daniel_taylor

"The one thing I've read in some piece of Canon documentation is that you must not remove an IS lens when the IS is active (actually on, not just turned on). I have no idea why. That said, you'd have to be pretty adept at removing lenses to manage to do that."

The IS element most likely "parks" itself just before turning off so that it can't move around freely. Removing the lens while it's in motion cuts power before it would have a chance to park. Then you have an element that's free to shake around as you move the lens.

Other than that, I see no reason to turn the power off to change a lens. The sensor is behind a closed metal curtain, it's not going to attract any dust.

"I've done the test with a 50mm f1.8 on a Rebel Ti and a EOS D5 and it is true.

If you want a full frame sensor you better have the absolute best glass available and be prepared to correct the vignette in p.s.

The full frame sensor will be softer in the corners than the film."

Please post your photos and test parameters.

"Vignetting occurs in digital sensors because they respond best to light that enters the cell at right angles. This is less of a problem with film because the surface is semi-matte and light entering at an angle is scattered into the surface. Halide particles don't care which direction the light comes from."

I keep hearing this explanation, but I have yet to see a side-by-side test where digital did in fact have more vignetting. And the few mentions from people on the web who should know (i.e. Reichmann) stated the opposite, i.e. about the same vignetting as film.

"Digital sensors do not cause chromatic aberation (sic), they are simply sharp enough to highlight CA in the lens that might be forgiven in a film camera."

Yes and no. Blooming on the sensor can contribute to and/or appear as "CA" even if it's technically something different.

Nikon will produce a FF body when a chip supplier provides them with a FF chip at a suitable price/performance point. Remember, Canon makes their own chips. That advantage is why they've had FF bodies. That Canon users can mount Nikon lenses on their FF bodies with an adapter should put to rest any theories about the F-mount or F-mount lenses being incapable of handling FF (after all, the adapter puts the lens at the exact same distance it would be at on a "real" F-mount).

"Interestingly enough, when the owner of the 1Ds MKII saw what it was compared to he too couldn't believe it either and insist on reshooting to maximize the results."

LMAO! BS Les. Are you telling me that my 10D records more data than the 1Ds mkII does within the APS-sized center of its sensor? Because composed as if it were full frame, the center of my 10D SPANKED your best film scan. I had NO need to reshoot that test. I could have hand-held it and still come out ahead.

At this point you're either lying, your friend used a 28-300 cheapo zoom, or you guys have no clue how to setup and properly shoot a 1Ds mkII.

"Yeah, he owns every Canon DSLR from the D30 (which he thought had long since spanked film!) and he isn't falling back laughing ;-)"

If he can't out shoot 35mm Velvia 50 on a desktop scanner with a 16 MP 1Ds mkII than I'm falling back laughing at his technique. Tell him he should immediately give his 1Ds mkII to someone who knows how to use it.

Come on Les, post the images! No cheating though! It's going to be painfully obvious you cheated if my 10D, composed as if full frame, out performs a 1Ds mkII.

"Of course to get back on topic, it is the year 2006 and film is still the R E S O L U T I O N K I N G ;-)"

Another example of the RESOLUTION KING is below. Enjoy!<div></div>

Wow...836 words from Les and he still didn't explain why his 4000 ppi 35mm scan failed to "annihilate" my lowly 6 MP "digiscam" picture.

Les, why don't you quit theorizing about 35mm Velvia vs. a 1Ds mkII and post your images? You have them, right? Post them! Unless, of course, you're so emotionally invested in your arguments that you just cannot post evidence which disproves them.

"But the bottom line is that, as far as pure line pair per millimeter resolution is concerned, 4000 ppi scanned Fuji Velvia 50 should be an average resolution match for a full frame Canon type sensor carrying 20.96 megapixels."

Really? Gosh, that's funny. See, I framed your map with my 35mm body, then put my 10D on the tripod with the exact same lens and mm setting so that the APS sensor was recording the exact same image that a 35mm frame would within APS dimensions, a simulation of what the center would look like in a 9.6 MP capture using 10D sensor technology. (This is all in the other thread.)

Here's a repost of 35mm Velvia "annihilating" a 9.6 MP digital image. Kind of like your flower shot "annihilating" mine. (As a test engineer, you should appreciate direct, empirical evidence of just how many MP it takes to put Velvia away. About 9.6 MP is sufficient.)

"So what do you think? Seems my observations, and I'm sure others as well, were not far off afterall don't you think? Or rather, don't you see . . ."

Scroll up and take a second look at those two flower crops. Don't YOU see?

