Scanning photos

[stiltman]

I hear ya VegasBud,

I would never consider jpeg as best format of chioce for storing or editing.
I either use Tiff (uncompressed) or PSD. I then save a copy to jpeg to send out or upload.

Using jpeg as your single editing and storage format is ludicrous


The first image shows the original JPEG.
The second image is the same file after being saved 100 times.
The file is noticeably degraded.

Discussion over

[SingSing]

Hey, Most digital cameras give you only jpeg images.

Note: We all know wave file sounds better than MP3. But how many wave player is out there ? NONE.

[lordsmurf]

Sorry VegasBud, but I think you're wrong.

MPEG-2 at a high enough settings is fully suitable for archival quality, under certain contexts. BJ_M used to point out that many editing situations would use "master" MPEG-2 files, including Sony professional workflows. MPEG in and of itself is not too compressed to be viable. That can be said for DVD-Video's MPEG-2 specification, but not of MPEG itself. The profiles, GOP and bitrate settings can give a near-transparent quality.

If your JPEG images have artifacts, then you're doing something wrong. A maximum level-12 JPEG "compression" in Adobe Photoshop CS3 is hardly all that compressed (file sizes are still very large, close to an LZW TIFF), and the introduction of artifacts is pretty much zero.

stiltman's "proof" is anything but proof. Anybody that saves over the same image over itself again and again is a moron. If you want to create a "working document" version, then use a RAW format, specifically the DNG (digital negative format). Using TIFF is the method of the past decade, not the current one.

Maybe 2-3 years ago, I'd have agreed with you guys. But not in 2008. Things have changed. Neither you nor I can argue against it, we must simply deal with it. The MP3 comment is not exactly the same, but it is comparable (although MP3 really is a lossy format).

Bigger isn't better. It's just bigger. TIFF is a perfect example of this.

[thecoalman]

I just did test with this image, this is my first save as .jpg at 100%. Note it's 153 KB:






Usually when recompressing a jpeg you'll lose a lot of file size particularly from images on camera. However for each incremental save the image went up 5KB. Here's the tenth save:





Note the image was closed each time I saved it saved then reopened so there's no possibility it was simply re compressed from the original. Truthfully I can't find any artifacts in the original or the tenth one but I didn't go over it with fine tooth comb either. What I did check was the red objects and edges as that is what usually gets it in the ass. Software used is Photoimpact 8 which is about 6 years old. :P The text and shadow had no anti alias or soft edge.

Origianl tiff, uncompressed:



01_jpeg_test.tif


I did a similar test a few years back using 90% and in that case the first numbers did show some artifacts after a few saves. I think one of the major problems in suggesting using JPEG is if the person is using the software improperly or it's just junk software.

[lordsmurf]

TIFF will look like crap in junk software too. Been there, seen it, years ago. Just an FYI.

[VegasBud]

lordsmurf,

There is a limited subset of available photo processing software which will allow you to insert a jpeg into a linear dng, which is used as a container. When the resulting dng is (re)saved, any editing changes made during the session can be written as xmp metadata directly into the dng, without altering the original image...if the software supports it. Lightroom is one program that can do it that way, or alternately just maintain the session state when Lightroom is closed. Either way, as long as the edited image isn't exported (rendered) to jpeg, you could perform unlimited editing, in as many sessions as you want, without any loss of quality beyond what existed in the original jpeg. If Lightroom can do it, there are very likely other programs that can do it also. Maybe the software you're using has similiar capabilities.

However, even in that environment, if a new image is generated by exporting (rendering) the edited version, and jpeg is chosen as the file format, there is a generational loss. It's only one generation from the original, regardless of how many times it was edited, so if the quality setting is at maximum, you may not see the degradation, but it is impossible for a lossy compression format (like jpeg) to produce lossless output.

Jpeg may be very adept at exploiting weaknesses in human visual perception to conceal the fact that it's discarding large amounts of real data, but that's exactly how it compresses. Artifacting may take several generations before it becomes apparent, but it exists even in the first generation...even using maximum quality settings. I think you're aware of that, which is why you said "the introduction of artifacts is pretty much zero".

