Tuesday, December 31, 2013

Scanning Torture (or Learning to Love Your Digital Camera)

Over the last 3 months I have been scanning a bunch more of my old 35mm slides and negatives using my Minolta Dimage Scan Elite F-2900 film scanner and Vuescan scanning software. It produces a 10mp image. From 1998 until a few months ago I had little by little scanned about 3500 slides/negatives. During the last 3 months of concentrated effort I have scanned an additional 6400+. I expect to do a few more over the coming weeks so the total from when I first started in about 1998 will be about 10,000. If you don't have a lot of experience scanning film (nothing like scanning prints) then you probably don't really know how much work it is to do the scan and then do all the preparation of the resulting file. The second part that is the preparation of the scan file is mostly left to be done. Out of the close to 10,000 scans I have done over the years I have only prepared about 1600.

In my experience the order of satisfying scans are, best to worst:

1. B&W negatives
2. Color slides
3. Color negatives

The most satisfying scans are of B&W negatives. You don't get the benefit of the automatic dust/scratch fixer that uses the infrared channel with B&W film though so they are a lot of work to fix up, but you don't have to be concerned about color. Getting very good color from a scan, especially a color negative scan, is difficult. Also, with B&W negatives you usually don't get the muddy/ugly/blotchy shadow areas that you often get with color negatives. Kodachrome slide scans are worse than scans of E-6 slides (Fujichrome, Ektachrome) and you can't use the infrared channel to get automatic fixing of dust/scratches. Color negatives are the worst. The dark areas are often muddy/ugly/blotchy and the whole image is grainier, but not good looking sharp grain like with B&W negatives. It is a blotchy, ugly grain with color negatives. The film base color (depending on film it is various shades of orange or pink) makes getting good color a real chore. Even when selecting the proper film type in Vuescan the result is not usually as good as with color slide scans. And even color slide scans often don't have color that is as good as the actual slide. By the way, I have a few C-41 B&W negatives (Ilford XP2) and they also don't scan so great. No color issues, but the grain is like the color negatives and the shadows are muddy/blotchy.

Back during the late 1990s there were people on the internet claiming that color negative film scanned much better than color slide film. I don't recall all that was claimed, but I think it had to do with lower Dmax or something like that.  I started shooting more color print film then because I expected that I would want to scan it later. I sure regret that. I suppose that is what you get when you listen to the internet "experts" who think they know the theory without having real experience. :-) This is the film I have been scanning:

B&W negatives:
Kodak Plus-X ASA 125
Kodak Tri-X ASA 400
Fuji Neopan 400 ASA 400
Ilford XP2 ASA 400 (C-41 process)

Color slides:
Kodachrome II ASA 25
Kodachrome X ASA 64
Kodachrome 25 ASA 25
Kodachrome 64 ASA 64
Kodachrome 200 ASA 200
Ektachrome X ASA 64
High Speed Ektachrome ASA 160
Ektachrome Elite 100 ASA 100
Ektachrome Elite 400 ASA 400
Fujichrome Velvia ASA 50
Fujichrome 100 ASA 100
Fujichrome 400 ASA 400
Fujichrome Sensia 100 ASA 100
Fujichrome Sensia 400 ASA 400
GAF 500 ASA 500 (I used only one roll of this for some indoor photos during Christmas 1974)

Color negatives:
various Kodak, Fuji, and Agfa films ranging from ASA 80 to 400

Okay, after all of the above I will finally get to my main point. :-) The quality of the raw files from the 10mp tiny sensor in my Canon S95 are so much better in every way. Although some may say the dynamic range is less, maybe a lot less, than color negative film I have found that in practical terms it is much better. As long as you are scanning to convert the analog film to a digital file then the result, in my experience, is less usable dynamic range than what I can get by shooting in raw with a tiny sensor digital. I can shoot for the highlights and bring up the shadows with the digital and still get better results than scanning color negatives. The color is also so much better it isn't even worth comparing. The noise/grain is also so much better it isn't worth comparing. The S95 at ISO 1600 or 3200 is probably better than scanning ASA 100 color negative film. Actually, even jpegs are better too. Another actually: my old 5mp Minolta D7i is better. I would much rather have a 5mp jpeg from the D7i than a 10mp scan of a color negative (also a color slide). Oh, the 10mp file from the scanner probably has no more than 5mp of real, useful data in it for most color negative scans since they are pretty noisy/grainy. The Scan Elite is a pretty good film scanner (cost me $1085 several years ago) and Vuescan gets even more out of it. I always scan using multi-sampling to reduce noise a bit more. A different scanner in some cases might be marginally better, but not much. It is just the limitations of converting analog film to digital.