"So sorry Danny, but I didn't mean to leave it out . . . it's called COPYRIGHT and I thought an experienced photog as yourself would know what that means. But hey, we're amongst friends and whats a few pics between good buddies huh, its all in fun right? ;-)"

Ever hear of fair use? Or do you just want to intimidate me into no longer posting comparisons which blow your asinine theories out of the water?

***

Again, my apologies to film fans in this thread. This is directed at Les and is not intended as an "attack" on 35mm as a medium, just an attack on Les ridiculous claims.<div></div>

If you are able to make this work you are going to have quite a business opportunity. There are actually a number of "niches" in digital cameras left unfilled. If you can hobble together some kind of consistent sensor/control board combination and fit it to various bodies, you'll be able to charge a mint. Off the top of my head...

* Color and B&W digital Leica (you're already working on one).

* Color and B&W Canon FD.

* B&W Canon EOS, Nikon, Pentax, Minolta, etc.

* Large sensor, low noise P&S (might be harder, less room).

"Just imagine what you would do if you had important data stored in boxes of 3.5" or even 5.25" disk today?"

Umm...I would put the disks in a 3.5" or 5.25" drive? (I have the former, and the latter can be bought on ebay all day long.)

While not in use much any more, this media is still usable/recoverable. There will be CD readers for a LONG time. Any format tied to a consumer media format as successful as the audio CD and video DVD has a long lifespan.

"How to archive digital images so future generations can view them...? Print them, using an archival inkjet printer on acid-free paper!"

This is the best option if you want future family members to SEE them.

"This isn't just because of doubts about the media, per se, but because of technical standards. What are the odds that a computer of 50 years hence, or even 25 years hence, or even 15 years hence, will be able to read a CD-R? For all we know, 25 years from now data will be stored on little holographic cubes that each contain petabytes or even exabytes of information in file formats we can't even imagine. Or everything will be stored "on the web" and computers (or whatever passes for computers by then) won't have any concept of local storage or device readers. This means that not only do YOU have to be a geek, but your heirs or descendents will have to be geeks, too or the chain will be irreparably broken."

Putting aside arguments of CD-R/DVD-R stability, what's claimed here is partly true. Digital storage in the future is going to be incredibly dense. So dense that you will be able to fit your life's interests...movies, home video, music, images, books, everything you've ever created...on an iPod.

But that's not a bad thing as portrayed above. When this transition occurs there will be plenty of time to get your CD/DVD media onto the new devices, much like audio CD is slowly moving to computers and iPods. After that your images will simply exist, stored on multiple high density devices and probably backed up over the Internet. As Scott points out, it will be harder to "kill" a digital file then than it is to backup one today.

So for now...

* Make your archival prints.

* Keep 1 copy on a HD and 2 copies on 2 separate CD-R's or DVD-R's. Keep the discs geographically separate.

* If you really want to be paranoid, add another HD and another CD/DVD to the mix.

And quit worrying about it.

I have 8 year old CD-R's that still read perfectly. In fact, I've only had a couple old CD-R's give me problems of any kind, and when I tried a second reader they copied fine (though I still burned new, just to be safe). So there goes that "professional" opinion.

There is certainly junk media out there, but that doesn't mean all CD media is going to expire two years post burn.

"I support a lot of Windows machines (mail servers, Citrix, few hundred clients), and our marketing dept all use Windows XP - no crashes - no bugs - no malware or virus's, etc."

In a *corporate* environment, Windows does fine. But that's because somebody is always watching, making sure every machine has the latest updates, the latest antivirus, all the right settings, firewalls, etc, etc, etc. If something gets through it's spotted quickly and removed. *I* don't have a problem with either of my two XP PC's for the same reason.

The average home user does not have this advantage, which is why the average home user has more spyware and viruses than productive programs. I've repaired the home PC of pretty much every friend and relative I know, some of them multiple times. They've all had some form of antivirus and antispyware, and it didn't do crap. Generally speaking the companies just can't keep up with all the problems, and some of the problems are so deep in the OS they need Microsoft to fix them.

I'll be the first to say that XP is easy to use, stable, a good OS. And the first to say that it is superior in some ways to OS X (just compare the Win32 and OS X threading API's). But the security is a glarring hole, and Microsoft has displayed nothing but incompetence in fixing it.

"Easy enough to not run Internet Explorer without the browser having admin rights either , which pretty much renders it immune to exploits."

Oh please...I can think of a half dozen ways to exploit a Windows PC regardless of how Explorer is setup. You can use Firefox if you want. If I want on, I'm still getting on. The hole I mentioned above is just the latest in a long string of glarring errors in Windows and the Win32 API, speaking from a security standpoint.

