howard_slavitt2

Eyal -- I tried the LX3 for a few weeks and ended up selling it because the quality just wasn't good enough for me after using a Canon 5D. However, I bought a Canon g10 about 1 week ago and absolutely love it. The quality, at 80 iso compares to a 5D and L lenses with respect to resolution. It is a fantastic and versatile camera. ISO 400 on the G10 shooting raw is better than ISO 400 on the LX3 in raw. I'd get the G10 if you can. The g10 is also better suited to rangefinder like shooting -- you can use the custom functions (2 of them) set to a particular focal length and focal distance, whereas the LX3 doesn't let you do this.

Get the 5D. I just sold my rebel xti and decided to get a second 5D. I've had my original one since they first came out. They're fantastic cameras. I was hoping Canon would upgrade it at PMA and I could buy the newer camera, but at this point I'll buy a second one. I need two bodies. I much prefer the way the full frame camera "draws" to a crop sensor -- depth of field is shallower, the dimensionality is different. Also, Canon's primes are designed for full frame, not crop sensor bodies. The 24-105mm is a great lens, as is the 17-40mm. For the weight, money, and flexibility, these lenses and the 5D is still, IMHO, far and away the system to beat.

Look at the blog at diglloyd.com . Yesterday's post has some great comparisons showing the resolution difference between a 1DsIII vs. that of a Nikon D3. The Nikon D3 should have roughly the same resolution as a 5D. These are about as controlled of comparisons as you're going to find. IMHO, the 1DsIII blows the Nikon D3 out of the water for resolution. So, if you're going to use a tripod all the time, mirror lockup and the best possible lenses, you'll get a huge quality jump by getting a 1DsIII. All of the above variables are critical, and if you cut corners on any of them, your quality will start to diminish rapidly . . . Oh yeah, I almost forgot, the 1DsIII will be diffraction limited by f8 (like, for example a 40D). You should start seeing resolution diminish by f10 or f11, and will certainly see a large fall off by f16. A few months ago, someone who removed the AA filter from their 5D sent me about 10 raw files. I worked with those files quite a bit, and compared them with the files from my unmodified 5D. I thought the increased resolution from having removed the AA filter was minimal. I also didn't like the change in the quality of the files; to me, and this is very subjective, they looked more digital and less natural.

I have this lens, and have previously owned a view camera. The problem with tilting with the lens is that it's almost impossible to see the effects of tilt in the viewfinder. Thus, you need to tilt based on experience (or knowledge) but not based on what you see. In general, I find that very little tilte and stopping down to f11 gives the best results to hold everything in focus. There's a spreadsheet that someone did floating around the web based on Scheimpflug calculations that shows, based on tripod height, how much tilt you need. I think the chart is based on infinity focus. In general, I use that chart, but only use about half of the calculated tilt and stop down to f10 or f11. My procedure is a bit more complicated than that based on a lot of practical tests I ran. If you want to get best results, my recommendation is to search for the chart (you should find it pretty quickly), then run various tests with notes on how much you tilted, focus point, etc., and examine the results to determine a procedure that works for you. . . . It's a special lens, and worth the effort, IMHO. Also Live View may make it easier to use in the future.

I'm a contrarian. Given that everyone is expecting the same thing, I now expect it will be something different. Or, at a minimum, I now expect that it will not be released until late next year.

Another big advantage for some is that the 24-105 weighs A LOT less than the 24-70mm, which is a monster. Others do not care about this feature.

 

"The 24-105 has longer reach and I.S. It's got NOTHING else over the 24-70."

So, Warren, your blog left me hanging -- how did you find the weight of the 1DsIII? In particular, how much heavier does it feel than the 5D? (I know, about 12 ounces, but how much heavier did it subjectively feel?)

DXO Optics pro automatically corrects for chromatic aberrations based on each specific lens and camera. It does a fantastic job. This is a problem with almost every Canon lens (I have mostly "L" lenses), especially the closer they are to wide open. It depends a lot on the lighting, backlighting with strong contrast between light and dark, makes these kinds of aberrations more likely, all the more so with the lens wide open. Lightroom does a very good job for correcting CA, but not as good as DXO Optics Pro.

Ken Rockwell has an excellent article on how to use the Canon 15mm fisheye well on his website. I'm normally not a fan of Rockwell, but his explanation and two portfolios with examples of it in use are excellent.

Actually, Stephane, I think you're incorrect. The 1DsIII has the same sized pixels as the 1DsII, just more of them because there's less space between the pixels (technological advances). So the 1DsIII should not place any more demands on lenses than the 1DsII, it just uses more of each lens (i.e. there is not as much space in between pixels, in-between coverage of the lens that is not used). Canon may have tuned the anti-alias filter to be weaker, thus not softening the image as much (for a tradeoff of more aliasing). . . . We'll need to wait and see how much resolution the 1DsIII truly offers.

 

quote from Stephane Bosman: "There is no way the difference between the 1Ds MkIII and the 5D can be likened to the one between 10D and 5D. The 10D is sensor-limited, the 5D is lens-limited with many lenses already, the 1Ds MkII with most lenses.

 

The 1Ds MkII will demand the finest lenses to exploit its very high pixel density."

I'm fairly certain that it's not. I've read somewhere that it's lighter than the one in the 1DsMkII and supposedly gives many times longer performance.