Each individual has their own goals, and way of getting to those goals. If your workflow works for you, Ken Rockwell's workflow works for him, and my workflow works for me, we're all doing what's right for us individually...even though our workflows may be completely different. Nevertheless, my opinion hasn't changed as to what constitutes "best practice" for 99.9999999999999% of the people scanning photos. I still believe the advice given by myself and others is the best way to achieve the highest quality results, and that jpeg is best utilized as a first generation output format which is only used when needed.

One quick thing:
I believe stiltman was merely demonstrating, in the clearest way possible, generational loss when using jpeg...nothing more

[videobread]

Re: Debris free 8 * 10's from the Pro's:
I brush my slides and then photoshop out dust, etc. when needed. Results are great on my Epson 4490 Photo scanner. Scanning at high resolution is not the issue for me. 600 is fine. Results at 1200 are a little better but not worth the increase in scan time.
I assume you are using photshop to get rid of junk.
I have thousand of slides to scan, but will never get to it. At least slides hold up at lot better than photos over time. Good luck.

[cyflyer]

Ok for what its worth, I'll add my two cent's worth. On the subject of dusty scans, I too have a Epson 4490 scanner which scans at pretty decent resolutions and picks up any stray dust on the scan. Solution ? I constucted a box out of that foam mounting material, dressed over with the self adhesive black shiny stuff, and keep the scanner safely tucked underneath. Works a treat, and keeps it clean even in the dustiest of conditions. Its simply and very effective. Try it.
As for the Tiff versus Jpeg argument, I too have to agree with everything VegasBud has said. Nuff said.

[edDV]

FWIW I took on the job of scanning family photo albums so each gets a copy.

First I found that removing photos from albums was potentially destructive and a huge labor sink. Second I discovered the dust issue.

My solution was to use a triage plan as follows:

1. First I scan the albums + loose photos a page at a time at 300-600dpi to a tiff for master and jpg for viewing. For this you need a large bed scanner. This gives the viewer the experience of viewing the album as is including wriiten comments. Individual photos can be zoomed at reasonable quality even with a 300 dpi scan.

2. Most are happy with #1. Some request detailed scans of specific photos. I then disassemble the album and clean the photo. This is where I face the dust issue. I found the better solution is to abandon the "office" and convert a bathroom to clean room scan duty. I clean the bathroom, photos, scanner and a laptop thoroughly and then maintain some humidity in there to control electrostatic cling. The exhaust fan helps suck out remaining dust but be sure to filter the under door incoming air. Better to seal the door and filter HVAC vent incoming air.

3. Those that still complain get the option of paying for a pro scan at a local photolab.

35mm slides are normally sent out for pro scan. Alternative is to rent a Nikon slide scanner and use the bathroom clean room technique but the dust issue is huge with 35mm slides. I do this for my own slides but I don't want to hear the complaints from the others.

[Des]

One last thing...since you used photographer Ken Rockwell as an "expert witness" for why jpeg is superior, I should point out that Ken Rockwell's choice of jpeg as a preferred file format is a direct result of his philosophy of photography. Throughout his website, he makes it very clear that he considers post processing of photographs to be a crutch for incompetent photographers. In his technique, all processing is performed in-camera by the camera settings he uses. When the photos come out of the camera, they're reviewed, selected, then shipped off to the clients. If a photo requires post processing, he discards it. If his workflow included post processing for increased flexibility and control, he wouldn't be advocating jpeg for a do-it-all format.

Funny this is the same workflow for shooting slides, which was the standard 'pro' format for decades. I run a photolab/store for 20+ years and have more than done (and do) my fair share of photography so I have some familiarity with all of this.

A good photographer does create the image in camera and does not have to post process anything. Exposure, light, subject and composition are what makes a good photograph. Digital photography has somehow managed to convince a large number of people that they need to post process their images.

Of course the other school of thought is a 'zone' approach in which you shoot with post processing in mind, in which case the two have to go togehter. Ansel Adams made this approach famous. Of course I am highly over simplifying it.