This post isn't meant as a complaint. It is just meant to remind us how much better digital is than 35mm film converted to digital. Not just a bit better, a whole lot better. Even a digicam is so much better. (Of course, a digicam's handling isn't anywhere near as good as a nice 35mm SLR and you have less control of dof, but those aren't the things I am talking about.)

I use Vuescan to create RGBI 64-bit raw scan files.  The result is a 10mp raw file (4000x2688 pixels) that is about 84mb uncompressed.  I did a bunch of scans before I realized that I had forgotten to select the compress raw option.  With raw file compression the file sizes are generally 50-60mb.  I have about 420gb of raw scan files right now.  Someday, little by little, I will run Vuescan again using the raw file as input and create an output tiff file.  Then use Photoshop to work on, clean up, etc. and finally import it into Lightroom. This is something that will take years, but can be done without access to the slides/negatives or scanner -- I can do it anywhere.  It is a lot of work and is not interesting to do at all.  I have done that for about 1600 of the files so far and have imported them into Lightroom.  I have close to 10,000 raw scan files and I doubt if I will ever prepare all of them, but whenever I have nothing to do I can do a few more.

The actual scanning time when using a dedicated film scanner is the trivial part.  The preparation for scanning (getting everything together, examining them with a 10x loupe on a lightbox, trying to determine whether I already have scanned this image years ago, selecting, taking negatives out of the plastic pages, cleaning, inserting in the holder, trying to get them properly lined up in the holder, scanning, and then taking out of the holder, putting the strip back in the page, etc.) is what takes so much time.  The easiest to scan though are color slides by a very wide margin.  Dealing with strips of negatives is a real pain in the butt.  Some are curled a bit so hard to get properly lined up in the holder and also even if you want to scan only one you have to load all of the frames in the strip. Selecting film to scan using the loupe and lightbox is much easier with slides than negatives too.

Happy New Year and Best Wishes for 2014!

Over 10,000 35mm slides and negatives scanned! -- Part 2




Sunday, July 21, 2013

Bali, Indonesia

I have been wanting to go to Bali, Indonesia for years and recently since I found myself with some free time here in Japan I decided to just get up and go.  It was a good trip.  Here is a short trip report:

http://www.bakubo.com/Bali.html

I have created a new Bali, Indonesia photo gallery on my website so please take a look if you are interested:

http://www.bakubo.com/Galleries%202/index.html

Friday, April 19, 2013

Canon G15 vs. Canon G1X

I just checked on DxO comparing the G15 and G1X (large sensor):

G1X:  about 2 stops better than G15 for noise and at higher ISOs about 1 2/3 stops better for dynamic range, but at lower ISOs the dynamic range is the same and at the lowest ISO the G15 is better

The G15 has a fast 28-140mm f1.8-2.8 and the G1X has a slow 28-112mm f2.8-5.8, so at the wide end the G15 is 1 1/3 stops faster and at the long end it is 2 1/6 stops faster.  In other words, especially as you start using the longer end of the zoom range the G15 will be using ISOs that are 2 stops lower than the G1X.  So, compare G15 ISO 200 to G1X ISO 800, G15 ISO 400 to G1X ISO 1600, etc.

The G1X is much bigger and heavier so for my uses of the camera it was not really in the running.  It was interesting to take a look at the G15 vs. G1X results on DxO though.  That G15 fast lens really makes a huge difference.

Thursday, April 18, 2013

Canon G15: Some thoughts on my new camera

For the last 1.5 years I have been using a Canon S95 with it's 28-105mm f2-4.9 lens and it has been pretty good as a carry everywhere camera, but I really missed not having a VF for use sometimes.  Before the S95 I used a Canon A590IS, A540, and A70 as my carry everywhere cameras.  Back when I bought the S95 I looked for digicams that had a VF, but there were very few.  The G12 had one, but it was too big/heavy and the lens was slow.  The S95 is small and has a faster lens, but no VF.
Last week in Osaka I had the chance to look at a Canon G15 several times and then on Friday I ordered one on Amazon Japan.  It arrived on Saturday, less than 23 hours after I ordered it -- and this was with the free shipping option.  Here are a few of the improvements of the G15 compared to the previous model, the G12, that caught my eye:

1.  Smaller size and lower weight.
2.  New 12.1mp sensor with better noise and better dynamic range than the G12's 10mp sensor.
3.  A 28-140mm f1.8-2.8 lens replaces the G12's 28-140mm f2.8-4.5 lens.
4.  A 3" 922,000 pixel LCD replaces the G12's 2.8" 461,000 pixel LCD.
5.  The OVF is a bit better.  I think Canon claims 85% coverage now and the G12 was a bit less.
6.  Faster AF and faster operation.

