GH2 Are all these new hacks and patches creating a placebo effect?

Dr. Dave,
I think the moral of the story is to go straight to what makes the MOST difference.
On my 9-foot projection screen - EVERYONE could see the gross pasty macro-blocking on dark foliage and sky-banding on outdoor shots. My non-video-pro friends noticed it straight away.
With the cbrandin 66mbps hack it was gone to the point where nobody reported noticing it. No big deal, problem solved, all done.
"Some" types of motion look more clear with GOP=1 while others are unaffected. Blowing fur or hair is a good example.

Your audio comparison is interesting, but misleading. I can record a vocalist on 3 different mics and no one can really discern "which" mic I used by listening to the end product...but as you well know- if I set up 3 different mics in the studio and have that vocalist do the same thing in each one- they are going to sound different. And we know that makes a difference in the end product characteristics. It's the same with photography. Following people's preferences can be illuminating, but the creator still needs to know their tools and know WHAT they are after.

Comparing cheap to expensive doesn't tell us if they got the engineering right, it only tells us if they used more expensive methods.
Your lens comparison just shows you found two very similar lenses.

I work in audio too - and I have a few secret weapons that are dirt cheap but they got the formula right...and I have a bunch of more expensive stuff that does not impress. Then again I also have some expensive stuff that just KILLS everything else because the designers got ALL of it right. (James Demeter, Anthony Demaria, etc...)

Lots of people tinker...some hit home runs when they tinker.....follow those people.
 
Last edited:
The problem is that lots of people just look at the footage as is. This codec is made to look good at first sight, so if your lighting is good and there are not too many motion it is going to look fine in motion. But just try too dig just a little in the shadows, or some blowing wind in the foliage and see the macro blocking showing instantly. So yes if you don't intend to do color correction, or film no too much motion or not challenging light it is ok. But if in 4 out of 5 situation like these it will make no difference but for the 1 out of 5!!!! So it depend on your use, if for any family, events conference etc. type of shots definitely no need of the hack or the 44 mbit will be much better, but if anyone wants to do any film-making, advance commercial type of shot, I can't see them not using the higher end gop1.

I posted a little comparison on the gh2 and Canon 7d dynamic range here http://personal-view.com/talks/discussion/1795/dynamic-range-test-between-gh2-and-7d#Item_1, beyond the comparison is the test when I push the shadows. Do you think that the stock gh2 would be able to stand that. So just with the hack you are getting at least 1 stop perhaps 2 stops more DR, depending on the noise level you can accept. Where the stock gh2 is clearly beaten by the Canon 7d flatter picture style DR, the hacked gh2 would at least be on par.
 
The test above also show the noise pattern of the hacked gh2. It is finely grained compared to the smeared macro-block of the stock footage.
 
The video is up.

So I finally finished the Color Profile and Hack vs. Not Hacked Shootout video. The link to my thoughts on the test cases are in the YouTube video below. However, that video alone should not be used to make a final conclusion since the examples in it have been re-compressed twice. It should be used as a guideline for what to look out for when you watch the original files as they came out of the camera.


You can download those files at the Vimeo link below. Please note that not all of the clips have been uploaded yet. I am still in the process of uploading them(Some of them are huge). If you don't see a sample that you need just let me know in the comments of the youtube video and I will try to get it uploaded.

http://vimeo.com/user442745/videos

Everything was shot under completely controlled conditions. I shot this in pure manual mode with Manual focus, a fixed shutter speed, and a fixed aperture.

The only in camera correction I made was that I increased the shutter speed 1/3 of a stop for Smooth since the camera meters that mode 1/3 of a stop more than any of the other modes.

The video shows the differences between the most popular color modes for video(Cinema, Smooth, and Nostalgic). In addition it shows the differences between unhacked 1.0 and 1.1 firmware and medium and extreme hacked 1.0 firmware.
 
Thanks for the test, I know it is quite some work to do that. But my conclusion are completely different. I will talk about only one point because I have my own test to back it up. It concern dynamic range. I did a test that I showed here http://personal-view.com/talks/discussion/1795/dynamic-range-test-between-gh2-and-7d#Item_1 on the gh2 dynamic range compared to the Canon 7d, but it also analyze the hacked gh2 DR overall. Look at this post where I push the shadows quite extensively http://personal-view.com/talks/discussion/1795/dynamic-range-test-between-gh2-and-7d#Item_12 it is a little bit extreme because of the tool I used in photoshop, but the noise and I would call that grain is very nice. The grainy type in motion compensate a lot of what you would define as posterisation and with a normal curve it looks very smooth. Now with unhacked footage all this would be a macro-block fest (Should try to do some example for people to see).

