. Donation is not required by any means, so please enjoy your stay.

Photo

Home scanning - yes or no?

This topic has been archived. This means that you cannot reply to this topic.
28 replies to this topic

#1 Michal1

Michal1

    Viking 1

  • -----
  • topic starter
  • Posts: 598
  • Joined: 25 Jul 2010

Posted 13 October 2010 - 09:35 AM

Is it better to make home scans or let our films scanned in a photo lab? When doing the home scanning one holds the process in their hands and it's for free no matter the resolution or bit depth of the output file. On the other hand, I've found several arguments contra which appears frequently on the web:
1) Any flatbed scanner can't reach the quality of a specialized film scanner. Only the most expensive ones are almost comparable.
2) You can consider your home scans to be good but only until you compare the result side by side with a better scaner.
3) It's very complicated to set the scanner's control properly. It takes a lot of practices. A photo lab staff gets it after several months of daily use.
4) With my shooting rate about 50 photos per year is the home scanning simply economically disadvantageous. The money paid by a medium grade scanner Epson V700 would be enough to pay the next 24 years of the photo lab scanning.

Because of these reasons I'm not sure about purchasing a scanner. Did you ever tried to do a comparison between a flatbed and film scanner for the astro aplication? Is it sufficient to get 16-bit/channel TIFF file from a photo lab? Is it sufficient to get a 8-bit/channel scan even if it's resolution is so hight that the grain or the film is cleary visible?

I'm looking forward to your opinions. Remember the argument 2) ;)

M.

#2 Dave Kodama

Dave Kodama

    Mariner 2

  • -----
  • Posts: 226
  • Joined: 04 Oct 2010

Posted 13 October 2010 - 10:15 AM

Michal,

I'm glad you started this thread. I would like to hear other opinions and get up to date as to what modern scanners are good or bad.

For myself, I made the decision over 10 years ago, when a good scanner was much, much more expensive than it is now. But I decided that commercial scanning (even by an astronomer friend with a commercial photo business) was not going to be good enough for the long term. To me, scanning is a very important part of the chain of post processing. A weak link here, means the whole chain is at risk. I certainly wouldn't be happy with having someone do the final processing of my astrophotos, and I'm not satisfied having someone else do my scanning.

It's true that I don't have a side-by-side scan comparison with a professional scanner (I don't know what that is today), but I do know that the professional scans I had done were not satisfactory. I'm aware that drum scanners have a greater dynamic range (good for underexposed transparencies), but the operator must be used to processing astrophotos. I think the problem here is that astrophotos need *individual* adjustment to boost contrast in some ranges and not in others. For example, you want to have high contrast in a comet tail or faint nebulosity, but not at the high end in the cores of stars.

While it might be true that the photo lab's technician would have more experience with the equipment, he will almost certainly be processing just normal daylight photos. It is certainly true that the scanner's settings can be complicated to set, but the technician probably is not familiar with how to set it for an astrophoto. This is just the same as using Photoshop. I know many people who are experts in using Photoshop, but they don't know how to process an astrophoto because that is not their specialty. Also, you don't always know if the same guy is processing your scans.

Having said all of the above, I do think it's easy to get the wrong scanner. The software is a critical component of the scanning process. If it doesn't allow you to get 16-bit/color scans, then it is worthless. That is the case of the (old) first 35mm scanner I got. It was a Nikon scanner which does fine (even today) to scan daylight photos, but it does not output 16-bit/color and does not allow adequate adjustment prior to the scan. By the way, this scanner is also stuck on my Windows 98 computer because of lack of driver support. Unfortunately choosing a scanner is a hit-or-miss process and most of us can't afford to buy a scanner as an experiment. This is why I'm glad you started this thread and I'm interested to hear what everyone is using now.

For myself, I'm using an Agfa T2500 scanner which is 10+ years old. Its resolution is adequate, but not great (2500dpi), and its driver has not been updated by the company, so I'm still using it with an old Windows 98 laptop PC and praying it does not break down! I also have as a "backup" and a normal scanner for scanning paper documents, an Epson 4450 scanner. I have only done a few astrophotos as a test on it so I don't have much experience with it, but it does have a modern (XP) driver and can output 16-bits/color and includes film holders for 35mm and medium format (up to 6x9cm) film.

