Would You Perhaps Be Interested In A Different GX80/85 Colour Profile???

[BTM_Pix]

23 minutes ago, jonpais said:

It's been a little while now since BTM_Pix shared his hack for the GX80, but still no sample footage. I'd be interested to see whether the profile really resembles Cinelike D in the G85. The following is straight out of the camera, no color correction, no sharpening, nada. It has not been exposed to the right, it is just for comparing how skin tones and dynamic range compare with your footage. 

 

Though not motion, I  put a post on the first page of two frame grabs from footage recorded with Cinelike D on both the GX80 and the G7.

To my eye they show identical handling of the same scene. Only to my eye though so there may be some on much closer inspection but I think for all practical purposes they are the same.

I don't know the exact family relationship senso and processor wise between the GX80 and the G7 so any small differences between the performance of Cinelike D on both are likely attributable to that.

Its impossible to do a direct comparison between what official Cinelike D would look like on the GX80 of course as there is no official one but just to stress that no other parameters were harmed in the making of the hack. 

I agree that I would love to hear from people who have implemented it to see what difference if any its made to their output and any difference they've had to make to their workflow either capture or post to accommodate it.

I'd be particularly interested if anyone has a Cinelike D based LUT to see that it maps across as expected.

[jonpais]

2 minutes ago, BTM_Pix said:

Though not motion, I  put a post on the first page of two frame grabs from footage recorded with Cinelike D on both the GX80 and the G7.

To my eye they show identical handling of the same scene. Only to my eye though so there may be some on much closer inspection but I think for all practical purposes they are the same.

I don't know the exact family relationship senso and processor wise between the GX80 and the G7 so any small differences between the performance of Cinelike D on both are likely attributable to that.

Its impossible to do a direct comparison between what official Cinelike D would look like on the GX80 of course as there is no official one but just to stress that no other parameters were harmed in the making of the hack. 

I agree that I would love to hear from people who have implemented it to see what difference if any its made to their output and any difference they've had to make to their workflow either capture or post to accommodate it.

I'd be particularly interested if anyone has a Cinelike D based LUT to see that it maps across as expected.

Right, but seeing stills is not the same as seeing actual, moving footage. Paul Leeming's LUT is the one to use if anyone is interested.

[BTM_Pix]

45 minutes ago, Grimor said:

Or move the BTM_Pix  hack to another topic.

@BTM_Pix Whats is the method your are using for sniffing/capturing packages from the camera-app communication? 

I can try it with G80 and share with you the results if you want.

Packet Capture on Android and then some manual requests from a browser and/or something of my own creation!

Any additional sessions can bring a bit more to the table if you invoke functions that the G80 has that the GX80 doesn't to see if it can be persuaded to implement them.

I don't have a direct feature comparison list of what's missing but I'm sure there is one somewhere online 

[jonpais]

Here's a clip that's been exposed to the right a bit, not seeing any bizarre color shifts or radical highlight rolloff. Image may appear somewhat more red and darker than when uploading to YT, not sure what's up with inserting media on this website. Guess I'll have to share via YT to show there's absolutely no color shift. It looks pretty close before seeing image full size, though.

[BTM_Pix]

13 minutes ago, jonpais said:

Right, but seeing stills is not the same as seeing actual, moving footage. Paul Leeming's LUT is the one to use if anyone is interested.

Absolutely not, no.

But having seen them both in motion when I have been doing it (other quick tests trying different stuff not static charts) I can definitely say there was nothing different going on to my eye. 

I'm away doing this assignment so I have Peli cases full of Nikon and Fuji's rather than Panasonics but I'll definitely do an A/B when I get home next week.

I'll put it through Film Converts Cinelike D profiles as well to check it end to end.

[kidzrevil]

1 hour ago, jonpais said:

Sorry KRE, but what I'm seeing here are crushed blacks and blown out highlights, not the extended dynamic range that I'm getting at all by using CineLike D. Perhaps I need to calibrate my monitor.

This IS cinelike d. It has been graded to taste. I get way better results exposing to 90 ire thats my point and you can see the same thing in the panasonic white paper and how cine d handles highlights

2 hours ago, Orangenz said:

The phrase "whatever works for you", especially in the repeated usage, really comes across as condescending towards the mod. With it removed your post is much more readable! I do appreciate that you've added a video in support of your workflow. I suggest your discussion of ire clipping levels move to another thread, new if you like, as this is distracting from the good work of BTM_Pix. 