Yes if you look at the image like this it is just going to look nearly the same but boost up the shadows a bit and the hack will show a lot of shadow detail. In some ways it looks a bit RAW like in photography when you start to boost the shadows, it is like a miracle, what looked completely black has a ton of detail. It is not raw like but still quite impressive.
 
Danielvilliers said:
Yes if you look at the image like this it is just going to look nearly the same but boost up the shadows a bit and the hack will show a lot of shadow detail. In some ways it looks a bit RAW like in photography when you start to boost the shadows, it is like a miracle, what looked completely black has a ton of detail. It is not raw like but still quite impressive.
Click to expand...

I don't believe in miracles. I get what you are saying even though I disagree with it. Can you download the original files for either V1.0 ISO 3200 or V1.1 ISO 3200 and compare them to the hacked ISO 3200 versions? Trying pushing each file in post processing and post your results. Until you have 1:1 samples of the exact same scene you can't really make a conclusion. Did you test Cinema vs. Smooth and Hacked vs. unhacked in your tests? It wasn't clear exactly what was tested.

I couldn't get the pictures to work in your post. It just displayed the thumbnails and not the originals. I must be doing something wrong.
 
mpgxsvcd said:
I don't believe in miracles. I get what you are saying even though I disagree with it. Can you download the original files for either V1.0 ISO 3200 or V1.1 ISO 3200 and compare them to the hacked ISO 3200 versions? Trying pushing each file in post processing and post your results. Until you have 1:1 samples of the exact same scene you can't really make a conclusion. Did you test Cinema vs. Smooth and Hacked vs. unhacked in your tests? It wasn't clear exactly what was tested.

I couldn't get the pictures to work in your post. It just displayed the thumbnails and not the originals. I must be doing something wrong.
Click to expand...

The test I done was not for hacked vs unhacked, but as I was showing how there was so much more detail in the shadows of the hacked gh2, I thought that anyone, who has tried to do it on the unhacked gh2 would clearly see the difference. Now I think the link that kholi posted above clearly shows it from your own file. If need be I can a hacked and unhacked gh2 test like this and pushing the shadows, but for me I thought it was mundane because I thought that there have been numerous example where it was showed that the second you just boosted a little bit the shadows the macro-bloking was evident. The reason I did a 160 Iso one is simple because it is normally where you have the highest DR and cleanest noise.

For me miracles do happen. When you use raw files in photography just adjusting some slide for example to recover highlight, sometime things that look completely gone are recoverable, the blown sky just comes back like magic. I think that our human eye sometime does not make for those very very subtle nuance. Look here for example http://www.dpreview.com/reviews/nikond700/page20.asp and the same for the shadows. Now the gh2 files we are working with are not RAW, but a lot of info that we just don't see is there mainly in the shadows. You should normally be able to down load my files in png on the personal-view site and play with them.
 
Last edited:
Danielvilliers said:
The test I done was not for hacked vs unhacked, but as I was showing how there was so much more detail in the shadows of the hacked gh2, I thought that anyone, who has tried to do it on the unhacked gh2 would clearly see the difference. Now I think the link that kholi posted above clearly shows it from your own file. If need be I can a hacked and unhacked gh2 test like this and pushing the shadows, but for me I thought it was mundane because I thought that there have been numerous example where it was showed that the second you just boosted a little bit the shadows the macro-bloking was evident. The reason I did a 160 Iso one is simple because it is normally where you have the highest DR and cleanest noise.
Click to expand...

I just realized you were Danyyyel from PV. Haha. I think your test validated a lot of findings, to me. It was very easy to see the difference, but what your test also showed me--and I didn't realize this, was how much macroblocking occurs in Canon footage. It may LOOK cleaner at first glance, but it's simply just blocky and smooth.
 