Dave

#3 Dave Kodama

Dave Kodama

    Mariner 2

  • -----
  • Posts: 226
  • Joined: 04 Oct 2010

Posted 13 October 2010 - 10:28 AM

P.S. One more argument comes to mind - Less handling of your film by another person is better. Only we know how much time, effort, and money went into each frame of film!

Dave

#4 WillCarney

WillCarney

    Surveyor 1

  • *****
  • Posts: 1,772
  • Joined: 08 Oct 2009

Posted 13 October 2010 - 11:28 AM

I have a fairly good scanner. It is made to scan negatives but I have never been able to get it too work well enough. It will scan pictures ok but not slides or negatives. I just take them to a local processor and have the negatives put on a CD.

William

#5 Nightfly

Nightfly

    Apollo

  • -----
  • Posts: 1,307
  • Joined: 20 Jun 2007

Posted 13 October 2010 - 04:37 PM

I have used an Epson Perfection Photo 4490 for several years and in August changed over to the Epson V600 model.

I'll let my images speak for themselves as to how good these scanners are. My only comments would be that a flatbed does well for 120 format and so-so for 35mm.

Enlargement will be the limiting factor on these scanners. For 35mm I would suggest 8x10 is the limit. For 6x7, I have produced 16x20's that have been very pleasing. I might be able to get 20x30's but they may appear a little soft.

These are limits imposed by the scanners, not the film!

#6 Dave Kodama

Dave Kodama

    Mariner 2

  • -----
  • Posts: 226
  • Joined: 04 Oct 2010

Posted 13 October 2010 - 05:33 PM

William, what scanner are you using? When you get a CD made, do you get 16-bit/color scans. I think Kodak used to do that.

Jim, what resolution are you scanning at? And does the V600 have glass near the film and between the film and scanner chip? One slight problem I noticed about the 4490 is that light scattered at the glass, softening the transition between a light and dark boundary. The nice thing about my old Agfa scanner is that it doesn't use glass for up to 4x5" scans.

Dave

#7 Nightfly

Nightfly

    Apollo

  • -----
  • Posts: 1,307
  • Joined: 20 Jun 2007

Posted 13 October 2010 - 05:59 PM

Dave,

There is a glass between the CCD and film, like the 4490. I think this is why it only does so well. The V600 also uses a LED light source versus the 4490's white cold cathode flourescent lamp. The CCD sensor is suppose to be better as well. I do see at least a modest improvement in the overall image quality.

#8 Dave Kodama

Dave Kodama

    Mariner 2

  • -----
  • Posts: 226
  • Joined: 04 Oct 2010

Posted 13 October 2010 - 07:03 PM

Jim,

What is your usual scanning resolution? Do you vary it according to the film you shoot?

Dave

#9 Nightfly

Nightfly

    Apollo

  • -----
  • Posts: 1,307
  • Joined: 20 Jun 2007

Posted 13 October 2010 - 07:29 PM

Dave,

At minimum I scan at 1200 Pixels-Per-Inch if I am just going to web post. Sometimes I scan at 2400 and resize to 1200 in PS. This gives me a workable fine without bogging down the PC. I almost always resize to 1280 x 1024 for final posting on the web. This also fits the native resolution of my monitor and makes wonderful wallpaper.

A 1200 scan is like a 9 Meg-pixel camera with the 6x7 frames. The V600 does this without any problems. At 2400 I get what is like a 34 Meg-pixel image. The quality is perhaps not equivelent pixel for pixel as there is some noticeable scanner noise, but then again any digital camera will produce that and more if trying to capture night images itself.

FYI. You may be aware. I make all my images available as a 1280 x 1024 image to fit the native resolution of the monitor. If doing a slideshow or wallpaper there is a significant difference in how the image appears. If I present the larger file as wallpaper of slideshow it is of less quality. Not sure if folks are matching their images to their monitors. It makes a noticeable difference.

For grainy films like Centuria 400, I don't bother scanning above 1200ppi. Daytime images shot with this film can do well above 1200, but astro images don't seem to hold the grain like daytime images.

