GH2 Cinema Mode is the GH2's best kept secret

I'm on the fence about -1 vs -2 for cc in post, but I use -1 unless noise is intrusive, so for me the settings relate to light, ISO and the presence of flat backgrounds which emphasize noise. Contrast has to come down at least to -1 for some noise reduction, but of course contrast relates a bit to the lens as well. With the Olly 45mm I have no hesitation in lowering the contrast slightly, other lenses don't yield the same terrific contrast across the frame at lower F stops. In any case, -1/ -1 is an OK place to start, new ppl to the cam have to start somewhere and many ppl will start with a kit lens that is slightly less sharp and contrasty. I'm hoping Panasonic will address the NR issue in the next firmware.

Contrast is easy to add but difficult or impossible to remove. So always set contrast to minimum on the GH2 (it'll not be flat like those ugly superflat Canon profiles that completely messed up the images instead of helping the grading process).

Saturation is where I don't get why people lower it. You can remove saturation easily, but adding is difficult or impossible (since it's not a RAW image file, it's a compressed H264 stream).

I always link this old example shot with my GH1 and 14-140: http://vimeo.com/17728149. In the start you have a super saturated digital looking image but through grading you achieve a more natural looking result. It's a little too sharp but that's because of the 14-140's innate sharpness and I was also using a 1/500 shutter for a special effect in 60p (since it was a flashback).
 
Last edited:
Contrast is easy to add but difficult or impossible to remove. So always set contrast to minimum on the GH2

This is true in theory, but on the GH2 the contrast control works in a particular way. Now, obviously, by altering contrast you get no more bits to play with - you are still limited to 8 bit samples; the contrast can only affect how these are allocated. To me, I feel as if I want to get it as close as possible in-camera so that as few of these 'bins' as possible are wasted. In the GH2, I believe that when you adjust the contrast curve, the overall dynamic range isn't really altered much at all, leaving the blackest and whitest points at around the same values, and it only alters the curve in the middle. When you lower the contrast, it boosts shadows and flattens the mid tones, allowing for a flatter curve in the mid tones and highlights. As a side effect, the shadows are raised (stretched, rather than flattened) which brings out more detail (which may be noise) in the shadows.

This may be desirable if you were going to boost shadows and flatten highlights like this in post anyway, since now you won't have to manipulate as much and you'll get less banding. But if you then want to go back to a more contrasty or less "flat" look in post, you'll need to crush that extra detail in the shadows out again, which will have some small impact upon your image quality. Some of the extra bandwidth set aside for these elevated shadows will have been allocated in vain, and you will potentially increase banding in midtones by stretching them out.

I'm not saying there is one contrast setting that all people should use - I believe you should probably consider what look you want to get in the final product rather than going for a one size fits all approach. If you're interested, I personally am using Contrast "0" a lot of the time right now.



I am very keen to see the work mpgxsvcd mentioned he is doing, measuring the various film mode settings.
 
Last edited:
This is true in theory, but on the GH2 the contrast control works in a particular way. Now, obviously, by altering contrast you get no more bits to play with - you are still limited to 8 bit samples; the contrast can only affect how these are allocated. To me, I feel as if I want to get it as close as possible in-camera so that as few of these 'bins' as possible are wasted. In the GH2, I believe that when you adjust the contrast curve, the overall dynamic range isn't really altered much at all, leaving the blackest and whitest points at around the same values, and it only alters the curve in the middle. When you lower the contrast, it boosts shadows and flattens the mid tones, allowing for a flatter curve in the mid tones and highlights. As a side effect, the shadows are raised (stretched, rather than flattened) which brings out more detail (which may be noise) in the shadows.

This may be desirable if you were going to boost shadows and flatten highlights like this in post anyway, since now you won't have to manipulate as much and you'll get less banding. But if you then want to go back to a more contrasty or less "flat" look in post, you'll need to crush that extra detail in the shadows out again, which will have some small impact upon your image quality. Some of the extra bandwidth set aside for these elevated shadows will have been allocated in vain, and you will potentially increase banding in midtones by stretching them out.

I'm not saying there is one contrast setting that all people should use - I believe you should probably consider what look you want to get in the final product rather than going for a one size fits all approach. If you're interested, I personally am using Contrast "0" a lot of the time right now.

I am very keen to see the work mpgxsvcd mentioned he is doing, measuring the various film mode settings.

Valid point.

I'm not sure the difference would be great, specially since the difference between the different contrast settings is so small.

But that still asks for some experimentation.
 
I have no empirical reason for doing this, but I go for " just a bit under". I want to be in the ballpark so I don't have to push anything too far in post, but I don't want too much of anything.
Unfortunately, this means often underexposing a tad more than I would like to avoid blown highlights.....I have had the experience with -2 sats that pushing sat in post it just went only so far....probably in combination with shooting a bit under in exposure.
 
Saturation is where I don't get why people lower it. You can remove saturation easily, but adding is difficult or impossible (since it's not a RAW image file, it's a compressed H264 stream).

Saturation is a complex thing. Increasing saturation doesn't just increase chroma and mean more bits allocated to it. Saturation is about colour purity, and increasing saturation means removing complementary colours from a colour... a highly saturated yellow has no blue in it, for example. Changing saturation doesn't affect H.264 encoding in quite the same way contrast does. The biggest problem of shooting with high saturation is you can clip the colour space the camera records in. My advice? Pick your personal preference between -2 and 0 saturation, but you probably want to avoid +2.
 
