[jgredline]

and i'm the know it all? if you say so. you're talking about a reflective meter, not an incident meter. high iso sucks on pentax cameras, you can't get around it. iso3200 looks just as good on practically everything since the *ist dl, it all depends on the person using the camera.

the rest is just hilarious.

Please, I am enjoying following this very good and pleasant conversation. Please turn down the attacks...if you don't like what he is saying, log off the thread.

[attack11]

advice should be corrected by people that know better. this is a resource website, not a tea party.

it's like when someone complains about p-ttl not working when they bounced it off the ceiling... sometimes it's better to speak up.

[attack11]

Are you saying iso 800 (underexposed by two stops), and recovered by pushing it two stops in post processing will get better results than shooting at iso 3200 and not needing any exposure change in post processing?

Is this true for iso 400 as well. If the proper exposure is obtained form iso 400, is it better to shoot at iso 100 and be underexposed by two stops and then recovering in the post process.

Thanks.

that's pretty much how it works. sensor noise is lower if you stay within the acceptable push boundaries of what you're working with; you get sharper images.

for example, it's the ideal way to shoot in a hockey arena.

[wolfier]

Okay, I understand it the purpose of the test now - as long as there is adequate light, ISO 3200 is not a problem. Thankfully I already believe this

Your test would be the same as when you set to ISO 3200 to freeze a motion in bright daylight (let's say, you want to use f/5.6 instead of f/2.8 for sharpness and DOF reasons).

In both this case and yours (where the room is dark so you increase the Tv) the result is the same - enough light => noise not an issue.

Cheers!

True. In dpreview I had mentioned how the 0.3 sec, makes it not applicable to real life situations. I see that I had skipped it here.

However, I think noise at 3200 is noise at 3200, what ever the shutter and aperture may be (with-in limits...not real long exposure). So, it is just a test to see what the noise is at iso3200.

Think of it this way. I wanted to make the room as dark as possible and do the test. If I turned on the table lamp, I could have gotten the shot at iso 3200, f4 and a healthy shutter speed like you would like (will be more applicable to real life situation). Now with that set-up, one might say, the room is not dark enough. Do you see where I am going.

So, given iso 3200 is iso 3200, and the aim of this test is to make the room as dark as possible and use a tripod to see what iso 3200 looks when ETTR, this is the ideal way to go about achieving those aims.

As iso 3200 is iso 3200, one would not gain any more knowledge by testing in a brighter room only to achieve a healthy shutter. It only introduces the notion that iso 3200 was used unnecessarily. Do you agree?

To answer your question. I use high iso exactly like you do, to get a good shutter speed. It is mostly in good light with thew 300mm attached to a 1.4 TC. Or available light indoor shoot where a much higher shutter tahn 1/3 will be required (could never do this with the K10D to my satisfaction, the K-7 handles it like a walk in the park below f2.8). So, an identical situation to the test will never occur in real life. Also, I will always choose noisier shot than a blurry shot exactly like what you say you would do.

P.S. Thank you for the civility in your question.

[pcarfan]

Okay, I understand it the purpose of the test now - as long as there is adequate light, ISO 3200 is not a problem. Thankfully I already believe this

Your test would be the same as when you set to ISO 3200 to freeze a motion in bright daylight (let's say, you want to use f/5.6 instead of f/2.8 for sharpness and DOF reasons).

In both this case and yours (where the room is dark so you increase the Tv) the result is the same - enough light => noise not an issue.

Cheers!

No, my test is done in bad light. Iso 3200, F4, 1/3 for proper exposure is very poor light. You can test it by setting your camera in manual mode and set F4, 1/3 and iso 3200 and find the light level that will not blow the highlights. You will find that it if fairly dim.

P.S: I think I was wrong in saying iso 3200 is iso 3200. In bright light iso 3200 will show less noise and in dim light like my situation here, iso 3200 will show more noise.

[Ash]

Okay, I understand it the purpose of the test now - as long as there is adequate light, ISO 3200 is not a problem. Thankfully I already believe this

Your test would be the same as when you set to ISO 3200 to freeze a motion in bright daylight (let's say, you want to use f/5.6 instead of f/2.8 for sharpness and DOF reasons).

In both this case and yours (where the room is dark so you increase the Tv) the result is the same - enough light => noise not an issue.

Cheers!

But you wouldn't really need to use ISO 3200 in daylight anyway. You'll find you won't have a shutter speed fast enough for such high sensitivity in those conditions. And when would you need a shutter speed faster than 1/4000?

Low-light is where it's mostly applied, and again mainly for being able to handhold a shot, otherwise you'd just set the camera on a tripod for stationary objects and allow for a longer exposure (and subsequently lower sensitivity).

[Alfisti]

that's pretty much how it works. sensor noise is lower if you stay within the acceptable push boundaries of what you're working with; you get sharper images.

for example, it's the ideal way to shoot in a hockey arena.