"My question is what inflated benchmarks Apple is going to post on their web-site when their Intel boxes hit the street. Of course Apple users will claim their Intel based Mac's are faster than PC's {ROFL}. That is of course if you're rendering crystal balls or doing weather simulations with Altivec because I've yet to see a benchmark of the Server version of OSX not getting it's nads handed to it by Windows 2003 and Linux running on hardware costing 1/2 the price."

Generally speaking, the code I have that runs faster on Windows does so because the equivalent Mac API's are "bloated" (i.e. GDI vs. Quartz). In a head-to-head CPU battle on code which pretty much just runs on the chip without depending heavily on an OS API or the HD, the Pentium 4 gets smacked around by the G5. I've got some C data processing code, cross platform, no Altivec, where the speed advantage is 4-5x Mac. Extreme example, but valid if you want to discuss the CPU's.

PC server hardware is cheaper per unit of performance because it's just, well, cheaper. But when you add in license fees to Microsoft I've more than once told a client their bill would be less with Xserve, which just comes with more out of the box. It all depends on what you want to do.

"MacTel's" won't be faster than PC's with similar configurations because OS X has more complicated libraries and API layers. That may change when Vista hits because WinFX is very much .NET technology. I like developing with .NET, but it's two or three layers of more bloat than, say, a classic C program accessing a simple C-style API would have. Vista will use CPU-sucking graphics libraries to render a "candy coated" user interface, just like OS X does. Both OS X and Vista will waste equally large numbers of CPU cycles drawing pretty graphics and managing abstract runtimes. And neither user group will care because they'll be running on dual and quad core P4's that have enough power to do all of that and still simulate a nuclear bomb blast.

Macs are not actually all that much easier to use than a modern PC with XP.

BUT...they are immune to the viruses and spyware that plague PC users. This is a much larger problem than many will admit. Show me a PC user who thinks this is minor, and I'll show you a PC user with about a dozen pieces of both on his system, which now happens to be running slower than an original 233 MHz iMac because of it.

For this reason alone, unless you're pretty knowledgable with Windows (i.e. can manually track and remove spyware since NO product currently captures all of it), I recommend home users go Mac OS. Security is just non-existent in Windows. (And before I get jumped on as a Mac fanatic, I spend 90% of my computer time on PC's.)

To give but one example: one of the more recent security alerts was that a hacker could gain access to your machine simply by getting you to view a WMF image, even one marked as a JPEG and viewed in a 3rd party browser. That Microsoft would make it possible to hook code into a DATA file and execute that code within an OS API just shows you how little thought went into Windows design from a security standpoint. There have been literally hundreds, if not thousands, of "holes" like this over the years. And they race to plug another one....

The only catch to this advice is: if you're going to buy a Mac, you might want to wait and see if Apple introduces anything at MacWorld Expo tomorrow. If they're finally releasing their Intel based Macs, you may want to get an Intel based iMac. Nothing is wrong with the G5, it's actually much faster per clock, BUT if you ever do need to run Windows/Windows software, an Intel Mac should be able to run it at 90% of the speed that a PC with the same CPU would, much like Virtual PC running on a PC. And you may not even need a Virtual PC layer, just a library installed on your system (WINE).

Hope this helps...

"Scott, let me simplify - 6X7cm Velvia annihilates 35mm equivalent which annihilates 6MP DSLR."

Below is one of Les' 4000 ppi scans of a flower "annihilating" one of my lowly 6 MP "digiscam" flower shots. Wow...that glorious grain, those soft edges, that dull look to the colors.

"Be secure in your thought that no one will blame you for using a lower resolving DSLR."

Be secure in the kowledge that no one will ever give a flying rat's a** that your 35mm film plus 4000 ppi scanner can theoretically record a few more lpmm than an APS DSLR. It's the *image* that matters. I agree some subjects printed to some sizes require a great deal of resolution to look their best. For those subjects/print sizes, you *do not* shoot 35mm. Like Scott I am simply amazed at the 35mm fetish crowd that obsesses over how many lpmm can fit on a miniature frame of film rather than just shooting a larger format if they need more resolution.

"As a test engineer, you note that I used the same lenses on both bodies, giving neither the advantage, same settings and such."

As a test engineer, you botched the 20D shot so bad that I didn't even have to try to out resolve it with a lower resolution 10D. Part of me can't help but think you purposely botched the 20D shot to make the differences seem more dramatic, but I can't get you to go on record with the aperture settings for all your shots, so we will never know.

"These examples are what I have observed and therefore put forward for public critique - not attack."

No, Les, you started off from post one on the attack, misrepresenting, misquoting. You wanted a p***ing contest. That is and was the greatest flaw in your methodology.

