Official Panasonic G5 topic

Vitaliy_Kiselev

Yep, it is very sad.

This is why I need to push more working on manual patches.

stoney

@LPowell FYI Gf2/Gf3/Gx1 can't really lock the shutter speed with Flicker Reduction menu option. Sometimes the bodies changed ISO and shutter speed when the exposure is too low or too high.

adventsam

In the user manual I thought I read somewhere that ae-lock can be set but only when the camera is in manual focus? is it a weird thing like that?

@ LPowell Sorry, I meant "I set the ISO to AUTO". I really did the test the way you described. Setting ISO to AUTO doesn't help to lock the exposure. Probably the ISO/gain in video is always automatically adjusted anyway. And so is the exposure. There seems to be no way to fix the exposure. You can't even see what exposure the camera chooses, you can't compensate it while recording and you can't lock it. @ oscillian Pressing AE lock both before and during recording have no effect. I will do some tests with manual lenses later on. That might still be a good solution for more manual control.

LPowell

@8088

I set the ISO to manual, the flicker reduction to 1/50 and the dial to "A".

That's almost, but not quite the same test I'm hoping you can do for me. What I'm interested to know is how the camera performs when ISO is set to Auto, not manual. In Auto-ISO mode using Aperture Priority with Flicker Reduction, I expect the G5 will adjust the ISO rather than the shutter speed, in order to maintain correct exposure. The Exposure Compensation button should allow you to alter the exposure it selects. If you then press the AE-lock button, it should lock down all exposure settings.

Now press the video record button, and observe what happens. If all exposure settings stay the same, then the G5 can be used this way to shoot video in semi-manual mode. If not, it will need to be hacked before it will be usable as a video camera.

oscillian

@8088 So with a manual lens it should work with only ISO changing. Did you try to press AE lock after you started recording?

@LPowell

Luckily I didn't buy my G5 for doing video. The lack of manual control is pretty disappointing.

I set the ISO to manual, the flicker reduction to 1/50 and the dial to "A". Just like you suggested. Turns out there is no way to control the aperture when recording a video. The aperture information isn't even displayed on the screen. Just the shutter speed info is there when in flicker reduction mode (it stays 50/60 obviously). So even when the camera is set to Aperture Priority, the aperture keeps changing. I can see the iris blades moving.

Exposure compensation can be set but only before recording. When recording everything is locked.

AE lock has no effect when in video mode. I locked AE before recording, but after pressing the movie button the camera goes back to auto exposure.

I am not sure about ISO but my guess is that it is always AUTO when recording video.

Even so the G5 is great. For photos.

@LPowell I will run the test first thing tomorrow morning (central european time).

LPowell

@8088 Could you run this test on the G5 for us:

Turn on the flicker reduction option to lock the shutter speed to 1/60 (or 1/50 in PAL mode).
Set the Mode dial to A (aperture priority) exposure mode.
Set ISO to AUTO.

Questions:

When you change aperture manually, does shutter speed stay constant while the G5 adjusts ISO to maintain correct exposure?
When you change the Exposure Compensation, does shutter speed stay constant while the G5 adjusts ISO to change the exposure?
When you record a video, do aperture and shutter stay constant while ISO varies with illumination?
If you press the AE Lock button before recording a video, do aperture, shutter, and ISO stay constant during the video take?

If the answer to all four questions is YES, the G5 can be used to record videos semi-manually much like the Nikon D5100, which allows you to preset aperture and shutter speed manually while auto-ISO controls exposure.

OASYS

@8088

So far I can tell that the firmware of G5 crops the Raw file from the very beginning. As you can see, the file size of G5 is also smaller than that of GH2. My feeling is that G5 actually uses different sensor from GX1/G3 because their image sizes are slightly different. So there's still hope to get the multi-aspect mode back with a hacked firmware. Maybe a complaining letter to panasonic can do some help as well.;-)

Thanks for your help again.

Cheers.

Thanks OASYS for checking this out. Does this mean the G5 sensor is smaller than the GH2 sensor? I was hoping that the G5 had the same multi-aspect possibilities as the GH2 and that those only had to be unlocked with a hack.

OASYS

Thanks @8088

Now it's confirmed. It turns out no luck. The exif shows:

Sensor width/height (16:9): GH2: 5208x2816 (17.2MB) G5: 4816x2608 (14.7MB) GX1: 4760x2592 (14.6MB)

Image size (16:9): GH2: 4985x2816 G5: 4652x2608 GX1:4597x2592

So it is restricted by the firmware. Also I checked the border beyond the image width. There're all junk pixels (see the right corner of the left picture). However, we can still see that the Raw file has larger field of view than that of video (right picture). The extra area should be for the lens correction.

