Download World Map High Resolution

New to Panasonic's G9 flagship is a high-resolution mode, which shifts the sensor by half-pixel increments eight times, and generates an 80MP final image. As with similar technologies from Ricoh and Olympus, it's not necessarily recommended for scenes with moving subjects in them. But we wanted to see if we could make it work.

This looks to our eyes to exhibit some improvement. We overall found that a shutter speed between 1/4 sec and 1/8 sec gave a reasonably natural look to the average pedestrian in motion - of course, for faster and slower moving objects, you'll have to adjust accordingly. Do take note, though, that there are some interesting colorful streaks in our moving subjects, and a reduction of resolution in static objects that can be seen behind them.

DOWNLOAD 🔥 https://urlca.com/2y4AKd 🔥

In any case, the high res mode on the G9 is something we want to continue to look into as we progress with our review. Raw support is coming shortly, and we're looking forward to examining the Raw files from both real-world shooting as well as our test scene.

The object of a really high resolution like 80MP would be to create really large prints using M43. Most folks would use this feature for landscapes etc. instead of trying to eradicate escalator people.

still however, i'd rather use a large sensor with a true high resolution of 80MP to deliver better results in one single click of the shutter button ... not even PhaseOne offers that though, unless via similar interpolating tricks, which has been available via software since the 1980s!

BTW - the Hubble pixel shift is a red herring. Hubble do not have any Bayer detectors. Hubble is doing pixel shift for totally other reasons. If I am not misremembering it was to fix dead pixels. Or maybe it was to increase resolution. Also. Anyhow - it was not to fix Bayer problems.

theoretically, AND practically, something similar is possible with film / print material as well, done many times in the past on a regular basis: a photo shot on 135 format film (or even smaller) can be printed at its optimum best-results size (say 8"x10" or a little bit bigger or smaller) and then re-copied using a large format film (using special copy film materials, filters, and processes) and then enlarged billboard size with amazingly high quality! (more on this on the comment below under star *)

This "trick" has only to do with nullifying the mosaic pattern of the Bayer CFA filter. Nothing else. Only some very few cameras do that. Hasselblad, Olympus, Pentax and now Panasonic. Only high(er) end cameras. All of them by mainly moving the sensor one pixel.

What you are talking about can be made with a program called PhotoAcute. There you take a number of images, moving the camera slightly between photos. Then you combine the photos, increasing resolution and removing noise.

And if some would remember, Olympus and KODAK both created the 4/3" format, sensor technologies and future technologies to overcome the resolution and color limitations with current bayer filtered sensors. So holding from that time the HR mode, waiting IBIS technology to improve so much.

the fact of the matter is, with interpolation or layer-rotation, the resolution doesn't change really and remains the same amount of pixels and only more color hues are gained, while with sensor shift, there are also some software processing tricks applied, which together do increase the resolution (thus file size) of the final image just as well ... hence a 30MP sensor yields a 100MP file and a 100MP sensor, a 400MP one in the end ... (such as in the latest Hasselblad model ...)

Sigma DP3 Merrill has the Foveon sensor which produces all the resolution required in a single exposure without having to "shift" the pixels. 16 bit file is 88.6 megs after processing and converting to tiff.

No. You cannot have very high noise levels with a good dymaic range and it shows. The dynamic range of these cams is poor. And personally I never liked the colours either. I wanted to have one due to all the raving reviews and started to look at tons of landscap pictures and they just looked oof with blocked shadows and strange skies that looked more purple than natural blue. But to each their own.

So a quick calculation indicates that at 80MP, diffraction limiting is f2.8. So realistically, best suited to lenses like the longer Voigtlander f0.95 trio (not the 10.5mm), the Nocticron, the O75mm/1.8, and perhaps the Sigma 30mm/1.4 stopped down to f2.8 where they achieve extremely high MTF values.

However, since the high-res mode is most appropriately used at 40MP, not 80MP, that brings your diffraction limit up to f4. Which opens up a lot more options, and means that something like the O12-100/f4 at the wide end might be good enough, as well.

Irrelevant.

Diffraction limiting means just that - it becomes a factor on a sliding scale. A 'diffractionlimited' f5.6 image will still have better resolution than a non-diffraction limited f5.6 20MP image.