The G15 noise and dynamic range is better than the G12 so that is a nice improvement and that along with the much faster lens helps the IQ a lot since lower ISOs are used.

The G12 had an articulating LCD, but the G15 does not.  I am rather happy about this change because I rarely use the tilt/swivel LCD of my other cameras and for a carry everywhere camera I value the size/weight savings much more than the tilt/swivel convenience for the rare times I would use it on this sort of camera.

It is the better sensor mated with a much faster lens that really got my attention since I wondered how it would compare to a larger sensor camera mated with similar effective focal length zoom lenses.  Before buying I investigated on DxO by comparing the G15 sensor to the Canon G12, Sony RX100, Sony A700, Sony A100, Canon 50D, Canon 40D, Nikon D300/D300s, and Canon 5D.  I chose these DSLRs to compare to because they all have bigger sensors and similar megapixels.  Also, just a few years ago I would have been thrilled for a digicam that could get close to those cameras.  I have owned the Sony A700, Sony A100, Canon 60D, Canon 30D, Canon 300D, and KM 7D.  Currently I am using an Olympus E-M5 and Panasonic G3.

I discovered that the noise and dynamic range of the G15 is better than the G12 (as claimed).  Here are the other results:

Sony RX100:  about 1 1/3 stops better than G15 for noise and dynamic range
Sony A100:  about 1 1/2 stops better than G15 for noise and worse for dynamic range
Canon 50D/40D:  about 2 stops better than G15 for noise and dynamic range
Sony A700:  about 2 stops better than G15 for noise and 1 stop for dynamic range
Nikon D300/D300s:  about 2 stops better than G15 for noise and 1 stop for dynamic range
Canon 5D:  about 3 stops better than G15 for noise and dynamic range

Okay, with all of that info then consider the following cameras/lenses:

Sony RX100 + 28-100mm f1.8-4.9
Canon 50D/40D + Canon 17-85mm f3.5-5.6 IS
Sony A700/A100 + Sony 18-70mm f3.5-5.6 or Sigma 18-125mm f3.5-5.6
Nikon D300/D300s + Nikon 18-105mm f3.5-5.6 VR
Canon 5D + Canon 28-135mm f3.5-5.6 IS

The G15's 28-140mm f1.8-2.8 lens at 100mm is probably about f2.5 or f2.8, so about 1 2/3 to 2 stops faster than the RX100 and the G15 noise and dynamic range is only about 1 1/3 stops worse.

The G15's 28-140mm f1.8-2.8 lens is 2 stops faster than the Canon 17-85mm, Sony 18-70mm/Sigma 18-125mm, and Nikon 18-105mm lenses and the G15 noise and dynamic range is also about 2 stops worse than the A700/50D/40D.

The G15's 28-140mm f1.8-2.8 lens is 2 stops faster than the Canon 28-135mm lens and the G15 noise and dynamic range is about 3 stops worse than the 5D.

The point of all of this is that it is pretty impressive that the combination of the new 12.1mp sensor plus the new, fast lens tests similar to the A700/50D/40D with similar focal length lenses.  Of course, I am not saying that the G15 will get exactly the same results as a A700/50D/40D.  Also, the G15 is a very different type of camera.  I just thought these comparisons were interesting.  I also looked at the dpreview raw comparison with various cameras and it seemed to show pretty much what DxO shows.
As an example, the G15 with its fast lens can use ISO 400 when the Canon 40D/50D, Nikon D300/D300s, and Sony A700 must use ISO 1600 and so on.

For a carry everywhere camera when I am out and about going through normal life I want something that is all self-contained (lens, flash, lens cover) and that is small and flat enough that I can put it in a jacket pocket or cargo shorts pocket.  The G15 is small enough, but it is rather heavy so I will have to see whether that heavy lump in a pocket works out okay.  At the moment I have the neck strap on it and have been carrying it around that way hanging from my neck.  I don't like that for normal life though.

