How important is high ISO?

[dllavaneras]

I've been testing out Noiseware Pro and its has been doing a good job so far and only costs about $40 dollars for the stand alone.

How would you rate Noiseware pro? I've searched for a comparison between this and Noise Ninja, but have failed to come up with objective results.

 

[wmmk]

just one thing, do you really need this level of a camera? If I were coming from P&S, I can't really comprehend even thinking about a 5D. It's your money, but as I always say, it's what's behind the viewfinder that counts. On flickr's camera finder, I see plenty of photos shot with budget and prosumer SLRs that have much better lighting than quite a few taken with 5Ds, 1Ds, d200s, D2Xs, and D2Xs. That said, if you're considering the 30D, think about the Pentax K10D. It's a better camera and a much better value, and I'll say this even though I know I'll be flamed by Canon/Nikon people because I really believe it.

 

[dllavaneras]

That said, if you're considering the 30D, think about the Pentax K10D. It's a better camera and a much better value, and I'll say this even though I know I'll be flamed by Canon/Nikon people because I really believe it.

Personally I don't believe it's better camera (I'm a Canon lover myself), but I'd love to see what arguments you have for making that claim. Still, my main reason for choosing Canon when buying a DSLR is the lens lineup. I need extreme macro for my job, and Canon delivers what I want.

 

[bhdean]

How would you rate Noiseware pro? I've searched for a comparison between this and Noise Ninja, but have failed to come up with objective results.

I ran a few of my ISO 800 pictures through it last night and it did a good job. Really, the only complaint I have right now is the inability to import my RAW files directly into the stand alone program. I'll probably download the Photoshop plug in trial version tomorrow and give that a go. Its a very easy program to use. Best part is they will give a 15 day trial on it. I'm going to shoot some test shots at ISO 1600 and see how it handles those. That is where I'm the most concerned about it. At 800, I usually don't get to much noise anyways.

I'm happy with its performance thus far. But, I have only been playing with it a few days. Hopefully, this weekend I'll get some time to take some shots; that will really test out its ability.

 

[dllavaneras]

Good to hear that. Noise ninja does a decent shot at saving my ISO 400 shots, which are just hideous. I sooo want a DSLR....

 

[coldrain]

just one thing, do you really need this level of a camera? If I were coming from P&S, I can't really comprehend even thinking about a 5D. It's your money, but as I always say, it's what's behind the viewfinder that counts. On flickr's camera finder, I see plenty of photos shot with budget and prosumer SLRs that have much better lighting than quite a few taken with 5Ds, 1Ds, d200s, D2Xs, and D2Xs. That said, if you're considering the 30D, think about the Pentax K10D. It's a better camera and a much better value, and I'll say this even though I know I'll be flamed by Canon/Nikon people because I really believe it.

The Pentax K10D is a nice camera... but better than Nikon/Canon offerings? To me a camera is good when its image quality it top notch.
The Pentax is not up there in that respect, very noisy in comparison for instance. And then there is the lenses we discussed in this thread. Is the Pentax 12-24 f4 out yet? And no 70-200 f2.8 Pentax I believe? I think you may be able to track down a Sigma 70-00 f2.8 if Sigma offers that for the pentax, but then without HSM internal motor...
Again, it is a nice camera, nice for Pentax fans. But better? I do not know why you think it is better. Cheaper and very complete, yes.

 

[coldrain]

Good to hear that. Noise ninja does a decent shot at saving my ISO 400 shots, which are just hideous. I sooo want a DSLR....

As far as I know, neat image does a bit better job than noise ninja. It loses less detail and makes the results a little less soft.

 

[compuwar]

As far as I know, neat image does a bit better job than noise ninja. It loses less detail and makes the results a little less soft.

Here's the most comprehensive head-to-head I can find:

http://www.michaelalmond.com/Articles/noise.htm

I've seen critiques for this here and there- biggest one was that Grain Surgery does significantly better if you don't let it pick the sample sites. Many other folks rate Neat Image higher than NN, but you'll find opinions on either side of the fence. It's enough of a moving target still, and differing images have differing noise enough that having two in the bag isn't all that bad an idea.

