charlie_vigue

The D70 is actually in front of the leaves I bet ..... [GD&R]

I'm interestec in the 24-60/2.8 and 18-50/2.8, both look to be good for a shorter std zoom on my 10D

Got mine from eBay member 'zenitar' (name was a dead giveaway) and it went super smooth. I love mine.

I think I understand what you're saying, but if selecting a specific post-process path makes a painter, does choosing a specific film/paper/process make a chemist? I'd say of course not, you?

Can you shoot Tri-X and then decide to make it look like Velvia?

This was out of a slightly foggy field. Oh well.<div></div>

Maybe the DCT compression throws away the high frequency noise? It is lossy, after all.

What I want to know is the MTBF of the microcontrollers in my microwave oven. I can't find that information anywhere, and I'm not buying a new one until I find this data! It's a conspiracy I tell you.

 

What? Door switches? Oh, nevermind then.

<i>I have some decent JPG shots that could have been much better if shot as RAW. It's because of those that I quit shooting JPG.</i><BR><BR>Indeed, a few of those broke me of JPG too. Of course, you can shoot 'both' on the 10D, more or less.

Digital is sort of funny too. The ISO setting isn't all that determines noise. Shooting a longer exposure at the same ISO results in a noisier image; some types of digital noise accumulate over time. So in digital land there's less penalty to switching to a higher ISO to shorten the exposure than there is to reduce aperture. Odd, eh?

 

The practical application is that one can use a higher ISO to capture sports action or similar things with minimal penalty in digital systems.

 

Temperature also affects noise in digitals. Maybe shooting ski competitions is the ultimate best case for digital photography?

Jean,

 

You make an interesting (but obvious) observation, but I don't see what it has to do with the case of encoding 12 bit data into a 16 bit format, unless the encoder was adding some sort of random dither or something. You're looking at apples and oranges. But I like both apples and oranges, so there's that .... good comment then!

 

Bas,

 

Thanks for the empirical cofirmation. The Canon RAW format even does a good job of losslessly compressing 12 bits/pixel (12 real data bits, nothing set to zero) down to about 8 bits/pixel with great consistency. In a quick test I just ran, loading a 24 bit 6mp TIFF and converting it to 48 bit, then saving with LZW compression blew it up from 18mb (24 bit, uncompressed) to 20mb (48 bit, LZW) while saving the 48 bit file uncompressed (36mb) and then zipping it results in a 14mb file. I would have thought with a bit of extra domain knowledge the TIFF folks should have done better, but they don't. PNG gets the 48 bit version down to 12mb, or 3:1 ratio over the raw data. That's more what a good compressor with some domain knowledge should do I'd think.

 

And, on the subject of data formats, the TIFF folk would have probably been best served (if they were concerned with data size) to arrange the data layout or compression specification for best compression rather than muck around with 8 bit, 9 bit, 10 bit, 12 bit, 14 bit, and so on, data depths.

If it's any consolation, people say the same thing about the 10D and 20D. It's a dSLR thing, not a Nikon thing, as others have pointed out already.

Bas,

 

What you say is almost true, except maybe the part about TIFF. I'd more suspect the folks who wrote that standard were interested in simplicity, and also probably understood that data in 16 bit packages which all have the four LSBs set to zero will typically compress really really well with any reasonable compression algorithm.

 

For a compressed set of image data, I doubt a native 12 bit format would be much smaller than a 16 bit with a 12 bit payload, and the format will be simpler. You're going to want to use compression anyway.

Andrew above raises a good point. I have the 28-135 IS and one thing using has taught me is this: Image Stabilization won't freeze subject movement. It seems so obvious in retrospect ....

 

Seriously, the 28-135 IS works so well that even in non-sports situations subject movement is often the limiting factor. IS only does so much.

I have the 10D as well. I use the Sigma 70-200 f2.8 handheld (in good light) or on a monopod, no problemo. And it costs about $600, with excellent optics. What you give up is the IS feature and some close distance focus ability. And the Canon name. This is generally considered one of Sigmas best lenses in most places I've read reviews, for what that's worth. I got the 70-200 and the Sigma 1.4 teleconverter EX delivered to my door for about $700.<BR><BR>

Something worth mulling over I think. A Manfroto monopod will set you back about $30 and is small, light, and fast to use.<BR><BR>

 

If you are tending towards the 100-400, I'd <b>not consider</b> the Sigma 80-400; it simply isn't in the same league overall I think. The Sigma 120-300 f2.8 is rumored to be exceptional, but it's not at all cheap. Well, compared to other 300mm/2.8 lenses maybe, but at $1700 it's huge and sort of costly. Coupled with a 1.4x TC it might make a really good 170-420/f4 .....

Taking it even further, rounding errors in manipulating 16 bit data can end up truncating it to a much smaller number of bits, just as working on an 8 bit image in 8 bit mode can do the same. 12-14 bits of useful data is a really good fit for minimal loss manipulation in a 16 bit domain.

On your 10D the difference between ISO 50 and ISO 100 is enormous. Try it and you'll see what I mean.

 

For the more generic question, a lot of those 'rules' have gone out the window with improved emulsions and digital.

 

[Hint: There is no 50 on an EOS-10D]

The Sigma 70-200 f2.8 is about the same money as the Canon f4, and is a reasonable alternative as well. No IS, but it is optically fine, well made and the HSM is fast and quiet. I like mine, I'd only get the f4 if size is a concern. Eventually I will want the IS, but I couldn't justify the extra $1100 over the Sigma. YMMV.

Or you could just get the f1.4 and have USM to boot.

I use it on the 10D as well. The only thing I miss is IS, I'd never buy the Canon non-IS version myself. If I were going to drown, I say drown in the ocean and get the IS version. Otherwise it's between the Canon F4 and the Sigma. The Sigma is great with the 1.4 TC as well, BTW.

 

The only major (for some) issues are weight (it's a lot of glass) and the minimum focus distance is (IIRC) 2 meters, so it's not really a good macro lens at all. I've even used it for a few portraits, although it is a bit intimidating in some cases.

 

One thing you might consider is the 135mm prime; it's fast, good and cheap, but not a zoom. Same for the 85mm.

Err, yeah, I mean Michael..... you know. :-)

Tim, can you keep us posted on the Sigma 24-70? It is of interest to me, I could see myself haveing both it and the 18-50, for different venues.

I got the original Vosonic 20gb XDrive (rebranded for RadioShack) for $100. If I was worried about losing data between the campsite and home I'd just buy two. They're small and the cost per byte can't be beat. Also, they double as a USB hard drive and card reader.

The Zenitar 16 is a good lens, and you can get one for $130-$150 with an EOS mount. At that focal length focus is easy since DoF is very deep. I love it. I lust after the 24-70 2.8L, but I think my budget will have me getting the new Sigma 18-50 f2.8 instead.

The same sort of comments came out with the 10D. User error.

 

- Learn the digital darkroom. The data collected by dSLR will be 'soft' before you do the work of turning it into an image.

 

- The Focus/Recompose method can produce backfocus in many cases.