@OASYS

Here you go. Let me know if you need anything else. http://preview.tinyurl.com/8r54dp3

OASYS

@8088

Yes, it would be great. A 16:9 Raw file and a short clip (MTS file) taken from the same scene would be enough. So we can also compare the angle of view and the pixels used for video and still mode.

I own an G5. Do you want me to send some Raw files?

OASYS

Hope someone can provide a G5 Raw file in 3:2 or 16:9 so we can see the possibility of multi-aspect function.

oscillian

The G5 could be a sleeper!

cmountman

May I ask? As The G5 is about half the price of the GH3 and will share the same sensor and frame rates that are ideal Is there plans for a hack? If so will it be any time soon as this seems more appealing to buy than the GH2 due to the new improved MOS sensor and ergenomic (grip).

OASYS

@lliasG

Thank you for the detailed explanation. It makes sense to me now. When I looked into DxOmark's measurements, I tried to figure out what will really influence our photos. It's surprising to see that in fact the SNR 18% of the best APS-C and M43 sensors are so closed (here I refer to GH2, G5, GX1, EM5, 600D, and NEX-5N..). They're almost within a small error bar. Then I turned to see color depth and DR. Though the differences are more significant, most people use a 8 bit jpg file as their final output. So again I wonder how DR and color depth would impact our photography in real life. Since I'm a user of GX1 and GH2, I really like to know if my cameras are really BAD in still quality as everyone says in the forum. Are they really far behind from the most advanced cameras in the final output? Now it's clearer for me.

Thanks again, lliasG and also Vitaliy's great work. Hope we can see a new Ptool very soon.

Here is an article I found for people interested in RAW and DR:

http://www.normankoren.com/digital_tonality.html

Cheers.

IliasG

@ http://www.personal-view.com/talks/profile/1650/OASYS "You're right. I assumed DxOmark measures the sensitivity of a sensor before any further processing such as gamma conversion. In a 12bit Raw file, it should only have linear data so in principle the maximal DR is 12Ev. That's what I cannot understand. (DR of G5 is 11.6Ev, that of EM-5 is 12.3Ev?)"

OASYS these numbers are the "Print DR" where Dxo theoretically calculates the DR not per pixel but after resampling 16Mp to 8Mp. This gives 1/2 stop increase. The per pixel DR is the "Screen DR" in our case 11.1 and 11.8 stops.

There is more to this....

Your assumption that raw files are linearly encoded is (almost) correct. I wrote (Almost) because there is a suspicion for some recent models (D7000, D800, D4) that the manufacturers started to use non linear data (at the brightest 1/3 stop). Some raw file researchers observed that the SNR curves which DxO provides don't exactly fit on the theoretical model .. thus the suspicion. But even so, they (the manufacturers) linearize the non linear data before they write the raw file so that all raw converters do their job on linear data... and in our case DxO measures are on linear (or maybe linearized) raw data.

Your assumption that a 12bit raw pixel limits the DR at 12 stops is wrong because per definition DR is Max_signal/noise. Well ... this is the so called "engineering DR", DxO uses a slightly different definition Max_signal/Level where SNR=1.
Keep in mind that noise is stochastic value and so it can be a fraction of the used unit. If noise could be zero DR could be infinite.... In our case (digital photo) there is always some "read noise" but even if in a magic way a manufacturer can eliminate it at zero, in a n-bit encoded raw there will exist the quantization noise which is the stdev of a uniform distribution http://en.wikipedia.org/wiki/Quantization_error and this equals to 0.29. So the max (engineering) DR that a n-bit file can hold is n+1.8 ..

http://forum.dxomark.com/index.php/topic,198.0.html

OASYS

@Vitaliy_Kiselev

Thanks for the clue. I shall do more research on it. Cheers.:-)

Vitaliy_Kiselev

In a 12bit Raw file, it should only have linear data so in principle the maximal DR is 12Ev. That's what I cannot understand.

It is wrong assumption. Data is not perfectly linear, and some processing is made for raw anyway.

OASYS

@Vitaliy_Kiselev

You're right. I assumed DxOmark measures the sensitivity of a sensor before any further processing such as gamma conversion. In a 12bit Raw file, it should only have linear data so in principle the maximal DR is 12Ev. That's what I cannot understand.

(DR of G5 is 11.6Ev, that of EM-5 is 12.3Ev?)

Vitaliy_Kiselev

@OASYS

You assume perfectly linear sensitivity and no processing.
It can be not the case here.

OASYS

Thanks for the explanation. In my knowledge, a 12 bit Raw file can record lightness from 0 to 4095. For each stop (Ev) is half or double the amount of light, we can do a simple math. For example, -OEv can store 4096, -1Ev can store 2048, and so on. Therefore, it should theoretically stop at -12Ev (2^12=4096). It's really strange that a more than 12Ev dynamics can be measured unless I misunderstood something. Please correct me if I'm wrong.

Howdy, Stranger!

Categories

Tags in Topic

Top Posters