By the way, a word about the OVF.  This isn't a TTL viewfinder so it isn't possible to have a 100% or close to 100% view.  For that you will need a high end DSLR or you can use the G15's LCD which does have a 100% view.  The G15's OVF is sufficient for my purposes since it always shows a bit more than you will get rather than less than you will get.  If they tried for something close to 100% then the parallax alone would cause the view to be a bit wrong much of the time and the cost of such an OVF would probably be very high.

Saturday, January 19, 2013

Nepal photos online

I have created a new Nepal photo gallery on my website so please take a look if you are interested:

http://www.bakubo.com/Galleries%202/index.html

I traveled in Nepal for a month in November/December, but it has taken me a bit of time to get some of my photos ready because I bought a new computer and then there was Christmas and New Years. It was my first time to go to Nepal so it was cool to see many new places, meet some new people, and take photographs in a new place!

In Nepal I went to Kathmandu, Patan, Bhaktapur, Nuwakot, Bandipur, Pokhara, Sauraha, and Chitwan National Park. I spent 2 weeks traveling with 5 people from Australia and Scotland and 2.5 weeks on my own. They were a great group of people!

Among other things, we went on a 2-day, 1-night whitewater rafting trip on the Seti River in Nepal and at night we set up tents next to the river and spent the night. In the deep, black sky the stars were glorious! I had planned to go on a trek somewhere for about a week, but then I got quite sick for 10 days, lost lots of weight, and even after I finally started feeling better I was weak so I ended up not doing any trekking. I did go on a couple of 7 hour hikes (one in the hills outside Bandipur to a little village called Ramkot and one at Chitwan National Park through the jungle). Even though I was sick for 10 days (even during the rafting trip) I still did a lot and didn't sit around in a hotel room somewhere. I didn't want to miss anything. I just didn't enjoy it as much as I would have if I had been well. For example, in Pokhara before sunrise we climbed to the top of Sarangkot to watch the sunrise and the first rays hitting the Annapurna range of the Himalayas. Beautiful!

Here is a panorama of the Annapurna Range of the Himalaya Mountains at sunrise in Nepal taken from the top of Sarangkot:

http://www.bakubo.com/panoramas.html

I ended up spending more time in Kathmandu than I had expected because I was recovering from illness. Also, the airline canceled my flight home so I had to stay an additional 3 days until another flight from Kathmandu to Kunming to Shanghai to Honolulu was available. The extra time in Kathmandu gave me the chance to do more street life photography there though!

This is what I took with me:

Olympus E-M5 + 3 batteries + charger
Panasonic G3 + battery + charger (backup body)
Olympus 14-150mm f4-5.6 + UV filter + polarizer filter + lens hood
Olympus 9-18mm f4-5.6 + UV filter + lens hood
Panasonic 14mm f2.5 + UV filter + lens hood
Panasonic 20mm f1.7 + UV filter
Olympus FL-300R flash + 2 nimh AAA batteries
Canon S95 digicam + battery + charger


Friday, January 4, 2013

Olympus OM-D E-M5 very slow wakeup

I use an Olympus OM-D E-M5 for street stuff and it is pretty good, but there is one thing that really, really annoys me about it.

The E-M5 is slow, much slower than a DSLR, to wake-up from sleep or when you turn it on. In my street photography I sometimes miss shots because the camera is waking up very leisurely. When possible I do my best to anticipate when I might want to take a shot and start the wake-up process, but sometimes things happen very quickly and by the time the camera is finally ready the moment has passed. In most ways the E-M5 is fast and very responsive, like a DSLR, but in this area it is like a digicam. Even if a DSLR was as slow as the E-M5 it would still have an advantage because while you are waiting for the camera to be ready to shoot you could still look through the OVF, do quick framing, adjust zoom, and then shoot as soon as the camera is ready. With the E-M5 you can't even do the quick framing and adjust zoom while you are waiting and waiting and waiting for the camera to wake-up.

Since the E-M5 does not have an OVF it uses battery power pretty fast.  If you set it to never sleep then the camera is ready all the time, but the battery will run down fast.  Also, the sensor and EVF will be on all the time even during the long periods when you are walking around watching for a potential shot.  The sensor will be heating up and getting noisier.

I just got back from a month in Nepal and there were a few times when I almost threw my E-M5 against a brick wall when I missed a sudden photo opportunity while the camera took its sweet time waking up.