Whoops sorry 

[jonpais]

1 hour ago, kidzrevil said:

This IS cinelike d. It has been graded to taste. I get way better results exposing to 90 ire thats my point and you can see the same thing in the panasonic white paper and how cine d handles highlights

I haven't seen the white paper, but I'm more interested in results, not white papers. I'd also like to see your original clips without any grading, it would be far more useful to readers of this blog than heavily graded clips. I'm seeing heavily crushed shadows and blown highlights in your final grade, which is fine, if that's your intention, but it doesn't prove your point that ETTR at 100% is a bad idea. BTW, I really do enjoy your clips, they show a world I've never seen before.

[jonpais]

I should add that I do appreciate kidzrevil sharing his clips, which is more than most do on forums, here or elsewhere. if I were a newbie, I would be very skeptical of those voicing opinions on matters about which they have no work to back up their claims. If a person has no website or YT channel, I pretty much dismiss their comments outright, for better or worse. It also helps if uploaders show their waveform monitors in their NLE, etc, as well as giving information about the settings used in camera, as well as what was done in post processing, something few people do, but which I always include in my YT videos.

[Georgios]

39 minutes ago, kidzrevil said:

This IS cinelike d. It has been graded to taste. I get way better results exposing to 90 ire thats my point and you can see the same thing in the panasonic white paper and how cine d handles highlights

Whoops sorry 

Is it possible to share the link of the white paper you are referring?

[BTM_Pix]

Just been asked to leave John Lewis midway through hacking their LX15.

Was able to handshake it and fake the connection but they pounced as I was about to hit Deploy!

I'd already set the alarm off doing their GX800. Status on that was that it took the Deploy command without error but called up the Standard profile.However, the lighting in their was so poor and I had to be so quick that I wasn't able to do full compares to the real Standard profile to see if it was actually Cinelike D. 

So file that one under potentially possibly maybe until I have another opportunity. 

I'm intrigued by getting it on to their small cameras, particularly the travel zoom ones.

[kidzrevil]

13 minutes ago, jonpais said:

I should add that I do appreciate kidzrevil sharing his clips, which is more than most do on forums, here or elsewhere. if I were a newbie, I would be very skeptical of those voicing opinions on matters about which they have no work to back up their claims. If a person has no website or YT channel, I pretty much dismiss their comments outright, for better or worse. It also helps if uploaders show their waveform monitors in their NLE, etc, as well as giving information about the settings used in camera, as well as what was done in post processing, something few people do, but which I always include in my YT videos.

Im gonna start doing that more often, posting originals with waveform and such so we can all collectively come up with a better understanding of the hardware. I usually just share links cause its super convenient as Im usually posting on here when im running around the city

10 minutes ago, Georgios said:

Is it possible to share the link of the white paper you are referring?

Here is the link. Panasonic is far from clear but I assume "half video levels" is half a stop ire below reccomended exposure that would put us at 1/3-2/3 underexposre 

https://eww.pass.panasonic.co.jp/pro-av/support/content/guide/EN/AG-AC90handbook_en.pdf

[jonpais]

I'm not sure that reducing gamma (or video level) by 1/2 is the same as reducing exposure by half. It's pretty certain from the screen grabs I've shared in another thread comparing Natural, Standard and CineLike D that the last thing you'd want to do is to underexpose CineLike D.

[Brousseb]

30 minutes ago, BTM_Pix said:

Just been asked to leave John Lewis midway through hacking their LX15.

Was able to handshake it and fake the connection but they pounced as I was about to hit Deploy!

I'd already set the alarm off doing their GX800. Status on that was that it took the Deploy command without error but called up the Standard profile.However, the lighting in their was so poor and I had to be so quick that I wasn't able to do full compares to the real Standard profile to see if it was actually Cinelike D. 

So file that one under potentially possibly maybe until I have another opportunity. 

I'm intrigued by getting it on to their small cameras, particularly the travel zoom ones.

I mentioned earlier in the thread that I was able to apply Cinelike D to my zs100 and it appears to be working, it's very close to my GX85 although not 100%, maybe due to the smaller sensor.

[jonpais]

23 minutes ago, Brousseb said:

I mentioned earlier in the thread that I was able to apply Cinelike D to my zs100 and it appears to be working, it's very close to my GX85 although not 100%, maybe due to the smaller sensor.