Yes like many here i have my dual personality in those two forums :-). My test came from the philip bloom shootout that was very misleading at first sight on the DR of the hacked gh2. I knew from my informal testing that the hacked gh2 had a ton more detail than the unhacked one and the Canon 7d, with those flat profiles. So I decided to put them in a controlled situation to show/prove it. I see so many assumption on the net, one being that the gh2 had limited DR compared to the Canons. Perhaps the stock gh2 yes, but the hacked one has gain at least 1 perhaps 2 more stop of DR in the shadows and has put it on par at least to the Canons. More so, that in my test the highlight seems to have more steps in the roll-off. Perhaps I should have push my research a little more with the stock gh2 to show how the hack indirectly (I say indirectly because the DR is the same, it is just the compression that is removing the detail in the shadows) contributed to a higher DR camera.

Now I think that with mpgxsvcd test, this matter will be put to rest. I should put my test here also, but the fact that the image are re-sized and only jpeg is a bit annoying. To summaries, the proof with the hacked gh2 is in the detail and I would say fine detail. Having worked a lot with RAW files as a photographer, I know that a lot is buried in the shadows and some digital camera the highlight also. So that you should not rely on the perceived image, with a clean signal you can retrieve much more in general. At least try to do it, but just don't just rely on what you see.

Where I wanted to further my test is in the number of stops or perceived one we could underexpose to get good recoverable shadows. The challenge will be on skin tone/human beings because scenes or object can accept more noise or degraded color than the background furniture in my test for example.
 
Last edited:
Thanks for doing the test mpgxsvcd.:cheesy: Grabbed some of his files - the stock and 187 mbps files and did a quick test for those who care. Please download the originals and view on a full HD screen for best comparison.






All the best - Seasons greetings.
 
Danielvilliers said:
Yes like many here i have my dual personality in those two forums :-). My test came from the philip bloom shootout that was very misleading at first sight on the DR of the hacked gh2. I knew from my informal testing that the hacked gh2 had a ton more detail than the unhacked one and the Canon 7d, with those flat profiles. So I decided to put them in a controlled situation to show/prove it. I see so many assumption on the net, one being that the gh2 had limited DR compared to the Canons. Perhaps the stock gh2 yes, but the hacked one has gain at least 1 to stop more DR in the shadows and has put it on par at least to the Canons. More so that in my test the highlight seems to have more steps in the roll-off. Perhaps I should have push my research a little more with the stock gh2 to show how the hack indirectly (I say indirectly because the DR is the same, it is just the compression that is removing the detail in the shadows) contributed to a higher DR camera.
Click to expand...

People have to realize that DR can be defined in a technical sense, and then in the real working world those numbers change... usually for the worse. I would prefer a camera with 9-10 stops of usable DR in the harshest production scenarios or the most perfect, over a camera that delivers 11-12 stops in the best scenarios but only 8 in the worse. Point being, I know that I can rely on camera A to give me the results I need in every scenario. Balance.

The Gh2 Hack's noise pattern is what changes the claim of DR to me, where usable DR is the only true factor in the finishing process of motion picture. Yes, I agree: the compression of the Canons is removing tons of shadow detail, and so the same for the unhacked GH2.

Now I think that with mpgxsvcd test this matter will be put to rest. I should put my test here also, but the fact that the image are re-sized and only jpeg is a bit annoying. To summarise the proof with the hacked gh2 is in the detail and I would say fine detail. Having worked a lot with RAW files as a photographer, I know that a lot is buried in the shadows and some digital camera the highlight also. So that you should not rely on the perceived image, with a clean signal you can retrieve much more in general. At least try to do it, but just don't just rely on what you see.
Click to expand...

Your tests are relevant, even more so if you do them again and offer some PNGs or TGAs. Either way, don't worry so much about it.

Where I wanted to further my test is in the number of stops or perceived one we could underexpose to get good recoverable shadows. The challenge will be on skin tone/human beings because scenes or object can accept more noise or degraded color than the background furniture in my test for example.
Click to expand...

Skin/human is the most important thing. That should be the baseline for most narrative filmmakers, I believe, but mileage may vary.

I know you've been asking me to grab the Epic and shoot some comparisons to GH2 footage. I don't think there's any true reason to compare Epic's 5K and 4K to the GH2, there's no contest what-so-ever. Compounding the raw workflow over that, well, you know the story. And while I am very comfortable with saying that Epic's 2K 3:1 mode and GH2 AVCHD High-bitrate's a damned close competition, you don't buy an Epic to shoot 2K, you shoot 5K. Same thing with Scarlet, only you're shooting 4K.

