ARGHHH!

[shots worth sharing]

<p>This (minimal pp--straightened, cropped, converted and brightened slightly) sucks. I'm not so concerned that the focus is slightly off--it's the extravagant noise. So much for my new year's resolution about controlling my swearing ;~) <br>

What went wrong? ISO = 2500, 1/250s, F/3.5, 1 EV. I'm thinking it may be the center weighted average metering--is that right? <br>

(Oh, and she missed her shot, too.)</p>

<p> </p><div></div>

[revolver]

<p>Dave, I often swear after seeing my "people" shots....Godforbid if they move evenso slightly.....I feel your pain but have no advice.</p>

<p>:)</p>

[shaloot]

<p>Have you tried running it through some sort of noise reduction program? This gym looks dimly lit, so maybe this would have called for some funky flash setup but we all know those are rarely allowed...<br>

I feel your pain though. I borrowed Jeff Crister's 50-135 to take with me to an indoor soccer game and I got all crappy shots. I was wide open, but only managed 1/150 and they were mainly blurrly... iso was cranked all the way to 1600 and the noise was terrible. It was lit by florescent lights, so really didn't have ideal lights.</p>

[pj_vesterback]

<p>My $0.02... the shot is very underexposed. Noise seems to be much more apparent with underexposure. I think the underexposure may be due to center weighted metering. You would do better by spot metering off the subject. But then again if your lens isn't fast enough, there's probably not much you can do.</p>

[r.t. dowling]

<p>I've seen some dimly lit high school gyms, but none as dim as that. Maybe you should hold some bake sales to raise funds to upgrade the lighting in the gym! I'm only half kidding. Today's lighting systems are brighter AND use less energy, saving the school money. The high school I went to recently upgraded the lighting in the gym and they now save thousands of dollars a year compared to the old system, and the new system is brighter and much more photography friendly (less color cast).</p>

<p>Also, 3.5 really isn't much slower than 2.8, so I'm not sure that a 2.8 lens would have made much difference (except to your wallet).</p>

[shots worth sharing]

<p>Sommana: Oh sure, I used Neat Image to clean up the finished product (this ain't it) and it helped but when you're working with something this bad there's only so much it can do.<br>

Thanks for the $.02, Patrick--the metering mode is the only way I can explain such drastic underexposure with those settings (but maybe she missed the shot because it was too dark to see the rim.)</p>

[shots worth sharing]

<p>R.T. And like I say: not just dim light, <em>strange</em> dim light. When she gets to upper school, she'll be playing in the new facility where the light is much better. <br>

It was a 2.8 lens, btw, but I stopped down for reasons of dof.</p>

[trw]

<p>The floor and walls are lighter than the people. You may need to use positive exposure compensation to get a good exposure of the people. Also, it looks like a mix of either old fluorescent with "warm" fluorescent or fluorescent with sodium vapour. Mixed lighting sources is never good, but beyond your control in this case.</p>

<p>If you can get a couple of manual hotshoe flashes (ie Vivitar 285HV), and some way to trigger them remotely (Cactus V2s if you're on a tight budget, cyberssyncs or pocketwizards if you need reliability), check out these three articles from David Hobby:</p>

<p> </p>

<ol>

<li><a href="http://strobist.blogspot.com/2007/03/q-speedlighting-gym.html">Speedlighting a Gym</a></li>

<li><a href="http://strobist.blogspot.com/2006/04/on-assignment-prep-basketball.html">Prep Basketball</a></li>

<li><a href="http://strobist.blogspot.com/2007/02/on-assignment-speedlighting-college-gym.html"> Speedlighting a College Gym</a></li>

</ol>

Also, follow the links to "Lighting 101" on the same site.