Saturation is a complex thing. Increasing saturation doesn't just increase chroma and mean more bits allocated to it. Saturation is about colour purity, and increasing saturation means removing complementary colours from a colour... a highly saturated yellow has no blue in it, for example. Changing saturation doesn't affect H.264 encoding in quite the same way contrast does. The biggest problem of shooting with high saturation is you can clip the colour space the camera records in. My advice? Pick your personal preference between -2 and 0 saturation, but you probably want to avoid +2.

Yes, of course. I shoot at 0 sat.
 
Do you have an ETA, yet?

looking forward to it.

Well the new firmware came out last night so that has jumped to the top of the queue. I hope to do the new firmware video tonight. A slow motion video that is done already. I just need to upload it after that and then I should be able to do the color testing video with the new firmware after that.

The color testing video is hard to do because it is so subjective. I am trying to just keep it to reporting the facts and let everyone determine what they like based on that data. Collecting that data is time consuming though.
 
mpgxsvcd,
Can you share in layman's terms what is different with your hack, compared to the others in that bitrate range?
 
Saturation is where I don't get why people lower it. You can remove saturation easily, but adding is difficult or impossible (since it's not a RAW image file, it's a compressed H264 stream).

Because if saturation starts to clip, (big RED or BLUE lights for example) it's better to lower it so it doesn't clip as much. Saturation is relatively easy to add back.
 
I'm looking forward to the video on film modes also! thanks
I am shooting the demo clips for the color mode tonight. I am aware of the yellow shift that everyone is seeing in the highlights for Cinema. I will be specifically looking at what causes that and if it can be remedied in camera.

I plan on testing Cinema and Smooth for all ISOs from 160 to 12,800 for stills and 160 to 25,600 for video(Yes I believe that it actually goes that high with the shutter priority mode exposure trick). I will test those ISOs for Cinema and Smooth for hacked(176 mb/sec I frames, 42 mb/sec stock GOP, and total stock) firmware vs. new Panny firmware. I will also run through all color modes at ISO 160, 800, 1600, and 3200 for stills and video.

In addition I will test the ISO bug for changing it from a lower ISO to higher ISO vs. changing from a higher ISO to a lower ISO.

This testing is going to take a long time but I plan on doing it right. I will setup everything in a pitch black garage with a tripod and with a single bright video light to give the scene maximum dynamic range. I will make sure that the camera doesn’t move at all inbetween shots and that the video light maintains an even brightness throughout the testing. I need to get some more AAs to insure that.

I will post the review video to youtube. However, I will also put all of the original video files for any results that are questioned to Vimeo for download.
 
I am shooting the demo clips for the color mode tonight. I am aware of the yellow shift that everyone is seeing in the highlights for Cinema. I will be specifically looking at what causes that and if it can be remedied in camera.

I plan on testing Cinema and Smooth for all ISOs from 160 to 12,800 for stills and 160 to 25,600 for video(Yes I believe that it actually goes that high with the shutter priority mode exposure trick). I will test those ISOs for Cinema and Smooth for hacked(176 mb/sec I frames, 42 mb/sec stock GOP, and total stock) firmware vs. new Panny firmware. I will also run through all color modes at ISO 160, 800, 1600, and 3200 for stills and video.

In addition I will test the ISO bug for changing it from a lower ISO to higher ISO vs. changing from a higher ISO to a lower ISO.

This testing is going to take a long time but I plan on doing it right. I will setup everything in a pitch black garage with a tripod and with a single bright video light to give the scene maximum dynamic range. I will make sure that the camera doesn’t move at all inbetween shots and that the video light maintains an even brightness throughout the testing. I need to get some more AAs to insure that.

I will post the review video to youtube. However, I will also put all of the original video files for any results that are questioned to Vimeo for download.

Thanks a lot for doing this as it is quite a bit of grunt work. I'm sure many of us will appreciate it.
 
Please, as you see your results, give us your opinion. Too many times we are presented with the raw data and told to draw our own conclusions, there is a real value to knowing what you think, especially since you are a long time poster and source of trusted information.
 
I disagree a little bit here. With things that are measurable like resolution I can give my opinion and feel safe in assuming that more resolutions better in everyone's mind. However, with the color modes I might prefer a little more red. Or I might prefer saving highlights over shadows but everyone else might think the exact opposite.

What I plan on doing for this test is giving my opinion in a video. For those who want the short answer and want to accept what I say without validating it then they can just watch the video and decide for themselves.

However, for everyone else I am going to post the original unedited footage and let them decide on their own what they like and what they don’t like. I just want to be the conduit for information and not really the person that changes their minds.
 
Last edited:
Okay, cam's out of my hands now for a while and I didn't get to do this test. I'm looking for this specifically:

An underexposed situation (night exterior or underexpose interior) where the primary subject, which should be a person, is at least a half stop underexposed, or a full stop. THe rest of the image can have portions exposed or over-exposed. It does not have to be a moving image.

If someone could record about 10 seconds of each and upload, that would be great.

Cinema -2 -2 -1 -2
Standard -2 -2 -1 -2
Smooth -2 -2 -1 -2

OR, if you already have footage like this with a human in it, please send.

I'm finding that I can't underexpose smooth or I lose what's important. It's more like over-exposing it (but not blowing anything out), leaving it like that, then dropping it back in post to get the right look. If Cinema mode allows you to underexpose, then it's gonna be what I use on interiors and night material.
 
mpgxsvcd,

HUGE, HUGE Jedi points for this effort Sir!

Huge...

I lost track of time last night and went to bed at 3:30 AM. I woke up at 7:00 sharp not a happy camper. At least I got more sleep than last week. I went to bed at 6:00 AM and woke up at 7:00 AM. It wasn’t even worth going to sleep then.
 
Back
Top