The other would be complex studio set ups since ith digital multiple exposures can't easily be done in camera.

As for JPEG vs TIFF vs RAW vs A(miga)IFF

If you don't have to edit the image, aside from some cropping, then high quality jpeg is considered quite suitable in the photographic world.

--dES

[lordsmurf]

It's the same for scanning -- scan it correctly to begin with.

[VegasBud]

Des,

Digital photography has somehow managed to convince a large number of people that they need to post process their images.

It might be that post processing of images on a computer, with readily available software, has simply become more accessible to the masses than when it required an arcane knowledge of the darkroom. I don't know about "needing" to do it, but there sure are a lot of people who seem to really enjoy it.

As for JPEG vs TIFF vs RAW vs A(miga)IFF laugh.gif

If you're finding this discussion of the concept funny, maybe you'll be more receptive to Roger Clark's take on the subject:
Digital Camera Raw versus Jpeg Conversion Losses
Dynamic Range of an Image - (the "Raw versus Jpeg" section begins about 2/3 of the way down the page)

If you don't have to edit the image, aside from some cropping, then high quality jpeg is considered quite suitable in the photographic world.

If you're willing to take out the "aside from cropping" part, I agree with you and the photographic world.

[VegasBud]

While I was digging the Roger Clark links (posted above) out of my very unwieldy e-library , I ran across this link, where he demonstrates scanning a SHARP print at 400 dpi does produce increased detail. His assessment is that there is an improvement in detail up to 600 dpi, but I don't honestly see much difference between his 400 dpi and his 600 dpi samples. He doesn't give the test specifics, so anyone who's really interested might want to run their own tests using their scanner and their prints.

[thecoalman]

His assessment is that there is an improvement in detail up to 600 dpi, but I don't honestly see much difference between his 400 dpi and his 600 dpi samples.

And again depends on the source and where you're going with it. Let's say for example our 6X4 image has a density of 600dpi. At viewing distance the human eye is only going to be able to discern 300dpi of detail. We're not going to lose anything if we scan at 300dpi and print the same physical size image. Where we're losing is if we print the image at a larger physical size.

So the question is does the print contain enough information to justify scanning it that high?

[lordsmurf]

A print is a relatively lousy source, so it's highly unlikely to benefit from a higher-res scan. If anything, it will simply show flaws of the print more than details in the image. This is the affliction that can turn one of my older $1,000 lenses into a lens that shoots mediocre on an APS-C crop body sensor -- all it does is enhance flaws in the pre-digital glass, even worse than some newers and cheaper digital-coated lenses.

The scanner's glass itself is just as important to this discussion. Too often people are bamboozled by compression or counting pixel (notably DPI and megapixels) instead of looking at the other issues that are far more helpful (or damaging) to the output quality.

A number of these scanners look pretty much the same at 300 as they would at 600 or 1200 because the glass doesn't really allow the detail to come out anyway.

If you use the wrong color profile in the image, or scan it with "cooked" default TWAIN plugin settings, then it won't matter if you use JPEG or TIFF or anything else. The damage was already done long before you hit "save as".

I scan at either 300 or 600 on my flatbed, because it's optically very good. The 600 is usually done because I am indeed printing larger than the source, and I need all I can get. I still have to go and seriously filter out the noise inherent to the smaller print, to make it passable at the larger size. The 300 is done for same-size prints.

Most photos can be printed at 200-240dpi, and it's pretty much transparent to 300dpi. Text is the only real problem, and that's better at 600dpi anyway. Of course, if your printer is lousy, that can hurt you too. I'm using a high-end color laser.

[Des]

Des,

Digital photography has somehow managed to convince a large number of people that they need to post process their images.

It might be that post processing of images on a computer, with readily available software, has simply become more accessible to the masses than when it required an arcane knowledge of the darkroom. I don't know about "needing" to do it, but there sure are a lot of people who seem to really enjoy it.

I'm not arguing that there is no room for the hobby of it, I'm stating that it is not a necessity for process to work.

