Official Panasonic GH3 topic

sohus

This is why they picked a Sony sensor:

"The new Olympus OM-D E-M5 sensor is an important contribution that brings micro 4:3 cameras roaring back into the game. During Photokina 2012, Olympus unveiled two new cameras in the Pen line, the Pen E-PL5 and Pen E-PM2, both of which are equipped with the same sensor as the OM-D E-M5. These Olympus high-end compact hybrids achieve scores that approximate those of their APS-C rivals (such as the Sony NEX). Simply put, the OM-D E-M5 is equipped with the best sensor we have ever analyzed for a micro 4:3, and to date the closest to that of the Sony NEX."

In other words: APS-C result with a 4:3 body and lenses (lighter and smaller). I like it!

I do think it will take a hit in terms of video quality but it will probably be as good or even better than most other DSLR camera's out there.

DrDave

DXO rates the Olly-presuamably-same-as-GH3-Sony sensor the best of the bunch, but of course that only measures the potential. Wondering if the GH3 HDMI will provide a clue for the cleaning up the GH2 HDMI..................... Of course I wouldn't use the Olly for video, so I hope the GH3 is better as far as that goes.

Aria

I noticed they tested the EM5 at 12EV. If the GH3 is at or slightly above that then it likely will have better DR.

thepalalias

@DrDave @Aria They also rated the sensor better on the GH1 than the GH2 and we all know how the result worked out on the video side. :)

DXOMark does not in any way, shape or form currently address the video-pipeline side of things. The issues with many DSLRs on the video side have to do with how the original data gets output at the video resolution in a given format. The pixel-binning/line-skipping/whatever else method used is a huge factor in that equation.

Like you said, DXOMark can show the sensor's potential, but we are not anywhere near realizing the potential of the sensor in most cases. The color bit-depth, sub-sampling, final output resolution, compression, etc. all mean that we are several steps further from the sensor's potential.

Even on the GH2 (which does a really good job by comparison to most cameras) the pipeline used for video means we are nowhere near the still image potential of the sensor. Even the hacked settings often cannot replicate the detail of JPEG in a high detail scene (let alone the color) and of course none of us are really expecting to be able to get RAW quality either.

So given that baseline, it really is much more about what the camera does with the sensor in video specifically than about what the image potential is like - unless we were recording in RAW at the original resolution and bit-depth.

EDIT: To see the gap between JPEG stills shot at 1920x1080 and video shot with the same settings try comparing the word "Pancakes" or "Ultimate Chicken Fingers" on the box in these pictures at full resolution.

http://personal-view.com/talks/discussion/comment/60506#Comment_60506

DrDave

@thepalais I slightly prefer the video of the GH1 to the GH2 at low ISO, however I can't tell them apart if they are unlabeled. I took a bunch of 100 percent crops of my own, mixed them with a bunch of the good online ones, and I could not tell them apart. I couldn't pick the cam (and I guess no one else could) in the recent demo lpowell did. So I am really buying the cam for the features, which for me are the usual aliasing, moiree, haloing, dynamic range, color and so on. 2K would be a very highly desirable feature. The improved screen is a desirable feature, I really like OLED and I wish I could plug my Samsung Infuse into my GH2.

Aria

I was mainly focusing on the DR of the sensor more than making a pure judgment of image quality which I don't feel I can make at this time. If it's true that the GH3 has a similar sensor to that of the EM5 then it's likely going to have a very significant increase in DR. I do think that is an improvement. When you take that and ALL of the other improvements as a whole the GH3 could be a very worthwhile upgrade.

None of this is definitive but it just goes to show that there's a real chance that the GH3 will indeed have improved DR and Low Light capability. My guess is that we should be able to record at lower ISO's and still bring out the image we need in post with improved DR and Color. It's just a guess from what i've read so far. I've read that the new codec holds up very well grading in post and that combined with the improvements in DR, Color and Noise should really help to achieve a better final product.

DrDave

I just can't understand why Olympus can't make a decent vid camera. IS for legacy glass, even if used just for photos, is amazing. Imagine if it could work for video without weird artifacts, and if Olly had 24p and good controls. I hope the GH3 is a huge success, sells millions, and nudges Olympus into making a real vid cam.

duartix

DxOMark is a fine tool for what it was set to be... a tool to reveal strong points, week points and sweet spots of each sensor RAW output.

Using it to derive conclusions to JPEG or video output is nonsense, as it is to qualify a sensor by one final single number. Most of the negative comments on DxOMark that surface especially at DPREVIEW, is made by mentally lazy or challenged people who look at the final score and couldn't spare 1m looking at the graphs to learn something.

I'm not affiliated with them but I have to say that with all else being similar (unlike banding on the GH1 as an example), it's a great tool to tell you what to expect from your photographic tool.