 

[trw]

<p>Also, if those are old (flickery) style transformer-ballasted fluorescents, they could be bright when the camera meters, then dim when the shutter opens... In north america, this usually limits you to a max shutter speed of 1/60 (one full AC cycle). Short of convincing the school district to upgrade to electronic ballasts, your only option may be using flash.</p>

[vegard_nervik]

<p>I think Patrick is right about the exposure, but your options were limited since you can't open up more, the shutter speed can't go down much and hardly no more ISO to be had.</p>

<p>The white-balance seems a bit off also in my opinion. </p>

<p>You could try to convert to Black and White. I usually don't like noise that much in colour, but I think noise in B&W is very nice sometimes. </p>

<p>I hope you don't mind, but I had a quick try. Adjusted the levels slightly and boosted the contrast a bit and played slightly with the shadows and the highlights. I am not very good at it, so the results should be much better if you try different options and spend a bit time on it (and get someone knowing PP to help, in other words not me).</p>

<p>With B&W noise works with you and not against you, but that is just my personal preference. Cropped it a bit tighter, but that was mostly because parts of the wall got burned out. That could have been fixed if I were more careful with my adjustments.</p>

<div></div>

[ronald_moravec1]

<p>Mixed lighting is terrible. They probably have a combination of fluorescents of mercury or sodium vapor none of which will work photographically. Then there is one side of the pic with one color balance, and the other side with another, Really fancy photoshop to correct.</p>

[michael_kuhne]

<p>Auto WB is notoriously bad with Pentax DSLRs. There is also considerable underexposure here, which should not have occurred with center-weighted metering. There is no real case of backlit subjects here. The central portion background has darker content as well as some lighter. I also much more frequently get under exposure with my newer K200D than with my K100D. Seems like the engineering efforts went towards protecting highlights.</p>

<p>For ISO 2500, you may need to run it through a denoising program. Make a copy to work on in a separate folder. If you have PS or PSE, go to enhance, then hit the auto contrast and see what happens. (as an alternative, increase contrast separately to your liking). then bring up fill flash, and give it a little of that. </p>

<p>I processed some hockey shots taken with my K200D recently in this manner with decent results. But I shot with ISO 1600, and did not have a pressing need to denoise.</p>

[michael_kuhne]

<p>I meant to say auto WB is notoriously bad for flourescent lighting with Pentax. Under most other conditions I find it to be ok.</p>

[shots worth sharing]

<p>Well, I did some shooting at my grandson's game with spot metering this afternoon and got similarly bad results. This really makes me appreciate the delicacy of the balance between various settings. I think I'll go back to manual pre-focusing the 43mm ltd and shooting wide-open: I'm certainly willing to sacrifice some sharpness for the sake of noise reduction.<br>

Oh, and here's the best I could do in terms of pp (but I think Vegard's B&W suggestion is a good one.)</p>

[shots worth sharing]

<p>here's the processed version.</p><div></div>

[mountainvisions]

<p>If you're shooting with the K20D, its not the camera. Likewise some recent ones with the K10D were equally clean. Of course I've been saying the K20D isn't tremendously better than the K10D in natural light at high ISO. But in good light, where you can shoot at 1/400-500th at f/2.8 to f/4 and get proper exposures both these cameras do just fine.</p>

<p>At 2500 my images shot last week were clear as day. I mean there is noise as expected, but not much. Where the K20D beats the K10D is in the shadows, much less noise, no annoying purple banding. So little in fact I hadn't used any software reduction other than +10 in lightroom.</p>

<p>The metering mode is never the problem. It simply sets or tells you the difference in exposure from standard exposure. It doesn't have any input on the image noise.</p>

<p>For instance, if I went outside with a camera with a broken meter, and used the sunny 16 rule my exposure should be close, and my image should be accurate. If I shot the same scene metering in spot, at the same exposure, I'd get the exact same image.</p>

<p>Likewise, using spot, center, or matrix doesn't change the image, only the way it is metered.</p>

<p>The reason your image is so noisy is you WAAAAAYYYYYYYY under exposed it. I'd probably 1-2 stops underexposed. Now even if your brighten it in RAW it's still going to be so noisy you will want to puke.</p>

<p>The reason, and this goes back to my K10D discovery which since then I've applied to all digital cameras I've used, is that digital sensors in general, have very little under exposure latitude at high ISO.</p>

<p>So if you are going to misjudge exposure, or have to choose between highlights and shadows at high ISO, I'd choose to burn out some highlights (you know like those shiny white jerseys they ALL seem to wear).</p>

<p>What I tend to do is spot meter off the white jerseys and open up a 1-1.5 (or rather 1.66) stops. You can check this with the blinkies, if 1.3 or 1.6 is too much. but usually between 1-1.3 is about right.</p>