#10 Dave Kodama

Dave Kodama

    Mariner 2

  • -----
  • Posts: 226
  • Joined: 04 Oct 2010

Posted 13 October 2010 - 07:58 PM

Jim,

Thanks for the info on the scanner and your workflow description. It's very interesting to see what others are doing to get their results.

I'm also wondering about TechPan. I've seen people argue that for TechPan, you need to use 4000DPI as a minimum to avoid aliasing. That will require another big step up for my computer...

For the moment, I'm always scanning at my max res of 2500 dpi. Whether or not that's a good idea... well I've been constantly buying new hard drives. Anyway, my basic idea was that if I scan everything at high-res, it's protection against film deterioration or other disaster that might happen to the original media.

For the web I generally post a shot no more than 950 pixels wide to avoid horizontal scrolling on a medium-sized screen. If the picture warrants it, I've recently decided to also have a link to an "HD" res shot (1920x1080) to fill a big screen TV or PC screen that is also used to view HD videos. This is starting to make a serious dent in my web space allocation, so I sometimes put the actual pictures on another hosting site (smugmug.com). But one of the frustrations of the web/PC continues to be that one can't even look at the full-res of a typical medium format shot, let alone a mosaic created from medium format frames!

Dave

#11 Nightfly

Nightfly

    Apollo

  • -----
  • Posts: 1,307
  • Joined: 20 Jun 2007

Posted 13 October 2010 - 08:15 PM


Since I upload to Flickr, the images are all resized to various sizes. This is not the optimum setting, but I don't want too much high res stuff online anyway.

You are correct of course that it is hard to represent what is actually on the film. The scanner is inadequate, the screen inadequate, etc... Keeping things stricktly analog would be the only way to truly present the images. That means old style wet prints. Not going to happen anymore.

I like my 5:4 ratio screen (1280x1024) It matches perfectly, the aspect ratio of the 6x7 frame. I'd love to upgrade to a 1600 x 1280!!! Never perfect, but always improving.

As for the original transparencies / negatives. One should plan to scan them again someday with a much more capable scanner. I'm not sure what the future holds for scanners, but there is lots of film out there still waiting to be scanned, about a hundred years worth.

#12 Nebhunter

Nebhunter

    Surveyor 1

  • -----
  • Posts: 1,926
  • Joined: 04 Oct 2003

Posted 13 October 2010 - 11:10 PM

I'd also have to say yes for scanning. HOWEVER - do not under any circumstances get a Microtek I800. I have struggled with mine for 2 years. Solid, well built, the Silverfast software that came with it is very powerful. But compression banding galore - and this pulls the stars into egg shapes, and adds wonderful CA colours to them.

The banding can be controlled to a degree - but the band usually ends up running right thru the nebula or star field - the most critical part of the frame.

Once I figure out what to replace it with - I might take it to the shooting range. The ONLY reason to hang onto it - it does a great job on scanning photographs. I would suggest sticking to the Epson line as they seem to work well for the money - new scanners. The Microtek M1 does not use glass to scan film - a dual bed design. But being bitten by the I800, I'm shy of anything from them.

It is a steep learning curve to learn how to use a scanner well. That said - most articles seem to suggest it's the software that makes the scanner work really well to get the most from it.

I'd can't confirm that, as I've rarely been able to get the most from mine other than scanning old photos.

igor

#13 Michal1

Michal1

    Viking 1

  • -----
  • topic starter
  • Posts: 598
  • Joined: 25 Jul 2010

Posted 14 October 2010 - 11:35 AM

I think it would be helpful for everybody, when we all post a sample of a raw unprocessed scanned image from our scanners or photo labs to see the difference. It would be the best to post a picture of some frequently imaged area using a usual equipement and settings, e.g. containing the area between the North America nebula and Gamma Cyg using E200 film and 100mm lens for medium format or 50mm lens for the 35mm format and exposure time about 1 hour. Naturally not everybody has exactly such a photo but better something than nothing. If you can upload a 16-bit file the better. A cropped image doesn't matter.

M.

#14 WillCarney

WillCarney

    Surveyor 1

  • *****
  • Posts: 1,772
  • Joined: 08 Oct 2009

Posted 14 October 2010 - 12:29 PM

William, what scanner are you using? When you get a CD made, do you get 16-bit/color scans. I think Kodak used to do that. Dave


I have a HP Scanjet 3970 which has 2400 DPI.