As for JPEG vs TIFF vs RAW vs A(miga)IFF laugh.gif

If you're finding this discussion of the concept funny, maybe you'll be more receptive to Roger Clark's take on the subject:
Digital Camera Raw versus Jpeg Conversion Losses
Dynamic Range of an Image - (the "Raw versus Jpeg" section begins about 2/3 of the way down the page) [/quote]

Well, in the scheme of life I put tragedies on top and everything else well bellow that so I see room for levity here. If you can't that's fine. To each his own.

If you don't have to edit the image, aside from some cropping, then high quality jpeg is considered quite suitable in the photographic world.

If you're willing to take out the "aside from cropping" part, I agree with you and the photographic world.

Cropping also occurs in the mechnical aspects of making a final print. My statement stands.

--dES

[lordsmurf]

maybe you'll be more receptive to Roger Clark's take on the subject:
Digital Camera Raw versus Jpeg Conversion Losses
Dynamic Range of an Image - (the "Raw versus Jpeg" section begins about 2/3 of the way down the page).

No, sorry. It's one of those web pages where I'm reading and in my head I'm thinking "blah blah blah, get to the ******* point" --- but it never happens. Some information is just upchucked onto the page, and that is supposed to suffice? I don't think so. Photographing test patterns and analyzing them in graphs doesn't mean anything to the conversation at hand.

[VegasBud]

lordsmurf,

It's one of those web pages where I'm reading and in my head I'm thinking "blah blah blah, get to the...point" --- but it never happens.

Yes, that's true. He's an imaging scientist, which influences the way he looks at things. He only provides his testing methodology, representations of the data collected, and limited analysis. He rarely provides a conclusion because that would introduce bias into his scientific studies. His style is definitely an acquired taste.

The idea behind including the links was to (theoretically) illuminate the subject better by providing more specific scientific details than had been included in the discussion, but that appears to have not been entirely successful.

I tried putting together a "get to the point" version, translated from geek to english, but in order to avoid making blanket statements, I ended up with a lengthy "part geek - part english" version which had several points, not one.

[VegasBud]

Probably it's best to just keep it simple, and have some fun while learning.

First, the "keep it simple" part:
If you want to preserve all the data in an original image, use a lossless file format, which can be uncompressed or compressed. If you want a smaller filesize than what compressed lossless file formats can provide, you have to give up some image data. The smaller the filesize, the more data you have to give up. Jpeg is popular because it offsets the data loss by removing more data from the part of the image which our eyes don't detect as well, and removing less data from the part our eyes are more sensitive to. Even though there is always some data loss with jpeg, the perceived image quality can be excellent the first time you save the image to jpeg (depending on the quality setting chosen). Each subsequent jpeg re-save increases the data loss and decreases the image quality.

The "have some fun while learning" part involves a trip to this page by Calvin Hass, where he will give you a free and useful program called "Jpeg-Snoop", which can tell you if a jpeg is an original or edited, repair damaged jpegs, etc., and also display all the inner details contained in a jpeg file that you're likely to want to know. That same page also acts as a springboard into learning more about jpeg, how it works, and how to best use it. Click any of the many links to learn more about a part that you're interested in. Even if you don't want to learn anything about jpeg, the "Jpeg-Snoop" program can still be fun to play with.

[vhelp]

Sorry to waik an old topic up, but..

I was looking for a topic that was asking a question about painshop pro x2 ultimate (I think it was about that suite) or some other photo editor suite (that I have been researching and debating purchasing) but I could not find it, so far. I want to get into the digital photo area because since I began with it several years ago with my canon c330 digital camera but put it on hold due to that camera's lack of true features and functions I was originally looking for, I had to put it on hold. Now, I know better (so I think) and want to get a digital SLR camera--but they are expansive, oh my! So I'm looking around, shopping and researching my options. Anyway. I came across this discussion, and thought I would drop by and throw up my 2cents worth.

EDIT:

ok. After finding the image and commenting further *and* making several attempts to upload .jpg image, I keep getting session timeout. Therefore, I am bailing out on this due to dialup (and human) fatigue plus I have other things planned and now notime for some of them. Sorry for waisting everyone's time and opening an old topic.

-vhelp 5204