GH1 Thorough GH1 FPN Test - Old vs. New bodies, Film Modes, and Emory's Settings

[CColton]

I tested to conclude what combination of body age (2010/2011), film mode, and recording mode produces the least FPN in a GH1.

Before you read or watch ANYTHING below, please keep in mind, I am not Zacuto, I've tried to run a pretty good test, but I'm not perfect, so if you doubt my testing below, please perform the same test and post your results.

My procedure:

1. I took my two bodies, one from pre-June 2010 and one I bought this week, hacked identically (Blackout-Powell 24p) and shot the exact same scene (an old computer lying around) with the exact same lighting, locking the tripod to ensure that both bodies would get the exact same shot.
2. If you're wondering how much light was actually there, know that every one of the clips you're about to see is brighter than the image I see with my eyes, and I have unusually good vision. I can barely read the darker parts of the text on the front of the computer in real life. So even though there's noise, this camera still does pretty well in the dark.
3. White balance differs slightly between the two bodies, I tried to manually match it, but it's not perfect.
4. Both cameras shot at 1080p24
5. All shots were with the same Panasonic 20mm f1.7 lens.
6. All Creative Movie Mode shots were shot at f1.7 with 1/30 shutter speed in order to mimic the exposure that Emory's settings would meter, since Emory's settings are based on Shutter Priority, not full manual.
7. When Emory's settings were used, they were identical to his settings linked in my signature, except in shutter priority, not Program. I held my hand in front of the lens before each shot in Emory's settings and hit AE lock, so that the camera metered the settings to the blackest image possible. When AE locked the camera reported f1.7, 1/30, and ISO 1600, which are identical to the Creative Movie Mode exposure settings. Therefore the only difference between the two modes was that Emory's uses the camera's metering and AE locking, whereas Creative Movie Mode was on full manual.
8. I tested two film modes. The first was Smooth with Contrast -2, Sharpness 0, Saturation 0, and NR -2. This film mode has been generally agreed upon (and in my opinion) as the mode that retains the highest dynamic range and most accurate color of any of the modes, and therefore is (in my opinion) the best for situations where light abounds. The other film mode used was Standard, +2, +2, 0, -2. I've worked this out as a good mode to attenuate noise and retain accurate color, with the Sharpness added to retain detail that's usually lost at high ISO. This is, in my opinion, the best settings for low light.
9. No color grading/correction was done.
10. My 2010 body has been heavily used, so in theory, its sensor might be more "burnt in" than the 2011 body.
11. The first half of the video is the raw results, the second half repeats those results split-screen between the old and new bodies to illustrate the difference in bodies.
12. Edited by import to DNxHD 220x (lossless 10-bit) in Avid Media Composer. Therefore the editing process was lossless, so any compression artifacts visible were the artifacts recorded by the cameras.

I've made a video that combines all the results side-by-side and split-screen that illustrates the differences well. However, since that video is encoded in DNxHD (lossless) the file is huge, and is still uploading. So the link below has all the original MTS files of the tests, renamed according to the settings used in each clip. Download them and see what you think. As soon as the compilation is done uploading it will also appear at the link below, so if you don't see it now, check back later. If you can't play the compilation test video, click on the imaginatively named "Install this if you can't play the video" folder and download the DNxHD codecs for Windows or Macs.

Click here to download the original test MTS files. The compiled comparison video will appear here as soon as it has finished uploading.

Please take a look and tell me what you think the results were. My conclusions are as follows:

• Emory's settings definitely help, even though we/I really don't know how or why. Bravo, Emory!
• The Standard settings have much less noise than Smooth, with the drawback of crushing some detail out of the blacks of the image.
• Honestly I can't tell a huge difference between the bodies. The 2011 body seems to maybe-kinda-sorta look better, but I don't know.

In summary, I'd like your opinion on which body is better, but Emory's settings in Standard +2,+2,0,-2 were definitely the least noisy settings on both cameras.

Please tell me if you agree or disagree with my opinions as to the conclusions that can be drawn from this test.

 

[CColton]

The new file will appear about a half an hour after this post.

 

[MadMonkFish]

Looks like you did a very comprehensive test

Regarding WB - did you set via selectable Kelvin & still see a difference between the two bodies?

I know it'll compress more & lead to pixel peeping, but do you intend to put on Vimeo?

 

[CColton]

Looks like you did a very comprehensive test

Regarding WB - did you set via selectable Kelvin & still see a difference between the two bodies?

I know it'll compress more & lead to pixel peeping, but do you intend to put on Vimeo?

Nope, just manually white balanced on a white card. Should've done Kelvin haha, my mistake.