This contradicts anything I have seen either on the web or first hand. As soon as you start pushing in post things get noisy REAL quickly.

[attack11]

This contradicts anything I have seen either on the web or first hand. As soon as you start pushing in post things get noisy REAL quickly.

it's only for <= iso800 'cause the sensor output goes to crap as soon as you get into the next stepping. it's a pretty standard film technique that transitions to digital very well if you get experienced with it. i routinely shoot iso1600 on a k10d without issue and usually no NR in PP (no NR in body either, i don't want in camera curves removing detail).

[Alfisti]

it's only for <= iso800 'cause the sensor output goes to crap as soon as you get into the next stepping. it's a pretty standard film technique that transitions to digital very well if you get experienced with it. i routinely shoot iso1600 on a k10d without issue and usually no NR in PP (no NR in body either, i don't want in camera curves removing detail).

Say what now? Are you shooting in 1600 or pushing to 1600 in post? Any time I push hard in post at any ISO I get a freakin mess of epic proportions.

[attack11]

what converter are you using? i rely on acr 'cause it's (imo) in the top 2 these days.

if i wanna get iso3200 i'll usually push iso800 2 stops and lately i just shoot iso1600 with spot metering and ev compensation.

edit: of course i get to deal with that awesome blue grid in mid shadows at iso1600. gotta love that

[CFWhitman]

I'm kind of at a loss here as to what the argument is about.

First it seemed like people were accusing pcarfan of pulling his ISO 3200 shots in order to reduce noise (i.e., overexposing them in camera and correcting exposure in post). To me, pulling two stops over correct exposure seems like a recipe for clipping highlights, but, either way, he confirmed that he was not reducing exposure by two stops in post, so he was not pulling the shots.

Now it seems that people are arguing against the validity of this test by saying that you could do just as well by pushing ISO 800 shots by two stops (i.e., underexposing them in camera and correcting exposure in post). This is an age old argument when it comes to digital photography, but what it has to do with the validity of his test results I don't know.

Yes, people have demonstrated in the past that it can be possible to get acceptable or even comparable results to a higher ISO setting by pushing a lower ISO setting. This is basically amplifying the signal with post processing software instead of in camera software -- something that is usually pointless except as an illustration of what is going on inside the camera unless you are trying to shoot above your camera's maximum ISO setting. That is, pointless because the in camera process generally works just as well or better than pushing the exposure in post.

This point that in camera high ISO settings are really pushes of their regular settings, though, can be made about every digital camera that ever had an adjustable ISO. It's pointless as an argument against the validity of this test because it applies equally to every other digital camera out there. I don't see why it should be brought up in this thread.

Then of course there is the point that enough light makes ISO 3200 shots acceptable. This is another thing that is true to some extent or another with every digital camera capable of doing ISO 3200. However, unless the images are very bright or he is pulling the shot (which he confirmed he had not done), then this is also pointless. Yes, a well exposed shot helps; it always does. It is true that the first shots had no illustrations of how a large dark area would turn out at ISO 3200, but the later shot helped. It might be nice to also see some shots where whole sections of the background were dark colored to see what kind of results you would get, but the images shown are still valid tests.

[pcarfan]

Charlie,

Thanks....I enjoy a good argument, but not a diatribe (easy to ignore who engage in this). However, almost everyone is engaging in a proper discussion, and I think it is a good learning experince for all....certainly is for me.

[rawr]

It seems one lesson of this thread may be that if this style of shooting is important to you, just get a camera with a better sensor - that can hold onto a decent dynamic range higher up the ISO scale - so you can confidently shoot 'natively' at higher ISO's, without needing to fall back to extreme EV adjustments or PP pushing to deliver an acceptable image.

[pcarfan]

it seems one lesson of this thread may be that if this style of shooting is important to you, just get a camera with a better sensor - that can hold onto a decent dynamic range higher up the iso scale - so you can confidently shoot 'natively' at higher iso's, without needing to fall back to extreme ev adjustments or pp pushing to deliver an acceptable image.

roflol....

[robgo2]

Yes, people have demonstrated in the past that it can be possible to get acceptable or even comparable results to a higher ISO setting by pushing a lower ISO setting. This is basically amplifying the signal with post processing software instead of in camera software -- something that is usually pointless except as an illustration of what is going on inside the camera unless you are trying to shoot above your camera's maximum ISO setting. That is, pointless because the in camera process generally works just as well or better than pushing the exposure in post.

This point that in camera high ISO settings are really pushes of their regular settings, though, can be made about every digital camera that ever had an adjustable ISO. It's pointless as an argument against the validity of this test because it applies equally to every other digital camera out there. I don't see why it should be brought up in this thread.

Is this correct? I thought that in the camera there is electronic boosting of the signal, which contributes to high ISO noise. It is not just a software enhancement.

In the case of the K-7, ISO 1600 is the actual maximum ISO, and ISO 3200 is merely an extension of it.

Rob