"So to get back on topic, no a well scanned 35mm, such as the 4,000dpi or better, is more then the equal of todays current DSLR all the way to the 1Ds in terms of resolution..."

My 10D composed as if it had a full frame sensor annihilated...no...SPANKED your best film scan, but 35mm is "...the equal of todays current DSLR all the way to the 1Ds..."?

To my knowledge you've still never posted your 1Ds mkII shots. If they didn't bury your Velvia 50 scan then I'm sorry Les, but the problem was "test engineer error."

***

Disclaimer: to the film fans in this thread, I do not mean to "attack" your medium. I have a photo album full of great 35mm prints, shoot with friends who use 35mm, hope film sticks around for a good long time, and think a good 35mm workflow produces great results at common print sizes. This post is directed at Les, not at 35mm in general.<div></div>

Is film dead? No. Will it be dead soon? Most likely not. It really isn't a difficult thing to produce and there are enough film bodies and film users to keep production going.

Is film becoming more difficult to use, especially in terms of quality control, because of less business and closing labs? Yes. Could the finer films fall by the wayside in favor of a couple consumer grade emultions? It hasn't happened yet, but it is a possibility. Will the lab situation get so bad that users are forced to use mail order for processing? I think this is inevitable in the future except for the largest cities. It's already a reality for many.

It's your choice. There are certainly plenty of high quality scanners if you wish to continue to use film. Good films and decent processing are still available, though you may have to deal with mail order for them.

Still, if you like the look of certain films, have your workflow nailed down, and wish to continue, by all means. It's the image that matters.

If you have Minolta glass you want to use and the money, just pick one up. What does it matter if you have both a Minolta and a Canon body?

You may end up choosing one over the other and sticking to one system. Or you may just cherry pick the lenses you want from each system. Either way whatever works....

Yes, please do visit the Film and digital resolution tests Part II. Make sure you get to the end. You'll see what a sharp 6 MP image looks like (vs. a soft 8 MP image). You'll see what a 9.6 MP full frame 10D would look like if it existed. And you'll see Les flip out when his deeply entrenched, emotional position gets knocked out from under him.

Les has flipped out. Seriously.

"Can anyone with experience (key word here) using one of the high-end DSLRs (e.g. 5D) comment on relative quality?"

You can download 5D samples at: http://web.canon.jp/Imaging/eos5d/eos5d_sample-e.html

Your scan is a good capture of Provia. I think you'll find the 5D sample portraits better. It's up to you whether it's worth it.

I will say that in addition to time savings, you will gain far higher ISO performance. The 5D will do pretty much as well at ISO 400, and won't lose much performance even at ISO 800.

"Except of course the FBI who actually studied it and took measurements and spent about a year getting comments before publishing their results."

What film did they use in their "study" Neal? On what films did they perform measurements?

We can all wait for hell to freeze over before Neal gets back to us with an answer. The reason is that this "study" doesn't cite any specific film, or other test parameters, at all. Why you ask? There was no study and there were no tests. One tech writer working for the FBI grabbed some average lpmm ratings for generic film classes (i.e. instant, color, B&W), then multiplied them by the wrong number for Nyquist to try and arrive at the equivalent MP for an internal field manual, not a "study".

"I guess it is natural for "artists" to value feelings over facts but when people ask for facts it seems a shame to substitute feelings without some type of disclamer."

Put a disclaimer on your spin of the "facts" Neal. The only fact that can be derived from that field manual is that the tech writer who authored it knew less about photography than a child with a disposable P&S. So don't spin it as any kind of "special study" by the FBI.

"> You didn't provide the final pixel dimensions for your file, so there was no way to reliably perform the adjustments you claim are absolutely necessary.

Well, the full frame I posted is exactly 1/8 of the original file, as stated in my posting (4000dpi down to 500 dpi). Depending on where you crop exactly, the picture area is about 5650 pixels wide."

"Depending on where you crop exactly..." was exactly why I didn't venture to guess what the original size of your image was.

"Do you think you could fairly compare negative/file quality in two prints, one 50 cm wide, the other 27 cm?"

I think your film scan, at pixel view, looks like crap. Got it? It's an ugly, mottled mess. A classic example of why Scott often says film breaks down past 2000 ppi.

You had to reduce it to 54% and run it through NeatImage to get close to what comes off a DSLR, pixel view, without any effort.

Which *** surprise *** brings us back to the original statement I made: "Yes, thank you. If you've got a good scanner, you use the finest films, and your reproduction chain is nailed down, you can get very good prints from 35mm. But it's a lot more work than with a DSLR."

