Low pass vs high pass filter
: Fish pond filter
Low Pass Vs High Pass Filter
- A band-pass filter is a device that passes frequencies within a certain range and rejects (attenuates) frequencies outside that range. An example of an analogue electronic band-pass filter is an RLC circuit (a resistor–inductor–capacitor circuit).
- A high-frequency sound or musical note
- a lofty level or position or degree; "summer temperatures reached an all-time high"
- at a great altitude; "he climbed high on the ladder"
- A high point, level, or figure
- greater than normal in degree or intensity or amount; "a high temperature"; "a high price"; "the high point of his career"; "high risks"; "has high hopes"; "the river is high"; "he has a high opinion of himself"
- A notably happy or successful moment
- A low point, level or figure
- less than normal in degree or intensity or amount; "low prices"; "the reservoir is low"
- A state of depression or low spirits
- in a low position; near the ground; "the branches hung low"
- A particularly bad or difficult moment
- an air mass of lower pressure; often brings precipitation; "a low moved in over night bringing sleet and snow"
Oldtown Alexandria Waterfront V.A.: dawn
[Sony A200/Tamron 18-250 ISO4
00 1/160s F8 180mm effective > dcraw H0W AHD3 > Gimp rotate & crop, contrast 18, USM r1.0 s1.0 th 05]
So "Mr. Confident" didn't take an ISO800 shot in this group. So I took the first ISO4
00 that I could find, not having ExifPro handy to read this particular folder of images and not wanting to take time to move it to where I could run ExifPro on it easily and thus I had to scan through the 20 or so shots by hand to find one at ISO4
00 with good exposure, did that and processed it much as before, with the specific Gimp settings above. Basically I always use radius 1 and with the G9 I usually use strength 0.7 or 0.8, but I settled on 1.0 because that works fine with both the A200 and the G9 at least at low ISO and I can just vary the threshold to keep things in line. In extreme cases lower the USM strength, raise the threshold...use UFraw and use VNG-4
-color demosaicing. If things are really bad use UFraw & lower the gamma and linearity coefficients until the noise fades into the black somewhat. I haven't run NR on any of my images in about a year and I've found USM masks to be a waste of time. They just sharpen the noise in the mask and leave the rest of the image looking dull, two clues that one has been used. It's better to just use a moderate USM on the whole image, that you can live with.
There's the usual "grain" in the sky here that I see with AHD 3-color demosaicing and any serious unsharp-masking, and it's here in a very moderate amount, which could be reduced easily. But other than that the fine-detail looks really, really good. Just barely enough contrast here to clear-up the image. It even looks stable at 100%. Hard for me to find fault with this though I know the sensor is not at its optimum ISO. I would point out that the shot has a reddish-brown cast (that's probably not allowed to be in CMOS shots anymore ;) and the colors are still somewhat muted even after all that contrast. Just too much cloud-cover for a good overall effect. I don't have a G9 shot to compare to this but it would undoubtedly be a little "whiter" and colder, whitebalance-wise. Easy enough to drop the temp by a couple-hundred degrees or so.
So it looks like with this camera the thing to do is to shoot tight exposure-brackets with the high side at 0, keep the low exposure-time at 1/FL, the high ISO at 4
00, trust the metering during the day, drop to around -1eV indoors and at twilight and leave ISO800 to really-dark -2eV shots. Have to rely on the hold, the camera is too noisy above ISO200 to push the shots around a lot and probably too noisy to do much zoom-cropping. But they can still be pushed and cropped *some* with good results.
That will still leave outstanding the issue of underexposed midtones, but that's a general problem with any camera. That's what happens when the sky is really bright.
The key to dealing with that at least in pp is to change the midpoint (which in Gimp means to change the gamma). Any changes of gamma will require a change in contrast to compensate which will effectively undo the change in gamma if the new midpoint is not well-chosen. If the new midpoint is too high then adding contrast will just push the midtones back into the shadows. If too low then adding contrast will push the midtones up into the highlights, giving the shot a falsely-"bright" look. It is crucial to pick a midpoint (a gamma) which puts the midtones square in the middle. That, likewise, is the same as picking the correct overall exposure. Changing the upper level will push the midtones, the shadows and the shadow-noise and blow-out the highlights unless the exposure was so low to begin with that there's no highlight-info to start with. But that's the same as setting the gamma and linearity way too low. And remember that ufraw uses the reciprocal of the gamma and linearity that dcraw (and almost everyone else) uses. So gamma 2.25 in dcraw is gamma 0.4
in ufraw. Lowering the gamma in ufraw brightens the shot, raising it in dcraw brightens the shot. The lowest that the gamma can be set to in ufraw is 0.1 and the same with the linearity. "Linear gamma" is 1.0 in both dcraw and ufraw. "Linear-gamma" is thus very dark in both.