 

[EstorilM]

Just to put things into perspective for you - as someone mentioned in here earlier, most sensors on the market today offer very good IQ. However, someone has to pull them apart in a lab and pick a winner, so like I said.. sometimes we need some perspective! When we're not at ISO1600 / 200%, underexposed, etc.. things aren't so bad haha. 😀

The OP asked about having to push your camera and stuff, and how it performed. Well, talk is cheap - so this is what I get in real world use. My goal was sharpness, and I needed a minimum of a ~1/800 shutter.
Camera Model: NIKON D200
Exposure Time: 1/800 sec
F-Number: f/3.2
Exposure Program: Aperture Priority
ISO Speed Rating: 1000
Focal Length: 36.00 mm
Lens Range: 24.0 - 70.0 mm; f/2.8
End result:

Here's a 100% crop of the shadow area, however keep in mind that I was panning with the subject, not the background, so there's motion blur in the rest of the frame (and the crop.)

I was more than happy with the camera's performance in this case, especially considering it's price point (again, more than $1000 cheaper than the 5D the OP was considering.) Besides, high ISO noise is about the weakest point of the camera anyways! ..which obviously isn't that bad!

 

[coldrain]

Just to put things into perspective for you - as someone mentioned in here earlier, most sensors on the market today offer very good IQ. However, someone has to pull them apart in a lab and pick a winner, so like I said.. sometimes we need some perspective! When we're not at ISO1600 / 200%, underexposed, etc.. things aren't so bad haha. 😀

The OP asked about having to push your camera and stuff, and how it performed. Well, talk is cheap - so this is what I get in real world use. My goal was sharpness, and I needed a minimum of a ~1/800 shutter.
Camera Model: NIKON D200
Exposure Time: 1/800 sec
F-Number: f/3.2
Exposure Program: Aperture Priority
ISO Speed Rating: 1000
Focal Length: 36.00 mm
Lens Range: 24.0 - 70.0 mm; f/2.8
End result:

Here's a 100% crop of the shadow area, however keep in mind that I was panning with the subject, not the background, so there's motion blur in the rest of the frame (and the crop.)

I was more than happy with the camera's performance in this case, especially considering it's price point (again, more than $1000 cheaper than the 5D the OP was considering.) Besides, high ISO noise is about the weakest point of the camera anyways! ..which obviously isn't that bad!

Did you shoot in RAW or JPEG? And what noise removal settings did you use? And what is the cause of that strange softness in the 100% crop that looks like the softness you get with bad CA but without the CA? It looks a bit weird, also the green leafs look a bit weird..

 

[EstorilM]

Did you shoot in RAW or JPEG? And what noise removal settings did you use? And what is the cause of that strange softness in the 100% crop that looks like the softness you get with bad CA but without the CA? It looks a bit weird, also the green leafs look a bit weird..

JPEG fine, it's event photography so I can't do any post processing on stuff really (people buy the pictures less than an hour after I take them.) In camera NR was medium, that was it. Most of the softness was motion blur (the horse was moving pretty quick) um, some was probably the in camera NR, some was probably the crappy Sigma lens haha.

That scene is a CA nightmare, f/3.2 with overexposed spots of light coming through little specs between correctly exposed leaves 😱 - but, not really any way around it with that lens (also why I wouldn't use 2.8.)

Don't pick it apart too bad though, I'm happy with it.

 

[compuwar]

Did you shoot in RAW or JPEG? And what noise removal settings did you use? And what is the cause of that strange softness in the 100% crop that looks like the softness you get with bad CA but without the CA? It looks a bit weird, also the green leafs look a bit weird..

Looks more like the cropped stuff is outside the plane of focus to me.

 

[EstorilM]

Oh yeah haha.. duh. oops 🙄 At f/3.2 even the logs are going to be slightly OOF.

 

[coldrain]

Looks more like the cropped stuff is outside the plane of focus to me.

It clearly is not totally in the plane of focus, but that is not what I mean.
With lenses that CA quite a bit, you get this uneasy double image unsharpness, that makes it very uncomfortable to look at. Like some camera shake unsharpness can give at times too. This crop has that "quality", and what is more the leafs and details have gotten a very "flat" and unnatural feel to them. That is why I wondered about the image processing.

The woman and the horse look very nice in the resulting photo though, when seen in this very low resolution at least.

 

[compuwar]

It clearly is not totally in the plane of focus, but that is not what I mean.
With lenses that CA quite a bit, you get this uneasy double image unsharpness, that makes it very uncomfortable to look at. Like some camera shake unsharpness can give at times too. This crop has that "quality", and what is more the leafs and details have gotten a very "flat" and unnatural feel to them. That is why I wondered about the image processing.