Vitaliy_Kiselev

@DrDave

As far as I remember, IS is not working for video (only software one).
Also sensor based IS have big troubles with long video shooting just because you can't make any good heat dissipation from sensor. So, panasonic made absolutely right decision. And as we can see, Sony also did the same, removing sensor IS from all E-Mount solutions.

mpgxsvcd

@Vitaliy_Kiselev

What is it about the IS that generates so much heat? I never understood why the IS would not allow the sensor to dissipate heat. Does the IS encompass the entire sensor or is it something that only touches the sensor in small discreet spots?

zigizigi

There's no need to physically move the sensor. Most likely Oly just reads 3D data from the motion sensors and calculates compensation angainst any shake based on that data in video mode. From what I can see in Oly clips it's more than enough for video. Probably physically moving sensor to compensate gives some extra point for stills, but who cares.

Panasonic already has similar technology in their latest video cameras that have 5 axis HYBRID O.I.S.

I don't think that AC90 actually has a mechanism to move the huge 3-CMOS censor block. Quite unlikely.

So why not having those motion sensors inside the camera and use their data either to fasilitate in-lens OIS or add some extra software stabilization. I dunno how it works exactly but whatever it is it works great and that would be a brilliant feature.

Vitaliy_Kiselev

@zigizigi

Software stabilization has it's own issues from introduced crop, to sudden blur and freezes.

@mpgxsvcd

What is it about the IS that generates so much heat? I never understood why the IS would not allow the sensor to dissipate heat.

It is very simple. Because IS is moving sensor to compensate, so, instead of fixed sensor and proper contact with massive heat spreader all you have is sensor assembly, that is made as light as possible. Hence, while generated heat is the same, heat can't go away as efficiently, and overheating occurs.

DrDave

Thanks Vitaliy--I was wondering why the video on my Olly looks a bit funny with the IS but the photos look good--it must be that it is running in software.

magnifico

My point of view about gh3, considering i want to upgrade from gh2. My 10 whishes (after VKs hack of GH2) are these :

1/ a firmware hackable and almost open source for some functions : GH3 -> ??

2/ VERY IMPORTANT : detail >1.1x GH2 : GH3 -> ??

3/ moire : GH3 -> ??

4/ alias : GH3 -> ??

5/ DR in video : : GH3 -> ??

6/ colors : GH3 -> seems OK (hope so)

7/ photo, i need RAW IQ > OMDs IQ. DXO gives 60 for gh2 and 71 for OMD. I just need a 75 mark for GH3:-) I know IQ is not a DXO mark only: GH3 -> ??

8/ size smaller than gh2 (we are talking about m43, man): GH3 -> NO

9/ sound 3.5 mm : GH3 -> OK

10/ 1080p60 : GH3 -> OK

Point 2 3 4 5 6 7 are not negociable for me: I need both of them. So what are my chances to upgrade?

svart

@magnifico I think you'll know when the rest of us know. The problem is that there is way too much conjecture about the GH3 and it's abilities right now. Panasonic was vague on a lot of things that we feel are important and went into detail about things that Panasonic thinks it's customers feel are important. I think the folks on this forum get too caught up in believing that we are Panasonic's most important customers. We aren't. While Panasonic seems to actually take a lot of information from this forum and use it in their decisions, we are only a small number of potential customers. It's actually amazing that they even acknowledge we exist at all.

magnifico

@svart I just wrote my check list before upgrading. It is just my personal point of view. Will be back to that list when we get actual information for each point :-) Note that my whishes shouldnt be the same if VK didnt hack GH2. (thx VK for fighting "obsolescence programmee")...

den15_de

Dear Vitaliy, tell us please if you know, is it possible in the future to expand DR a little bit at GH2 and GH3 with some hack? The suspensed DR is the main thing that don't allowed me to make GH3 preorder.

Kihlian

8/ size smaller than gh2 (we are talking about m43, man): GH3 -> NO

I´ll find that a bigger body is better, better to hold and looks more professional for the customers.

magnifico

@Kihlian I am amateur, I dont care about customers. Those who are shot by cam do not like big cameras :-)

roncoker

Quote from Voldemort interview.

Yoshiyuki Inoue Senior Engineering Planner For Panasonic GH3. The format after processing for HDMI is 4-2-0

Vitaliy_Kiselev

@roncoker

And strangely, it could be false for final products that will be tested.

Aria

@Vitaliy_Kiselev, do you mean that Pany could change the GH3 so that it will output 4:2:0? I ask because Mr. Inoue intimated that it was baked in the hardware cake already. Perhaps his English translation was off, but he said that "This depends on the system internal structure. This one is optimised for the still picture quality also." I guess they limited it on purpose for some technical reason having to do with the stills side of things. It doesn't make sense to me but that was his explanation. Whatever the reason Mr. Inoue seemed pretty certain that it was 4:2:0.

Vitaliy_Kiselev

Made

http://www.personal-view.com/photokina/coverage/gh3-wi-fi

guys, please add things that you heads or saw about Gh3 Wi-Fi in it's current form.

Vitaliy_Kiselev

@Aria

He can be pretty cernain, but many times managers get information from strange sources.

ghkqn

I wonder how the GH3 (and all the other cameras supporting wifi) deals with the differences in channel implementation of 802.11b/g/n between US and other countries.

Howdy, Stranger!

Categories

Tags in Topic

Top Posters