Share? Anyway, I can't find the link here, but here's my comparison. If anything, it looks like you'd want to give less exposure in Natural or Standard to save the highlights, Cinelike D can obviously benefit from even more exposure. Highlights are clipping faster in the other profiles, am I wrong? Forgive my ignorance about reading graphs, but I was educated in the USA.

https://www.google.com/amp/s/jonpais.wordpress.com/2017/05/25/shouldnt-you-be-shooting-cinelike-d/amp/

[Brousseb]

8 minutes ago, jonpais said:

Share? Anyway, I can't find the link here, but here's my comparison. If anything, it looks like you'd want to give less exposure in Natural or Standard to save the highlights, Cinelike D can obviously benefit from even more exposure.

https://www.google.com/amp/s/jonpais.wordpress.com/2017/05/25/shouldnt-you-be-shooting-cinelike-d/amp/

I have a comparison video at home, I'll upload it once I get a chance.

[BTM_Pix]

28 minutes ago, Brousseb said:

I mentioned earlier in the thread that I was able to apply Cinelike D to my zs100 and it appears to be working, it's very close to my GX85 although not 100%, maybe due to the smaller sensor.

Sorry I was writing a big reply when you posted that so completely missed it

Nice one

That's a TZ100 in the UK isn't it?

[jonpais]

8 minutes ago, BTM_Pix said:

Sorry I was writing a big reply when you posted that so completely missed it

Nice one

That's a TZ100 in the UK isn't it?

 

[mercer]

1 hour ago, jonpais said:

I haven't seen the white paper, but I'm more interested in results, not white papers. I'd also like to see your original clips without any grading, it would be far more useful to readers of this blog than heavily graded clips. I'm seeing heavily crushed shadows and blown highlights in your final grade, which is fine, if that's your intention, but it doesn't prove your point that ETTR at 100% is a bad idea. BTW, I really do enjoy your clips, they show a world I've never seen before.

White papers are written by the people who designed the curve. They understand the best exposure methods to use so skin tones hit at the right place on the curve, so your highlights don't clip, etc... One of the reasons there was so much orange skin with the CineLikeD Profile, back in GH4/G7 days, was because everybody was exposing it to the left... of course they were also dropping contrast too.

With most Log and Flat profiles you need to know where middle gray hits on the curve. Just exposing to the right doesn't give you that. Now if you're using Leeming LUT and his settings, you are probably good because Paul has done extensive research, but if you're not, it's best to nail your exposure in camera, or go slightly under and pull up. I don't have a Panny camera at the moment but I remember when I would ETTR, the footage was clipped on the RGB parade. Pulling it back didn't bring back that highlight information. But underexposing a 1/3 of a stop or hitting dead center and you could pull up the highs and still get some nice rolloff.

[jonpais]

4 minutes ago, mercer said:

White papers are written by the people who designed the curve. They understand the best exposure methods to use so skin tones hit at the right place on the curve, so your highlights don't clip, etc... One of the reasons there was so much orange skin with the CineLikeD Profile, back in GH4/G7 days, was because everybody was exposing it to the left... of course they were also dropping contrast too.

With most Log and Flat profiles you need to know where middle gray hits on the curve. Just exposing to the right doesn't give you that. Now if you're using Leeming LUT and his settings, you are probably good because Paul has done extensive research, but if you're not, it's best to nail your exposure in camera, or go slightly under and pull up. I don't have a Panny camera at the moment but I remember when I would ETTR, the footage was clipped on the RGB parade. Pulling it back didn't bring back that highlight information. But underexposing a 1/3 of a stop or hitting dead center and you could pull up the highs and still get some nice rolloff.

I do own a Panny camera, I expose to the right, and footage isn't clipping the RGB parade. I believe the screen grabs I've shared show that if anything, highlights are more likely to clip in the Standard and Natural profiles, which also have considerably less dynamic range. And I'm not convinced that Panasonic's recommendation to lower gamma levels equals reducing exposure. And the only time my clips are orange with today's Panasonic cameras is when my white balance is off.

[mercer]

25 minutes ago, jonpais said:

I do own a Panny camera, I expose to the right, and footage isn't clipping the RGB parade. I believe the screen grabs I've shared show that if anything, highlights are more likely to clip in the Standard and Natural profiles, which also have considerably less dynamic range. And I'm not convinced that Panasonic's recommendation to lower gamma levels equals reducing exposure. And the only time my clips are orange with today's Panasonic cameras is when my white balance is off.

And you're using Leeming LUT and his settings... right?

But how else would you lower Gamma in camera without lowering exposure? It either has to be lowered with a lighting setup designed so the midtones hit a specific spot on the curve, or through your in camera exposure. Now unless you have an extensive lighting kit, you have to do that in camera. I guess you can argue that you may be losing some highlight information in that circumstance.