The photo processor saves the pictures on CD's in JPG 1800x1215, 24bit images.

William

#15 Dave Kodama

Dave Kodama

    Mariner 2

  • -----
  • Posts: 226
  • Joined: 04 Oct 2010

Posted 14 October 2010 - 12:50 PM

William,

I'm not familiar with that scanner and online info isn't very complete. Does it actually have a backlight for scanning film? I recall some very early HP scanners came with a kludgy reflector which was supposed to reflect light from the normal scanning bulb up and behind a piece of film. This was a terrible "feature".

If you can get your photo processor to give you scans in 48-bit (16-bit/color) scans in TIFF format, that would give you a much better starting point for processing your images. You will have better color depth (essential for adjusting color and contrast), and no JPEG compression artifacts.

Dave

#16 Dave Kodama

Dave Kodama

    Mariner 2

  • -----
  • Posts: 226
  • Joined: 04 Oct 2010

Posted 14 October 2010 - 01:00 PM

Michal,

Back in the 1990's the people on the Astrophoto Mailing List (APML) organized a scanner evaluation. They concluded that it was essential to remove the scope, seeing, and film from the testing, so it was decided that a standardized resolution test image should be used. I believe this was a glass slide with a fine pattern deposited in metal. This was passed around to each person in turn along with one color astrophoto slide. It took months to circulate, and I think the test was only partially successful. It was possible to see how various scanners did in resolution, but not in image depth and noise performance for dark slides. And it did not really address the issue of how much control you had via the software provided with the scanner. And of course another problem is that it only tested scanners owned by the various people on the list and not new products.

The evaluation of different scanners is a difficult task, and I don't think posting raw files will help much without being able to look at the original piece of film.

Dave

#17 WillCarney

WillCarney

    Surveyor 1

  • *****
  • Posts: 1,772
  • Joined: 08 Oct 2009

Posted 14 October 2010 - 02:05 PM

William,
I'm not familiar with that scanner and online info isn't very complete. Does it actually have a backlight for scanning film? I recall some very early HP scanners came with a kludgy reflector which was supposed to reflect light from the normal scanning bulb up and behind a piece of film. This was a terrible "feature".

If you can get your photo processor to give you scans in 48-bit (16-bit/color) scans in TIFF format, that would give you a much better starting point for processing your images. You will have better color depth (essential for adjusting color and contrast), and no JPEG compression artifacts.Dave


The scanner does do this. It's useless for film.
I'll ask next time about the files. I have a roll right now that needs processing with Hartley pictures.
William

#18 Rick Thurmond

Rick Thurmond

    Mariner 2

  • -----
  • Posts: 216
  • Joined: 30 Oct 2006

Posted 14 October 2010 - 03:31 PM

I do my preliminary scanning on a Epson 3200. I'd say it doesn't focus too well, and doesn't handle dark or light extremes. It won't resolve anywhere close to 3200 dpi. But it handles medium and large format film. All the dirt on the glass shows up on the scans, and some scanner artifacts creep in too. The Nikon Coolscan V is a much better scanner for 35mm.

For the very best pictures I use a lab that has an excellent drum scanner. The scans I get back from him are perfect: all the resolution the film has, all the dynamic range, excellent color, low noise, and no scanner artifacts. Not all labs that have scanners use them correctly. I've sent films to other labs with drum scanners and an Imacon and gotten back absolutely unusable scans.
Rick

#19 Michal1

Michal1

    Viking 1

  • -----
  • topic starter
  • Posts: 598
  • Joined: 25 Jul 2010

Posted 15 October 2010 - 04:03 PM

Dave, thanks for your hint. Of what you are writing makes sense. Without scanning the same pattern is not possible to do a proper scanner comparison. But I think it's not completely worthless. After seeing the Nightfly's high resolution Summer triangle I was surprised how different it was from the scans I get from my photo lab (Jim, thanks for posting).
I'd like to get some kind of answer to the question in the name of this thread. A good argument for home scanning is that it is recomended by famous astrophotographes, but I'd rather to see something more material/rational.