The woman and the horse look very nice in the resulting photo though, when seen in this very low resolution at least.

Probably motion blur in the leaves in the background. If you mask them out, the OOF areas don't look nearly so strange.

 

[EstorilM]

..anyways, I was looking for noise artifacts. 😀

 

[coldrain]

Probably motion blur in the leaves in the background. If you mask them out, the OOF areas don't look nearly so strange.

Well, at 1/800sec and a perfectly frozen jumping horse I do not think that what we see is motion blur from leafs, especially since the doubleness of it is especially visible in the logs themselves.

On closer inspection I see it actually is CA that gives that double image feel to the crop. I do not know if the CA has been partially processed away, but look at the vertical lighter plank on the right, for instance. on the top of the crop its side has a certain thickness, and a slight red tint to the left most edge of it. The moment it crosses the lighter trunk, the thickness of the plank almost seems to half, and the red edge is gone.

Below the metal wheel, the edge of the plank has a very red tint to it.
Now look the right of other dark parts, like the vertical metal bar... again a distinct red CA line. On lighter elements you seem to see green bleeding in to the left, like with the highlights in the leafs in the background. So my educated guess is that it was some quite strong CA after all.

This still leaves the flat feel of things, which seem to suggest some noise reduction algorythm.

..anyways, I was looking for noise artifacts.

The flatness of the leafs, metal parts and all could be called noise artifacts... They are not noise, but artifacts from de-noise-ing... just like sharpening artifacts are not sharpness but for instance strange light/dark halos around contrasting objects 😀

*edit* : Sorry about picking it apart, haha! I totally missed the part where you explained the CA nightmare. I can understand why you are happy with the shot.

 

[compuwar]

..anyways, I was looking for noise artifacts. 😀

There may be some chroma noise in the leaves, it's really difficult to tell because of the motion blur going on there (or maybe I'm out of pixel peeping practice!) It's a very good shot though. The time I shot a show jumping arena I realized how tough these shots can be to nail. Have you looked at it at 8x10 printed? That's my usual output size (though I can't say I've ever had my D2x set over ISO 800 with the new firmware it should be pretty comparable, so if I hit the chance, I'll give it a go.)

I'm sure the rider would be more than happy to pay for a copy to hang on the wall. That's what really matters!

Well, at 1/800sec and a perfectly frozen jumping horse I do not think that what we see is motion blur from leafs, especially since the doubleness of it is especially visible in the logs themselves.

On closer inspection I see it actually is CA that gives that double image feel to the crop. I do not know if the CA has been partially processed away, but look at the vertical lighter plank on the right, for instance. on the top of the crop its side has a certain thickness, and a slight red tint to the left most edge of it. The moment it crosses the lighter trunk, the thickness of the plank almost seems to half, and the red edge is gone.

Below the metal wheel, the edge of the plank has a very red tint to it.
Now look the right of other dark parts, like the vertical metal bar... again a distinct red CA line. On lighter elements you seem to see green bleeding in to the left, like with the highlights in the leafs in the background. So my educated guess is that it was some quite strong CA after all.

This still leaves the flat feel of things, which seem to suggest some noise reduction algorythm.


The flatness of the leafs, metal parts and all could be called noise artifacts... They are not noise, but artifacts from de-noise-ing... just like sharpening artifacts are not sharpness but for instance strange light/dark halos around contrasting objects 😀

If I put my nose right up to the LCD, I can see some CA on the leaf edges and the light colored stick. I'll save you the hassle of pointing out the 1DS mkII has an extra filter to suppress sensor reflected CA 😱

Nikon Capture's CAC will likely clean it right up should you need to print at 3' x 4' or whatever... At normal viewing distance, I can't imagine it's even a factor.

 

[EstorilM]

There may be some chroma noise in the leaves, it's really difficult to tell because of the motion blur going on there (or maybe I'm out of pixel peeping practice!) It's a very good shot though. The time I shot a show jumping arena I realized how tough these shots can be to nail. Have you looked at it at 8x10 printed? That's my usual output size (though I can't say I've ever had my D2x set over ISO 800 with the new firmware it should be pretty comparable, so if I hit the chance, I'll give it a go.)

I'm sure the rider would be more than happy to pay for a copy to hang on the wall. That's what really matters!

