@fatpig Ive been trying to understand what the exact resolution required to pull it out under mjpeg mode without having to resize and be abe to play it straight back off the sd card plugged into a viera tv. Its unusable and yes very sad.
Barrry Green's article was fairy helpful but working out the dimensions pre NLE is taking a lot of testing. I think youre probably right even though its not an anamorphic lens (see lpowell and B Greens findings) there may well be a way of blowing it out properly. Presently its too wide so you mess with the vertical res and then you one image sitting inside the other.... tres difficult... looks like post is the only way to do it for now unless anyone can find a suitable adapter.

@fatpig Try sticking it on a panny viera screen and its vertically squashed (side by side 3D). You can work on it in an NLE/post and sort it but I was attempting to change the sizing incamera so I didnt have to do that (of course this was purely for me and other viera/3D TV owners). We may need a new thread to discuss this further.

I've been experimenting with small resolution increments, just for 16:9, which some seem to work so far without lowering the quality settings Driftwood provided.

The hope is, it down sizes from full size sensor image -> x by y, and stores that, so the idea is to shoot slightly wider than I need for the odd clip that I want to process through Deshaker for some more stability, so that it can crop without (or even less) loss of detail/sharpness for 1080p.

@Frame I can confirm - the MJPEG doesn't suffer of banding problem, as opposed to AVCHD. I've attached two sample screenshots (ISO 320, F0.95, intentional out of focus). I know, these are jpegs, but anyway. First one (Driftwood, GOP1, 174mbits/s) has concentric bands around the picture on the wall, and the second (MJPEG. 100mbits/s) one just doesn't have this problem at all. I think this is all about compression methods of the two very different formats. Previously I was really disappointed that even best available hack doesn't fix the banding issue that I face everytime when I shoot in the darker places with the smooth/gradient background. After I tried the Mjpeg mode I was really surprised in a good way. So here's the workaround for me. If I need to shoot in such environment as in the samples I'd use MJPEG, otherwise - AVCDH would be the choise, most of the time.

@Euggie2000 are you saying in low light situations mjpeg has no banding?

@kenoah in low light banding is either absent or very low compared to AVCHD mode.

@LPowell Thank you for the explanation! Oh, does this MJPEG tunning also improve the photo JPEG quality? Or is there a way to do this? Because many people complaining that the camera JPEG is not too good.

Second this! If you have RAW, why not shoot RAW? Compared to high-rate video, the space needed is still small…

Its quite straightforward, but yes a wiki from the testers will probably appear soon. Take a look at Lpowells low light setting which are good or my higher bitrate/quality mjpeg settings in some of the Quantum patches (dont forget to put it back to 720 width/height =1920 x 1080 and 480 width and height =1280 x 720). Basically you have four tables for Quality (bitrate) and four tables for, well, Quantisation.

My hunch has been that E1/F1 pushes Q chrominance and E4/F4 pushes Q luminance. Lessen the table values for better Quantisation (closer to 0 is better quality low compression but requires more bitrate in Quality setting, higher table values = higher compression, poorer quality but uses less bitrate).

Judge allocating bitrate in the quality settings according to how low your tables go. Too high on the Quality bitrate/low on the tables will hinder recordings/stop them prematurely.

Use jpegsnoop (search google for downloadable proggie) to look at the jpegs produced by the GH2 (when you hit record in mjpeg mode a first frame snapshot is produced - analyse this jpeg with jpegsnoop to see the quality/quant info).

If your recording never gets past the snapshot youre using too low a table setting/too high a Quality setting! Good luck!

@csync Yes, and these same 100Mbps MJPEG settings are included in Flow Motion v1.1

Hey,

i've a question. I use a 1.33x adapter and your anamorphic patch. For me, it seams like the image is much noisier than before.

I read somewhere, that this is "normal" and i have to reduce the noice later with my editing program (i just have premiere pro). Is that right or wrong? Sorry if im asking stupid questions.

Please check my video:

Also, i dont 100% the advantage of using MJPEG instead of AHVCD 24p mode (which is much more cinematic) - is there an easy explanation why i should use MJPEG?

Isnt it better to do the post processing (in example exporting the image in 1080p, 1.33 PAR via Adobe Premiere Pro) than let it be done by the camera (cropping the picture to 1920x810).

Hope this is the right place to ask this kind of questions.

Best regards Emin

Thanks a lot @LPowell - i guess i "just" trust you in that. I basically get both points. I will try to use 800 ISO or less for my next shots and post the result here.

best regards emin

I did some tests and it seems that Gh2 use for .mov files rec601 (SD) not rec709 (HD) color profile. I heve not tested vimeo but I not think they use rec601. The computers and I think the new plasma, led and lcd tv with HD uses rec709. Rec 601 is standard for < 720p , Rec 709 is for > 720p. So this is a bug in GH2 camera. (Is possible in post-production to convert from one to another.) Maybe we will see a hack which has MJPEG 4:4:4 and Rec 709.