DxO outputs to linear DNG files. I have been processing Canon 5D files in DxO (just using chromatic aberration and vignetting corrections) and outputting them to LINEAR DNG files, and then opening them in Lightroom. They are essentially identical to files that I open for the first time in lightroom. The difference is that the DxO, linear DNG files, have already been demosaiced, and have also had the few corrections I've done to them applied. The Raw DNG files, are not demosaiced; for those files Lightroom does the demosaicing.

You'll need to verify this, but I've read somewhere that the G9 can take about 1 1/2 images per second, including raw, and that it can essentially write raw files as fast as you can take them. It apparently has no internal memory but processes raw quickly and quickly writes the files to a card.

I would look closely at the Canon G9. It has MUCH faster RAW write speed than the Ricoh. The downside is that it only goes to 35mm equivalent focal length on the wide end. But you can get an external screw on lens, to get it down to about 27 or 28mm if I recall correctly.

I use the EE-S screen on my 5D. It's normally dark, as a previous poster said, but I had it treated/brightened by Bill Maxwell (he's well known and does screens for all sorts of cameras, including view cameras, Rolleis, etc.). It's now a pleasure to use for manual focus in all lighting situations; it's bright and images pop into focus.

Get a Contax 24-85mm f3.5-f4.5, converted to work like an autofocus lens on Canon EOS. http://en.conurus.com/ Seriously. I have one, it's fantastic. The Canon 28mm f1.8 is also excellent -- my copy of it bests the copies of the Nikon 28mm f2.8 AIS, and Contax 28mm f2.8 that I tried. The Leica Elmarit-R 28mm f2.8, current version, is reputed to be the absolute best 28mm SLR lens available, but there are varying reports that it hits the mirror on the 5D requiring a mirror shave.

Don't fix it replace it. The odds of them being able to fix it so that there are no effects on the optics seems slim to me. I'd never feel comfortable using a lens after what happened to yours.

Look for Clayton Jones' website. He has a series of articles on how to use ABW with non-Epson papers. His system works great (for me on an Espon 2400 and now on an Epson 3800), better than the QTR Rip worked for me on an Epson 2200. It's also a lot more flexible.

I am looking to get a Nikor 35mm f2.8 pc lens, the latest, black knob version.

I will be using it on full frame, not a 1.5 crop camera. I haven't been able

to find much information specific to the latest formulation for the 35mm pc

lens. In particular, I am interested in: (1) what is distortion of the lens

like, how does distortion change when fully shifted? (2) how well controlled is

chromatic aberration, and (3) how sharp is it when stopped down to f11 and

shifted, say, approximately 7mm? Thanks in advance for the help. . . [Not

that it particular matters to answer the above, but I will be using it with an

adapter on a Canon 5D with an adapter, and I will be using the shift for rise

or fall and then taking about 5 images in vertical (portrait) orientation,

panning around the nodal point to stitch together in Photoshop CS automerge,

which I find does an almost perfect job of stitching panoramic images.]

There's no comparison between photomerge in CS2 and in CS3. Photomerge in CS3 is incredibly good, and easy to use. It gives you quite a bit of flexibility as well. I've tried a lot of programs, but never spent the time to learn PTGUI (I tried, but it was slow and not especially intuitive to me at least). I love photogmerge in CS3, and am now starting to build a lot of my workflow around it. The "persepctive" stitching option in CS is fantastic, the equivalent of a flat stitch using a shift lens (but having created the image using a pan head and nodal plate).

I've had both. I recently sold the 2400 to buy the 3800. The 3800 prints larger (17" wide), has bigger ink cartridges, AND has more consistent color. The 3800s are all calibrated at the factory to a known color value. The 2400s are not. It makes a big difference in being able to use the canned profiles. They're extremely accurate with the 3800. The quality of both is almost identical. My 2400 and my 3800 are giving me similarly accurate black and white out of the box -- both are fantastic.

This is normal. I'm very experienced with Macs, and had the same question (I have 7 GB, and only 3 GB are recognized by CS 3). It turns out that CS 3 only recognizes 3 GB maximum, something having to do with being a 32 bit vs. 64 bit program. However, the extra ram, though not recognized within Photoshop, is used for a swap disk to speed things up, the system, and other programs. So it all contributes to making your workflow faster.

The newest Pentax 35mm (the more expensive one) is generally considered to be the best 35mm wide angle made for any 645 system, better than the Contax or Mamiya. The older Pentax 35mm is not well regarded. I'd spent the extra money and get the newer one.

Doug Fisher knows. He makes some great film holders to maximize the film scanning ability of flatbed scanners.

I also come from a large format background and hate distortion. I've not noticed any with this lens. I use it for close focus, mid-ground focus, and infinity focus. The biggest problem with the lens is that the corners are soft until about f8/f11. The softness is due to curvature of focus, i.e., it's not at all a flat field design. So for instance, I recently shot an "environmental portrait" at f2 or f2.5; the person in the middle of the frame was in focus; most of the rest of the frame was out of focus; but the lower corners, which were in the foreground and very close to the lens were in sharp focus. That said, other than the example given, it's not a problem with this lens in 99% of situations, unless you want everything sharp corner to corner at f4 or f5.6. (I use the lens almost exclusively at f2.5 - f3.5, or at f9-f16.) Great lens: inexpensive, tiny, light.