<p>FWIW, you did the best you could with the lighting you had. As I tell my editor, give me good light, I'll give you good photos. It's really that simple, I don't care if you get the best sports photog in the world into one of these dank, dark middle school gyms, he/she won't get you anything you will be blown away by with the crappy natural lighting.</p>

<p> </p>

[shots worth sharing]

<p>Thanks, Justin--the sort of continuing feed back I get here is extremely helpful. I think I need these exercises in frustration to focus my attention on technique.</p>

[htarragon]

<p>Dave,<br>

This is the reason spots photogs have those humongous lenses. They need to be fast. I put my MF 50mm f2 on my K10 and shot at f2 and f4</p>

<p> </p><div></div>

[htarragon]

<p>And this is f2:<br>

I also did an Infrared conversion in Elements and that gave the best brightness/contrast without the weird colors. If you look at the histogram of the picture in the component colors you can see how deficient they are on the right side. Wait until she plays at a better Gym! When she's in High School and taller she'll make the shot!</p><div></div>

[htarragon]

<p>OOPS. ISO 800 125/sec. Well lit room</p>

[graham_ferguson]

<p>Hope you don't mind but I couldn't resist running your shot through noise-ninja: (I also removed the green-cast thrown by the bad lighting and green wall in PS):</p>

<p><img src="http://img213.imageshack.us/img213/5843/basketballfiltered1zk1.jpg" alt="" /><br>

To my eye, the remaining grain just looks like high-speed film stock. Please pardon my impudence at jumping into a thread like this, but I thought Noise Ninja could assist.</p>

 

[shots worth sharing]

<p>Thanks, Howard--what I'm getting out of this challene is a real world sense of what those camera setting numbers mean. And I'm taking to heart Justin's admonition (from another thread) that most photos look awful at high magnification.</p>

<p>Nice job, Graham--I don't regard it as impudent at all! I've just started using Neat Image but I'm still at the "what does this thingy do?" stage (and, for some reason, it's behaving a bit strangely with this shot but I was able to figure a work-around.)</p>

<p>I guess my major "take-away" from this thread is a finer calibration of reasonable expectations for this sort of shooting ;~)</p>

 

[htarragon]

<p>Dave,<br>

I have learned something and it has made me think.<br>

1) When I first saw the shot and saw the ISO number I thought something was wrong. ISO 400 hundred film will give you enough light to shoot indoors in a fairly dark room. So I tried some shots in my room which has 6 85 watt recessed ceiling floodlights and 2 300 watt halogen floor lamps. (I only use them all when I decide I need to sauna). ISO 400 was insufficient. You saw the results with ISO 800. So, film and sensors are different in their ability capture light. I guess I still had the assumption that they were similar.<br>

2) Seeing the individual histograms also surprised me i.e. the deficits in specific colors. That also contributes to less "light".<br>

3) The amount of post-processing needed to bring the image to what our brains would have pre-processed in a nanosecond to make the image normal.<br>

4) I do street shooting and even in shaded daylight my shots are often soft because haven't set the shutter speed high enough although it does give the sense of a fleeting moment. Or could it be the autofocus? Hmmm.<br>

Thanks for this thought provoking discussion.<br>

Howard</p>

[shots worth sharing]

<p>Very interesting observations, Howard--thanks!</p>

[mountainvisions]

<p>The problem that I've noted with these ultra fast lenses, is good luck getting a sharp shot.<br>

<br /> Even if your camera keeps up wide open, the problem is that the DOF is so small that even the slightest camera or subject movement means focus plane is altered.</p>

<p>So white those pro sports photogs DO have big lenses with fast apertures, they OFTEN don't shoot them wide open unless absolutely necessary. In your dark gym, they most likely would.</p>

<p>It's really a question of how much of the subject you want in focus. I suppose if I was shooting a long jumper the feet would be fine out of focus, as long as I nailed the face. But with team sports you might want a bit more DOF to capture the opponents and such.</p>

<p>On a daylight field sport, I bet most pros are shooting at around f/4-5.6 or even higher, a lot of pros shoot with shutter priority to shutter speed they prefer. For me in baseball, for daylight, it's 1/1250th (or faster). So I generally will adjust the ISO to keep my aperture somewhat wide, but ithin a range of f/4.5-7.1 in daylight shooting.</p>

<p>Isolation is great, but so is a sharp shot that doesn't make the subject look soft from too shallow DOF.</p>