Pixel peeping is the intent of the test; for people to see what they can do to reduce noise on a GH1. And I might put it on Vimeo on 1080p but that's still compressed, which will take away from the accuracy of the tests.

 

[CColton]

Now the DNxHD test is up!

 

[goodgoings]

Thank you Colton for taking the time to do the test!

I took a quick look at your files and even with/without Emory's settings I don't see much FPN in both models in the scene you shot. Maybe you have good models compared to the one I had. FPN will rear it's ugly head when shooting something in low light, large and white, and panning across it(refrigerator door). You can actually see the bands stay stationary as it pans.

 

[CColton]

You can see some stationary bands on my cameras but you have to be looking for it to see it. If you used anything like NeatVideo on it that could clear it up.

 

[CColton]

After shooting in a practical situation today for a client, I revise my opinion. For various reasons I was forced to use only my 14-140 f4-5.6 panny lens inside, so I had to push the new body to ISO 1600. I changed to my modified Standard +2,+2,0,-2 film mode, and there is very little noise. I revise my opinion to agree that newer GH1 bodies have less FPN/banding (noise at high ISO). On mine, there is noise, but it looks like film grain at the worst, provided your shot is well exposed. If 1600 is still getting you very underexposed, you'll see more noise. Also, on Smooth the noise is more noticeable.

 

[Svart]

Is there any way to post screen grabs? I can't watch videos at work..

 

[CColton]

Frame grabs are in a folder on the files.coltonvideography.com, as well as the following:

 

[DaLiV]

[*]My 2010 body has been heavily used, so in theory, its sensor might be more "burnt in" than the 2011 body.

2011 bodies have very diffirent firmwares
and depsite some peoples words - I can say - that firmware are with much bigger changes than simple adding of antihack check points.
diffirencies
1.34(was unhackable)-1.32(main publicity used)
are bigger than
1.33(very last available firmware, distributed by panasonic france. but even that release already has color schemes changed)-1.32.

1.34 firmware still not leaked to public

 

[stonebat]

Emory's setting crashed black which I like in low light setting.

I guess GH1 is trying to boost dark area. Lower S/N. The banding seems artifact.

 

[booggerg2]

why do the 2010 caps look almost black n' white? .. definitely has less noise.. but the color is gone as well...

Thanks for putting together the test!

 

[CColton]

why do the 2010 caps look almost black n' white? .. definitely has less noise.. but the color is gone as well...

Thanks for putting together the test!

Don't compare the color, as stated, I did not white balance correctly. After using both bodies for a couple days, the color is the same.

 

[CColton]

2011 bodies have very diffirent firmwares
and depsite some peoples words - I can say - that firmware are with much bigger changes than simple adding of antihack check points.
diffirencies
1.34(was unhackable)-1.32(main publicity used)
are bigger than
1.33(very last available firmware, distributed by panasonic france. but even that release already has color schemes changed)-1.32.

1.34 firmware still not leaked to public

Once you hack the camera, whether an old or new body, they now have identical firmware, so the age of the original firmware doesn't matter.

 

[havasuphoto]

It does make you wonder why Panasonic would go to all the trouble, and expense, to create Firmware 1.34, just to thwart a hack. That doesn't make sense. So, the question remains-what there anything new added in firmware version 1.34, that was not included in version 1.32??

 

[CColton]

It does make you wonder why Panasonic would go to all the trouble, and expense, to create Firmware 1.34, just to thwart a hack. That doesn't make sense. So, the question remains-what there anything new added in firmware version 1.34, that was not included in version 1.32??

Who knows.

 

[faultiersid]

Finally, my 2011 GH1 arrived today

Made a quick comparison of it and a June 2009 GH1. Here are some screengrabs which speak for themselves - I guess The Focus is a bit off, sry, was really a quick test but it shows what is important.
Btw - I have three 2009 GH1s and they have all the same amount of fpn. One of them is made in Japan (Gold edition), but unlike the rumors the fpn was as bad as the others. Thank God I tried the 2011 version


GH1 2009 ISO 400

GH1 2011 ISO 400 (still unhacked)

GH1 2009 ISO 1600

GH1 2011 ISO 1600 (now hacked to the same settings as the 2009 version)

 

[goodgoings]

faultiersid, did you use Emory's settings on your 2011 GH1?Do you mind telling us the first letters/numbers of the serial on the 2011 model? And was it from B&H or PD?

 

[CColton]

After further shooting, my new GH1 is noticeably better. I'm comfortable shooting 1250 ISO Creative Movie Mode with my Standard +2,+2,0,-2 film mode without a problem, 1600 has some noise but it looks like film grain to me.