Ok, here is my contribution. This (10MB !) is an example of a total scan I get from my local lab. It comes from a 35mm film and is scanned at 2400 pixels/inch. At this resolution the E200 film begins to be grainy. This is a cut of it around the Veil nebula (368kB). The scanner produces bright horizontal stripes across the frame. In this lab all the work is made by one guy which allows an individual approach.

I let my only medium format photos scanned in an another photo lab. It's a big photo store. They don't use a film scanner but a minilab. I guess no human have seen the photos during the process of scanning. Although the result looks quite well for landscape photos, in astrophotos the stars are surrounded by dark rings, probably produced by unsharp masking, and the sky background is too high.

I'd very grateful for uploading a 16-bit image. I'd like to try the difference in processing of 8- and 16-bit image.

Rick, do you know the name of the drum scanner working best for you, please?

Michal

#20 Dave Kodama

Dave Kodama

    Mariner 2

  • -----
  • Posts: 226
  • Joined: 04 Oct 2010

Posted 15 October 2010 - 04:33 PM

Michal,

OK, give me a few days. I think I can do something more useful to you. I'll scan the same piece of film on 3 different scanners (Nikon LS1000, Epson 4450, and Agfa T2500). The Nikon only outputs 8-bit/color. The others can output 16-bit/color. I'll crop to a small section and put it up on a page for anyone to download.

It's too bad you live so far away. I could scan one of your shots to give you the best comparison...

Dave

#21 Michal1

Michal1

    Viking 1

  • -----
  • topic starter
  • Posts: 598
  • Joined: 25 Jul 2010

Posted 15 October 2010 - 05:16 PM

Dave, you are very kind.

It's too bad you live so far away. I could scan one of your shots to give you the best comparison...


I'm touched...

#22 Nightfly

Nightfly

    Apollo

  • -----
  • Posts: 1,307
  • Joined: 20 Jun 2007

Posted 15 October 2010 - 05:26 PM

Michal,

You large 2400 scan of Cygnus looks speckled, not grainy. That's noise. That's a pretty good scan minus the noise. I took the liberty of downloading and doing some work to it. If you would like to see let me know and I can post or PM me and I can email to you.

#23 Michal1

Michal1

    Viking 1

  • -----
  • topic starter
  • Posts: 598
  • Joined: 25 Jul 2010

Posted 15 October 2010 - 06:26 PM

Jim, I was said a higher resolution is unnecesary, because the film grain becomes visible. I'm happy you confirmed the scan is good.

Of course I wonder about your experimenting. If you want, I can send you a second frame for stacking. I played a lot with this frame. The image you can see in my Praktica gallery is the best of 14 different versions.

#24 Michal1

Michal1

    Viking 1

  • -----
  • topic starter
  • Posts: 598
  • Joined: 25 Jul 2010

Posted 24 October 2010 - 09:57 AM

What are the parameters you can set up on a scanner? Surely, there are some software settings, like the setting of black&white points. But in which way you can physically influence the process of scanning? I expect that the intensity of the light, the gain of the sensor and the scanning speed can be adjusted. Is there something else?

Do you have an experience with the Digital ICE technology for the film scratches removal? Does it have problems with stars in astrophotos?

M.

#25 Dave Kodama

Dave Kodama

    Mariner 2

  • -----
  • Posts: 226
  • Joined: 04 Oct 2010

Posted 24 October 2010 - 03:09 PM

The parameters depend greatly on the software supplied by the manufacturer. I think that's the primary reason why not all scanners are good for astrophotography (essentially always working with underexposed film). The hardware might be fine, but the manufacturer doesn't put enough effort into the software.

Most scanners I've used have some control of gamma, exposure, contrast, and color balance, but not enough to get the final scan close enough to avoid additional adjustments later, which is why it's really important to be able to get lots of bits/color or a true raw scan.

In my experience, there is no control of the light intensity. You can see why if the light source is a fluorescent tube. It might be possible with a scanner which uses LED's for a source, but I haven't seen it.

Dave


CNers have asked about a donation box for Cloudy Nights over the years, so here you go. Donation is not required by any means, so please enjoy your stay.


Recent Topics






Cloudy Nights LLC
Cloudy Nights Sponsor: Astronomics