panasonig g6 softer in 50p compared to 25p?

[Noa Put]

I saw a user report that showed notably softer image at 50p compared to 25p, I also read the review of the website author which doesn't mention this but honestly am not taking him too seriously, especially because he doesn't find it nesassary to comment on that user report.

So I rather ask here to see if there are any g6 users that have seen this behaviour, is there something going on in 50p mode that softens the image?

[William Hohauser]

Without direct experience with the camera I feel comfortable guessing that putting 50p in the same bandwidth stream as 25p would result in some sort of compromise in image quality.

[Noa Put]

I don't see that behaviour in the gh3 so that's why I"d like to have it confirmed by someone who knows.

[Steve Game]

Originally Posted by William Hohauser

Without direct experience with the camera I feel comfortable guessing that putting 50p in the same bandwidth stream as 25p would result in some sort of compromise in image quality.

That was my first thought but usually, 50p is offered as a 28Mb/s stream whereas 25p is normally only available up to 24Mb/s.

Edit: Although the Panasonic UK website fails to state the recording bitrates, I've found a review that give the following info:

50p (PSH mode) - 28Mb/s
25p (FPH mode) - 24Mb/s
25i (FSH mode) - 17Mb/s

The data rates are similar to Sony and Canon modes so if there is any subtle difference in image sharpness, it may be that the pre-compression processing has a bandwidth limitation.

[Pete Carney]

I've shot a ton now with the G6 and tested all the different modes yesterday. There's no difference in detail of the different modes.

I will suggest that guy has an issue with importing 1080-60p on his machine. 60p on a 30p timeline requires that you turn off frame blending or resample as some of the software calls it, or you get a blend of two frames from 60p down to 30p. Vegas for instance requires you select "disable resample" on every 60p clip or you get ghosting on every frame with significant movement and very soft images when there is very little movement.

This camera is proving to be incredible after years and years with miniDV camcorders then Canon DSLR's the last few years. There simply is no camera made with all the features and image quality the G6 produces, especially for the long range type work I tend to do most often.

Here's a couple samples from last weeks America's Cup training in San Francisco.

Emirates Team New Zealand on Vimeo

1080-60p in a 30p timeline

Oracle Team USA on Vimeo

1080-60p slow motion to 24p

Cheers,
Pete

[Noa Put]

Thx pete, just what I wanted to hear, the g6 will be a great b-cam to my gh3 and I"m saving money as well. (great lookng video btw)

[William Hohauser]

Originally Posted by Steve Game

That was my first thought but usually, 50p is offered as a 28Mb/s stream whereas 25p is normally only available up to 24Mb/s.

Edit: Although the Panasonic UK website fails to state the recording bitrates, I've found a review that give the following info:

50p (PSH mode) - 28Mb/s
25p (FPH mode) - 24Mb/s
25i (FSH mode) - 17Mb/s

The data rates are similar to Sony and Canon modes so if there is any subtle difference in image sharpness, it may be that the pre-compression processing has a bandwidth limitation.

Actually that confirms my point although I didn't know about the bit rate increase for 50p. It's twice the frame rate yet less than a 20% increase in bandwidth. That will cause some additional compression to be applied.

[Nigel Barker]

Originally Posted by William Hohauser

Actually that confirms my point although I didn't know about the bit rate increase for 50p. It's twice the frame rate yet less than a 20% increase in bandwidth. That will cause some additional compression to be applied.

It may not be all that much given the way that MPEG4 or indeed any video CODEC works. It may be twice the frame rate but it's not twice the number of frames like if you were using film. These cameras use he same bit rate compression for 24/25/30p & I don't think we see any quality differences.

[William Hohauser]

Not to get nitpicky but it would certainly depend on the image being compressed. A hand held shot where every frame is unique will require more bandwidth to maintain quality at 50p, a tripod locked sit down interview may not require much more bandwidth between 25 and 50.

[Steve Game]

I'm not so sure that additional bandwidth required (for an equivalent picture quality) is that much more. To reconstruct an visually acceptable picture stream in a given scenario of detail and movement with respect to time, 50p video compression will need less inter-frame data to reconstruct each frame than 25p. In other words, as each frame will have 20ms of movement at 50p, it will only be 40ms per frame for 25p.
I hope my clumsy explanation above makes sense.

[Noa Put]

I rather see an image to prove a point, it's nice to know the technical reason behind it but they mean nothing if you can't actually show what it's like.

[William Hohauser]

I don't have the camera but the water test would prove what the camera is doing between 25p and 50p. Simply a shot of actively moving water on a lake or pond, filling the screen, taken at 25 and then 50. The reason that this is a good test is that there is no data from the main frame of the GOP that can be used as reference as the water is constantly changing. Each frame has to be new or strange artifacts resembling old web video will occur. But since the reference frame which is alloted the highest amount of data has passed, the following frames will be degraded in some way.

Check the blocking of the compression by zooming into the image and going frame by frame if it's not apparent immediately, a moving picture will cover many ills (which is good for us). If the images from the two frame rates are similar then Panasonic has implimemted some method to cope with high stress compression situations like this. It's situations like moving water that some people are so interested in external recorders that have three times the data rate or in the case of the GH2 create hacks with names like "FlowMotion" to up the data rate.

What does it all mean? Is the G6 a "bad" camera if 50p turns out to be slightly softer than 25p? That all depends on your use of it. Last month as a studio manager I witnessed a documentary filming using the interesting technique of surrounding their interviewee with four G6 cameras on tripods while a Canon DSLR model (maybe a 5D) took the main shot, all arranged in a neat semi-circle. They were shooting in 24p but I don't know what each camera shot ended up being. When I was in the studio as they were setting up, each camera had the same medium shot, of course from a different angle. How they will integrate the footage is an unanswered question (and I probably will never film this way) but they were confident that the results will edit well with the Canon. If you are getting results to your satisfaction and other people can look at your work and enjoy it without being distracted by camera "defects" then it's a great camera.

[Rob Hargreaves]

According to Philip Bloom running "overcranking" in camera yields a better bitrate, than shooting 50p and conforming in post to 25p (PAL land) as the bitrate is halved...

Canon XA25 Review on Vimeo

I wasn't aware of this can anyone confirm this?