Yeah that crop is by no means pretty - I'll have to look for a better example, however it's hard to find a sample at high ISO that doesn't have non-subject motion blur, simply cause the only time I need to use high ISO sensitivities is when I'm doing action shots. 🙁

Show jumping can be tough, depending on the course especially - it's hard to memorize the order and direction of the jumps (which block eachother usually) - also a PITA to put yourself in the best spot to get as many jumps as possible, with the right lighting, going the right direction. That's why I'd rather be out in the fields / woods doing the cross country jumps 😀

 

[compuwar]

JPEG fine, it's event photography so I can't do any post processing on stuff really (people buy the pictures less than an hour after I take them.) In camera NR was medium, that was it. Most of the softness was motion blur (the horse was moving pretty quick) um, some was probably the in camera NR, some was probably the crappy Sigma lens haha.

If you don't mind me asking, how many riders are there usually, and how many shots do you take and prints do you usually sell?

 

[EstorilM]

Well, at 1/800sec and a perfectly frozen jumping horse I do not think that what we see is motion blur from leafs, especially since the doubleness of it is especially visible in the logs themselves.

On closer inspection I see it actually is CA that gives that double image feel to the crop. I do not know if the CA has been partially processed away, but look at the vertical lighter plank on the right, for instance. on the top of the crop its side has a certain thickness, and a slight red tint to the left most edge of it. The moment it crosses the lighter trunk, the thickness of the plank almost seems to half, and the red edge is gone.

Below the metal wheel, the edge of the plank has a very red tint to it.
Now look the right of other dark parts, like the vertical metal bar... again a distinct red CA line. On lighter elements you seem to see green bleeding in to the left, like with the highlights in the leafs in the background. So my educated guess is that it was some quite strong CA after all.

This still leaves the flat feel of things, which seem to suggest some noise reduction algorythm.


The flatness of the leafs, metal parts and all could be called noise artifacts... They are not noise, but artifacts from de-noise-ing... just like sharpening artifacts are not sharpness but for instance strange light/dark halos around contrasting objects 😀

*edit* : Sorry about picking it apart, haha! I totally missed the part where you explained the CA nightmare. I can understand why you are happy with the shot.

Yeah like I said, it's not pretty. BUT I wasn't taking a picture of logs, I was taking a picture of a horse haha. Also, the first thing you said about the horse's motion being stopped so the leaves can't have motion blur doesn't make sense - the horse was moving (fast) the leaves were stationary - I was panning with the horses movement, not so with the leaves.

..but yeah, I think I mentioned it in another post, that my only complaint with this lens is the CA. I haven't had any problems with other lenses, at least ones that could be attributed to the sensor itself.

I could have probably walked out of the tree line and shot a jump in the sun, but everyone had a picture like that already - the woods added some different scenery, which people really liked. I'm just glad I didn't get jumped on.

 

[Lovesong]

Hey all, I just wanted to say thank for all your help, advice, and pictures. I've pretty much made up my mind. EstorilM, don't listen to compuwar or coldrain, that is fantastic action shot. Thank you all again.

 

[compuwar]

Hey all, I just wanted to say thank for all your help, advice, and pictures. I've pretty much made up my mind. EstorilM, don't listen to compuwar or coldrain, that is fantastic action shot. Thank you all again.

Hey! I said it was a "very good shot!" How do I get to be the bad guy? 😱

Plus, coldrain is right, if you're *really* close to your monitor you *can* see the CA in the picture. Which just goes to show that you don't need perfection to get a good shot.

 

[EstorilM]

Hey! I said it was a "very good shot!" How do I get to be the bad guy? 😱

Plus, coldrain is right, if you're *really* close to your monitor you *can* see the CA in the picture. Which just goes to show that you don't need perfection to get a good shot.

Haha thanks.

Ok, I found this.
Blah, now I want to sell this thing for an 18-55 2.8 (NIKKOR!) I'm starting to buy into the 3rd party lens cliche.

 

[Lovesong]

Hey! I said it was a "very good shot!" How do I get to be the bad guy? 😱

Plus, coldrain is right, if you're *really* close to your monitor you *can* see the CA in the picture. Which just goes to show that you don't need perfection to get a good shot.

I know. This exchange actually helped me answer my questions more than anything. You can dissect an image all you want, but what matters is that it's a nice picture at the end of the day.