[richtrav]

Hi all,

This is my first post here but have been lurking for months. Two weeks ago I bought a K-7 and have put it through its paces pretty well. Overall I LOVE the camera, easily the best I've ever owned, but the banding present in videos or extended exposures is frustrating. I live in South Texas so high temperatures are going to occur - I get the red thermometer usually after a minute or two of shooting with Live View on in the afternoon. For what it's worth I apparently have an earlier model (ser # 3377xxx).

I've seen some posts imply that newer cameras are less affected by the banding. Can anyone confirm this? If so I'd like to give a newer body a try; the K-7 is nearly all I could hope for in a DSLR in 2009.

Video in particular is nice, especially compared to dad's Canon T1i, it has more of a filmlike appearance instead of a videocam look to it without the need to do some major tinkering. I also love the ability to "film" at full res and then being able to either crop it or leave it letterboxed for a 1080 image. But my first shots are always better in quality than ones after a few minutes outside. Any advice is appreciated (other than not to shoot outdoors, that's not a practical option). Thanks

Richard

[wolfier]

My K-7 was preordered, and when I used it to continuously take pictures (about 80 exposures of 15-30 seconds each) of lightning, without resting, without DFS, there was no banding...

[gfmucci]

It would really be nice for Pentax to issue an "official" statement about why some (many) cameras exhibited this problem, and what they did, if anything, to eliminate the problem.

As it stands, it appears to be a random problem without a definitively known cause. Their silence is deafening.

[LeDave]

It would really be nice for Pentax to issue an "official" statement about why some (many) cameras exhibited this problem, and what they did, if anything, to eliminate the problem.

As it stands, it appears to be a random problem without a definitively known cause. Their silence is deafening.

There will be an answer, let it be. For though they may be parted there is still a chance that they will see, there will be an answer. Let it be.

[wolfier]

By the way, please post an example of the "banding".

My take is, there's always a threshold on how long LV or videos can be taken before something overheats. It's a physical limitation and cannot be overcome, regardless of brand.

If there's a fix, probably it'll disallow you to continue shooting video when a lower threshold is reached, and I'll take banding over a stopped video.

There's a threshold set on every DSLR that can do LV or video - if the sensor goes beyond that temperature any video or picture taking will be forced to stop, giving the sensor a rest. This is what I've read.

The video time before it happens probably is a sample variation, and Pentax probably picked some median threshold. I guess in your case the threshold is longer than when banding occurs. Just return the camera and get a new sample..

[nostatic]

It would really be nice for Pentax to issue an "official" statement about why some (many) cameras exhibited this problem, and what they did, if anything, to eliminate the problem.

As it stands, it appears to be a random problem without a definitively known cause. Their silence is deafening.

It would also be nice to know the actual number of cameras that exhibited the problem instead of inflated assumptions based on the internet merry-go-round of complaints. You get one or two real issues posted, then they get reposted, then new threads get started by people not directly affected, and suddenly you have a "widespread" problem.

I'm not saying that there isn't a technical issue, but it never ceases to amaze me how much 2nd/3rd/4th hand information gets passed around and suddenly there is a corporate conspiracy of silence.

It doesn't excuse Pentax, but hopefully the people who actually had a reproducible problem reported them directly to Pentax so their engineers could work on it. If they can't reproduce the issue, they can't really issue a statement. Or maybe they have and they are just keeping quiet until they can figure out a fix. Wouldn't be the first time a tech company did that...or the last.

[richtrav]

OK here is an example, this is a still clip taken from a video. The camera had been shooting outside for about 8 minutes on a Texas afternoon:

http://i73.photobucket.com/albums/i202/richtrav/videobanding.jpg

To be fair that is about as bad as it gets, but it looks worse in motion than on a still.

Here is another, less egregious example from a timelapse taken at night, where there is a distinct band down the middle (this was taken at 5500' elevation in Mexico, so our nice toasty Texas sun was not the culprit)

http://i73.photobucket.com/albums/i202/richtrav/IMGP2053.jpg

Really if these occurred after long sessions of shooting video I wouldn't complain, but it crops up pretty quickly in the Texas heat (did the same in Manzanillo last week too). Aside from this issue I love the camera and its pictures and videos. I know some have blasted the 1536 mode but I love the flexibility it offers (being able to crop the top and bottom wherever you want, pillarboxing it for a 1080p screen or boxing/downsampling for a supersharp 1080 x 720 video or somewhere in between, the ability to control the sharpening when it gets downsized to 1280, etc etc). That is of course if you don't mind post-processing; if you want easy good looking video straight out of the camera set it to 1280 with no adjusting to the presets and let 'er rip. They're noticeably better looking than the default videos from Canon's T1i or Nikon's d5000/90

Sure I wish it produced the quality of a 2mp downsized still (who doesn't?) but no camera in 2009 that costs less than 5 digits can, including the 5D or GH1. I poured over many sample videos from all those cameras and to my eyes the K-7 can keep up with either of them in decent light. The stills extracted from the 1280 mode are usually good enough to be used straight out of the camera, and 1536 stills are even better if downsampled to about 1200 x 800.

The explanation falconeye gave for the video implementation makes perfect sense - I used to have an old Kodak DSLR with no AA filter and it produced similar images: sharp but with some moire patterns present. By sampling every 3rd or 6th pixel the K-7's AA filter would not be able to do its job of reducing moire. The "Christmas light" speckles are easy to deal with and I think I remember even seeing a program that can deal with it automatically; on the other hand, the wavy luminance moire that you rarely see in windows or on clothing is very hard to treat (though to be fair I've seen moire on other video cameras too, it even crops up on TV). I'm still tinkering with the settings for video: what looks good on a computer screen and what looks good on a typically preset HDTV are often pretty different. Turning the sharpening down helps the artifacts in 1536 mode, and making use of the easy to access contrast and hi-low key buttons can really squeeze more dynamic range out of a scene (or conversely be used to "crush" the darker regions to get less grainy looking videos indoors).

Speaking of all this, does anyone know how the K-7 is outputting videos from the HDMI port in 1080i mode? Before the video starts it appears just like it were in 720p mode but once it starts playing the image contracts slightly and looks a bit crisper. Is the K-7 keeping the width at 1536 pixels and stretching the height slightly to a full 1080 pixels? Whatever it is doing looks good, though panning does cause some stuttering. I'm guessing the stuttering is a limitation of the K-7's playback circuitry: it's not in the actual recorded video, nor present when being played straight off an SD card onto a computer screen.

Bottom line, yeah I think I'll give another K-7 a try to see if it performs better. I don't have to have perfection, just something I can use for more than a few minutes outside in summer