FWIW, while Panasonic doesn't actually say it, based on the sub-pixel movement, number of photos, and end-result resolution, they are using "super-resolution" techniques. Because of this, diffraction limits as discussed above do not really apply... or, I should say, only apply to the 20MP sensor's resolving ability.

A good test is to go outside, set up the camera on your tripod with a long lens such as 150mm, focus on a distant object with live view 14X magnification, and tap the camera to see how long it takes for it to settle. When I did this using a modern high-end tripod and heavy duty pan-tilt head, I was surprised at how long it took.

The sad thing is not that it cannot do what it should not be able to do. The sad thing is that it cannot do what it should be able to do. It has some faulty algorithm for using subsampling for increasing resolution. It can be seen along all dark thin lines. They have a constant width area that is slightly brighter. Not like usual sharpening halos, but like something that looks extremely artificial.

Since the new versions of DFD on both GH5 and G9 (480fps vs the previous 240fps) focusing has improved, the question arises: will our "old" V1 lenses work fully on this new cameras or we have to throw them in the bin- or just use them on the old cameras...My guess is that the new V2 lenses received a faster controller to cope with the high speed and precise positioning of the af motors and the optics remained the same - but since Panasonic has released this document promising firmware update- it may be the case only of a new firmware update would be needed - which did not arrived yet.. But will that ever arrive? Or will we be forced to buy new lenses to benefit from the precise and fast AF and other improvement? Or let's ask differently: what will we loose using the old lenses with the G9 ?

@Roland -- I presume you are citing FF lens equivalents. On mFT that would be using a 42.5mm lens? Well that's far from the wide AOV that I need for most shots and I still don't see any reason to bother with it, when High Res does a great job (with some PP work). I will give it a try, although I think it will only work properly for WA architecture using a shift lens. Here's a relevant article - -with-tilt-shift-lenses-to-create-high-resolution-images/

BTW today I tried HDR plus High Res: I shot bracketed High Res exposures with camera on a solid tripod and tethered to a computer running Olympus Capture to control the camera without touching it. After processing the Raw files in Olympus Viewer 3, then blending in HDR EFX Pro2, the resulting image is even sharper and more detailed than the individual HR shots! There are halo artifacts around some high contrast edges, however, which may require different exposure and/or processing settings to alleviate.

@Roland -- Please read the thread again. OP wrote: "This seems like a solution in search of a problem. Why not just stitch images, if you want higher resolution? Less hassle, no need for tripod." That is what I replied to. Subsequent posts brought up the issue of noise, but the discussion was not exclusively about noise.

@Roland -- I am aware of how they "fix" the problem, but not without errors, smearing of detail, distortion, and other problems -- no comparison to a single high resolution shot taken with a highly corrected UWA lens. That is why these expensive lenses are prized by photographers.

I visited the Kolor Autopano Giga gallery, which is quite striking but doesn't offer high res samples, only downsized screen images. It features very wide panoramas with aspect ratios that are not representative of typical rectangular images that my clients want. It also doesn't indicate how much post processing work went into them. The closest one is Bruxelles, by Thierry Guinet," but it's not very sharp. I have yet to see a stitched UWA image other than shift-stitched that can deliver clean, convergence-corrected, undistorted high res standard rectangular images of an architectural subject -- in the way I know this can be done efficiently with High Res capture. I'm certainly open to the potential (I will do some tests) and to discovering samples that anyone might want to link here.

The technology has been around since 2001. First from Jenoptik's Eyelike digital backs, then later from Hasselblad (The H5D-50c & 200c MS allow still- life studio photography at moir free 50 or 200 Megapixel resolution. Four & six-shot technology ...)

You have to remember that Foveon captures 100% detail (RGB in every pixel) in a single, native res image. Bayer high res is still interpolating multiple images, each containing only 50% detail (RGGB square). Foveon is known to have detail well beyond its MP rating because of this.

You don't understand how the pixel shift tech works. Oly's and Panny's pixel shift accounts for the bayer pattern by taking 8 shots and shifting the image one pixel to get all the information AND to increase resolution. e24fc04721