I side * twice * that 35mm on a desktop scanner with a proper workflow will produce good prints. But somehow you took this as an invitation to get into a p***ing contest over digital vs. film detail under pixel peeping conditions.

"And if you could read,..."

And if YOU could read, you would know that this thread was about D70 files vs. Gold 100 negs printed at a local lab. Not peeping pixels from a desktop scanner at about twice the magnification where film starts to break down. The person who started the thread wanted to know why his film prints didn't look so good. It's because to get film to look good requires more work than simply printing at the local lab (unless that lab happens to be one of the dying breed of labs interested in putting in that extra work for you...and you're willing to pay). On the contrary, simply dropping a CF card in a Frontier generally produces very good results.

Did you put more work into your sample scanned image than simply dropping it off at a lab? Yes, you did.

"> I made a single post offering the opinion, shared by many, that 35mm film required considerable effort just to match a DSLR.

And as in all posts before you didn't state what you compared, and your claimed *facts* are mostly just opinions, too."

So it's not a fact that sitting at a desk scanning and post processing an image, a good 10 minutes a frame, is not more work than plugging a CF card into a Frontier and taking the prints home???

"> Maybe you have a thing for grain. Maybe you feel that the extra effort makes the images more meaningful to you. Maybe ...

As I said before: When *I* compare the prints *I* can make from files from *my* DSLR with prints *I* can make from *my* film scans *I* prefer the prints from film scans."

Good for you. When Anton compares the prints *he* can make from files from *his* DSLR with prints *he* can make from film printed at the local lab (equal effort), *he* prefers the prints from the DSLR.

Some of us tried to explain to him why. Needing everything film to be superior in every way, you jumped in with an irrelevant p***ing contest.

You know, at this point I think I'll be glad when film is dead. No more stupid contests because somebody dared to claim that digital was in some way easier or better....

"We are hiring a photographer for our wedding and have the option of either having the wedding shot digitally and the images given to us as TIFF files on a DVD or we can choose to have it shot on film and we'll have the negatives."

First off, what are they shooting when shooting film? I'm going to tick some people off when I say this, but I would never hire a photographer to shoot *my* wedding on 35mm. The grain and tonality just cannot compete with MF or DSLR digital, especially when the lights get low at the reception.

If you're going to spend the money, make sure the capture is MF or digital. If your choice is between digital and 35mm, choose digital.

"We like the idea of having high-resolution digital files so that we can do our own postprocessing. However, when it comes to photos lasting a lifetime, would film always have an advantage over digital or are there ways to store digital files safely?"

It's easy to make bit-for-bit perfect copies of a digital source. This is not possible with negatives. Any reproduction will always involve some degradation. So if you're serious about archiving, digital gives you the tremendous advantage of being able to produce multiple copies for off site storage. I don't care how archival a film is (and I don't think the color portrait films are very), one copy is dangerous.

Buy *good* CD-R or DVD-R media and make 3 copies of your wedding photos. Keep two on site and two off site *at different locations.* Parent's houses for example. (If you have the HD space, keep a copy there and move one of your on site copies off site.) You just insured yourself against fire, theft, flood, etc, etc. Make sure they're stored away from sunlight in jewel cases, no lables or writing on the disc itself.

Check them every 3 years just to be safe. That way if one disc was poor to begin with and is degrading, you'll notice it (it won't read or the computer will struggle to read it) and be able to make another copy from one of your other 3 sources.

"Why not belt and suspenders? Shoot it in film and scan the negatives."

Because this is much easier said than done. Professional scanning is not cheap. Neither is home scanning, plus doing it yourself is very time consuming. Unless the wedding photographer is offering both film and the film scans in a package, the scanning will likely never be done. You're back to one copy in a house that can burn, be flooded, etc, etc.

"A friend of mine just had to have some 5 1/4 flopies converted and most of the bids were in excess of $100 each!"

You can still buy 5 1/4" floppy drives! Were they PC floppies or one of the older 8-bit computers? Tell your friend that if he has any more 5 1/4 floppies to convert to 3 1/2 or CD-R, to buy a drive and do it himself. Or if he doesn't want to go through the hassel, I will GLADLY do it for him much cheaper than that. He pays shipping and handling and, depending on the number of floppies to convert, we'll work out a deal.

"I plan on shooting a lot of pictures and I don't want to get bogged down with processing costs I don't need..."

If that's the case, then why 35mm?

Shots from my 70-200 f/4L and 300 f/4L IS have a level of sharpness and detail, in 8x10, that's hard to match, and that's when they're wide open. Color and bokeh are also fantastic.

I wouldn't trade them for the consumer zoom.