Panasonic GF1 hacked firmware 1080p 4:2:2 sample? Is this 4REAL?

[GRENCH]

Panasonic GF1 hacked firmware 1080p 4:2:2 sample

http://www.youtube.com/watch?v=wr8kKWZqoZw&feature=watch_response

Also I just saw this 1 on youtube as well. It claim to output 80 mbps on a gh1. I'm just getting familiar with these hacks so please don't knock my questions. Just looking for some useful info so that I can make some decisions.

http://www.youtube.com/watch?v=aAVdStF8hNQ&NR=1


Is this 4 Real? Can't tell by youtube footage. Has anyone actually used these firmware upgrades. I have a GF1 and was thinking about getting a firmware hack for it and came across these links. I have done some research and have read a few post on the GH1 but nothing on the GF1. If know of 1 please let me know, but for now any insight would be much appreciated.

Thanks

 

[dex4]

Panasonic GF1 hacked firmware 1080p 4:2:2 sample

http://www.youtube.com/watch?v=wr8kKWZqoZw&feature=watch_response

Also I just saw this 1 on youtube as well. It claim to output 80 mbps on a gh1. I'm just getting familiar with these hacks so please don't knock my questions. Just looking for some useful info so that I can make some decisions.

http://www.youtube.com/watch?v=aAVdStF8hNQ&NR=1


Is this 4 Real? Can't tell by youtube footage. Has anyone actually used these firmware upgrades. I have a GF1 and was thinking about getting a firmware hack for it and came across these links. I have done some research and have read a few post on the GH1 but nothing on the GF1. If know of 1 please let me know, but for now any insight would be much appreciated.

Thanks

720p on the gf1 is going to be the best resolution because to get 1080 the camera upscales 720 so its not true 1080.

I am fairly new to this but Ipowell had a 1920x810 which I think currently has the best resolution so far out of the sensor.

 

[mpgxsvcd]

Sure it is for real. The GF1 is capable of recording 1920x1080 @ 30 FPS with 4:2:2 color once it is hacked. However, that doesn't mean it is better than the hacked 100+ mb/sec MJPEG video the GF13 can produce. Like Dex4 said the GF1 1080p is just upscaled 720p. You are much better off doing that in post processing if you need 1080p. However, the hacked 720p is extremely good. Check out the videos below. These were all done with a GF1.

Proof that the GF1 upscales
http://vimeo.com/13135705

High Quality Indoor footage
http://www.youtube.com/watch?v=HeRGcZwOwDs

High Quality Moon Video
http://vimeo.com/16982743

 

[GRENCH]

Thanks dsx4 & mpgxsvcd.

I'm noticing that these hacks are at different mb/sec. Some higher than others. The highest I saw was 100mb/sec. Correct me if I'm wrong but the higher your mb/sec rate is the better your image should be. So why are most of these post only 30 mb/sec and 44 mb/per sec.

Also I wish people would stop doing image test with their lenses wide open. A stopped down image would allow for us to see a broader image. Depth of field overboard on most of these test. very distracting. A tripod would be nice also.

 

[Camera Expert]

Reliability issues. The higher the bit rate the better the video quality but it'll also increase the chances of the camera stop writing so if your shooting a very important event and you need the camera to be rolling non stop for a while, setting the bit rate extremely high may not be a good idea at all especially if your getting paid.

 

[Lpowell]

Like Dex4 said the GF1 1080p is just upscaled 720p. You are much better off doing that in post processing if you need 1080p.

Sorry, mpgxsvcd, but I don't think your video is proof that the GH1/GF1 scales its MJPEG images from a 1280x720 frame. It shows only that with high quality post-processing, you were able to produce results comparable to a 1080p MJPEG frame from an upscaled 720p image.

The camera clearly must downscale its video frames from the 4000x3000 image sensor's native resolution. The MJPEG encoder has no inherent preference for a 1280x720 frame size and there is no technical advantage to using that frame size as a scaling standard. While Panasonic has not disclosed its scaling algorithms, the one used to produce MJPEG frames is capable of multiple resolutions: 1280x720, 848x480, 640x480, and 320x240. Given this requirement and the camera's directly programmable scaling patches, I think it most likely uses a single general purpose algorithm that directly scales the 12 megapixel sensor output to any specified frame size.

There is an additional advantage to scaling the frame to the desired size in-camera rather than in post-processing, since it is done before the image is encoded in MJPEG. Even the highest MJPEG quality levels will degrade the tonal range of the image, and upscaling can make these defects worse. The camera's direct access to the RAW 12-bit Bayer Filter output of the image sensor allows it to produce native quality levels that often cannot be matched in post.

 

[mpgxsvcd]

Sorry, mpgxsvcd, but I don't think your video is proof that the GH1/GF1 scales its MJPEG images from a 1280x720 frame. It shows only that with high quality post-processing, you were able to produce results comparable to a 1080p MJPEG frame from an upscaled 720p image.

The camera clearly must downscale its video frames from the 4000x3000 image sensor's native resolution. The MJPEG encoder has no inherent preference for a 1280x720 frame size and there is no technical advantage to using that frame size as a scaling standard. While Panasonic has not disclosed its scaling algorithms, the one used to produce MJPEG frames is capable of multiple resolutions: 1280x720, 848x480, 640x480, and 320x240. Given this requirement and the camera's directly programmable scaling patches, I think it most likely uses a single general purpose algorithm that directly scales the 12 megapixel sensor output to any specified frame size.

There is an additional advantage to scaling the frame to the desired size in-camera rather than in post-processing, since it is done before the image is encoded in MJPEG. Even the highest MJPEG quality levels will degrade the tonal range of the image, and upscaling can make these defects worse. The camera's direct access to the RAW 12-bit Bayer Filter output of the image sensor allows it to produce native quality levels that often cannot be matched in post.

That video shows a 1920x1080 JPG taken straight out of the camera and a 1080 MJPEG sample. The 1080 MJPEG looks almost identical to the high bit rate 720p MJPEG. The 1080 JPG looks light years better. What more proof do you want? I think I posted the original file for download as well. There isn't much more I can say if you still think the GF1 can do native 1080p. It simply cannot.

 

[Lpowell]

What more proof do you want?

Your test does not conclusively demonstrate that a 1280x720 frame was used as the starting point for producing the 1080p MJPEG image. That hypothesis is based on the assumption that 1280x720 is the native resolution of the MJPEG video modes, but there's no technical evidence in the hack of that being the case.

To prove your claims, I'd want to see evidence of periodic scaling artifacts in the upscaled image. 720->1080 is a 2:3 scaling factor, and scaling algorithms can produce excellent results at small-integer ratios like this. Upscaling 720p by say 10% (10:11 scaling factor) would produce periodic distortion on straight lines that should stand out more clearly. It would be more convincing if you were able to show the presence of these particular scaling artifacts in a non-standard size native MJPEG image.