If you need to rescue the midtones, instead of using an "HDR" tool, try just raising the gamma in ufraw and then raising the exposure to overcompensate, then add contrast. You will have to add a *lot* of contrast to make it look natural. Likewise in dcraw reduce the gamma (or use H7W), increase the brightness (or lower the highpoint) and then add contrast.
So this brings me back to the fact that dcraw uses a blackpoint of 128 for the G9 but a blackpoint of 0 for the A200. The only reason to do that would be if the true SNR of the G9 is 8x higher than that of the A200 at the same ISO. But that would mean that one can artificially "enhance" the SNR of any camera just by throwing-awa
Georgetown Waterfront, D.C.
[Sony A200/Tamron 18-250 autowb -0.7eV ISO100 1/1000s F8 39mm effective, handheld > dcraw, camera wb, H0W, AHD 3-color & sRGB gamma & linearity (2.4 & 12.92) @ 0eV > 16-bit uncompressed tiff > Gimp crop, rotate, brightness +60, contrast +56 (still needs more but things were getting blown-out), strong USM r1.0 a1.0 th02]
This shot just plain has more than 8 stops of DR and I've yet to see a software HDR tool that handles this problem well. Why a GND would be better I'm not sure. But I think that you can compare the results and see which works better pretty-easily. In my opinion the first problem with HDR tools is that the blend-in too much of the shadows from the lower exposure, which are inherently more noisy than the corresponding pixels from the higher exposure. It's a simple fact that SNR increases with exposure, it doesn't matter whether you shot at 0eV or +2eV: the shadows are going to be more noisy than the midtones and highlights. Shadows at +2eV are less noisy than shadows at 0eV, true, but that doesn't mean that the two different exposures can be blended-together without the noise in the shadows of one showing-up in the midtones of the result. 2nd is the potential for blur and an unrealistic tone-curve in the result. Panos often have these problems as well. I mean, honestly, this ain't bad for a single shot, though I'm not entirely sure whether I used the Promaster or Tiffen 2A UV filters. The 2A has a slight yellow-orange cast when held up to the light but it's definitely more effective at reducing haze. The scene itself is still going to be hazy, though, and it's not like I have found two shots where I can point at one vs
the other and say "see that's how you can tell that's with the Tiffen 2A vs
A UV filter is, beyond just protecting the lens, supposed to produce shots with much-less "flash" and glare especially in hazy bright conditions. For example the clouds here should have better definition and the water should show less white on the surface. And since it is *supposed* to do that, that is supposed to outweigh a slight color-shift. But the Promaster hardly has any detectable color-shift at all. But hell, they definitely both work ok. So far I haven't seen any shots produced with either one that would make me not want to use it, like light-scattering or internal-reflectance, and they both are much cheaper than the "sophisticated", "high-performance" multi-coated UV filters that I used to buy from say B+W and Hoya. But I had no problem with those filters, either.
I could literally buy all 4 and do an exhaustive test & see if there is any significant difference whatsoever between them. One difference is in the way the boxes are moulded to fit the filters. The Tiffen box has a thin sheet of styrofoam in it and nothing to hold the filter in place except the top plastic covered with a sheet of paper and a Tiffen foldout product-guide which wedges the filter in place. The B&W filters come in a paper box with a rigid plastic box inside that with a pair of foam pads with a hole cut for the filter. Promaster & Hoya are like Tiffen but with the foam inserts that B+W uses, and the Promaster filter is literally quite flimsy, the glass doesn't fit very snugly into the ring. But then the Tiffen 2A has a gap in the coating near the ring that's about 2mm long. Anyway next time I have a free $50 I'll buy one of the Hoya "Digital" MV UV filters that I used to have, its main selling-point was its ultra-low reflectivity when held over a black piece of paper. Now that I can work with the raw sensor-data directly I can easily compare the image-data from all 3 filters shot at the same exposure. The trick is to shoot them on the same scene in the same hazy conditions at the same time LOL
...again you either buy the best that you can afford and hope that it somehow proves to be worth the money and you don't break it stupidly, or you buy an affordable one and hope that it doesn't suck, the shots look ok & it actually protects your lens.
Anyway one last interesting result of all this raw-analysis. It becomes apparent just how important 14-bit D/A vs
12-bit actually is when you think about the gamma correction that has to be applied to get into the proper colorspace. Just pull up a raw file with linear gamma and look at the histogram, and imagine that getting corrected to fill the available DR of whatever format you want to display the image. The linear-gamma form is going to fill roughly the lower quarter of that DR for a well-exposed shot with strong highlights. It's even worse for a low-eV shot that's mostly shadows if not shades of black. Those extra two bits can make a big difference in IQ in high-DR scenes or shots at low exposure. But think about it: even *then* you're only using a fraction of the sensor's total DR. Coming nowhere-near close to saturating the sensor, taking most if not all of the image-
See also: water filter cartridges channellock oil filter pliers sigma uv filter polaroid nd filter nose filters allergy free url filter in sink water filter pur water filter problems