Both will downconvert to 1080/2K and look better (technically) than GH2 Hack Footage, or footage originating in 2K for many reasons that isn't just related to resolved detail.

That's all to say that you have a seven-hundred dollar camera that outperforms large-sensor cameras in the 5K-15K range as far as image is concerned. You're getting away with murder, so go and kill something; preferably a decent script. =D
 
Kholi said:
I don't think there's any true reason to compare Epic's 5K and 4K to the GH2, there's no contest what-so-ever. Compounding the raw workflow over that, well, you know the story. And while I am very comfortable with saying that Epic's 2K 3:1 mode and GH2 AVCHD High-bitrate's a damned close competition, you don't buy an Epic to shoot 2K, you shoot 5K. Same thing with Scarlet, only you're shooting 4K.

Both will downconvert to 1080/2K and look better (technically) than GH2 Hack Footage, or footage originating in 2K for many reasons that isn't just related to resolved detail.
Click to expand...
I'm looking forward to trying Scarlet's 3K mode, downsampling to 1080p to intercut with the GH2. At 3K resolution, the Scarlet has a 2X crop factor, very close to the GH2's field of view. Can't wait to try out my Nikon AF-D's on that camera!
 
DPStewart said:
Your audio comparison is interesting, but misleading. I can record a vocalist on 3 different mics and no one can really discern "which" mic I used by listening to the end product...but as you well know- if I set up 3 different mics in the studio and have that vocalist do the same thing in each one- they are going to sound different.
Click to expand...
If I put my Coles 4038 and my Sennheiser MKH 800 on a mic stand and adjust the patterns so they match I will pick them out between 95 and 100 times out of 100. Same with a Beyer M160 and a Schoeps MK41. Using headphones, however (audio peeping). With speakers it would be tougher. An untrained listener would find this test very challenging. And it is possible that we can train our eyes to reliably pick small differences in Video as well, better than the average viewer. In the PB experiment, which most people failed miserably, PB suggested that those who had identified some of the cameras were trained in such a way, doing color for a living. However, the best of these (one person) was 50 percent correct, that is, 5 out of 10. And 5 out of 10 is, well, half.

It would be really fun to test the Vixia G10 ($1200) against the Vixia M41 (way cheaper). I bet I couldn't pass that test either. I also bet I couldn't tell the M41 from the GH2.

In your example about the pasty macro-blocking, I have no reason to doubt this, and it should be very easy therefore to tell the difference in a double blind test. It won't answer the question of why we can't do better in the PB experiments. I suspect if PB, as he has indicated, releases more footy without labels, people will chime in (only some people) and we will know much more.

Realistically, all I'm saying is that until someone repeats this sort of test--under lab conditions--and people repeatedly provide the correct answers, it is logical to assume that we can't tell the difference.
But I look forward to more tests!

As far as the Vivitar and the Olly being closely matched, they aren't. The aren't even the same focal length. That means I can't tell a 45mm from a 55mm, yikes. The curvature of the lens (which affects the way we see objects is vastly different. Coatings are different. They are very different lenses in all respects--that's why it is so annoying that I (or 200 other people) can't tell then apart. My bad.
 
Kholi said:
My advice to those making statements like these is to make sure you're not being blatantly hypocritical. Practice what you preach, and make sure that you've got equal amounts of evidence (hopefully in your body of work, that you can show people) to back it up. Once you cross the line into passively insulting people, you should be ready to back that up.
Click to expand...


OK, it was a snipe - and, upon reflection, an unworthy snipe. I suppose I was miffed because you moved my Nex-7 thread - and, as this is the season of goodwill, I apologise unreservedly for sniping.

To be honest, I really don't know why it bothered me. Of all the wories and concerns in my life, cameras come at number 780937... ;) What *is* it about these gadgets that makes us so insanely tribal?...

As far as my opinion about a 'placebo effect' goes - well, I was being honest, I do think that there is an awful lot of 'seeing what you want to see' in these things.

But that doesn't matter - people are free to see whatever they want to, and I was ungracious in deriding the phenomenon.

All I can say is that having recieved my replacement GH2, I am bowled over by the stock high ISO performance. I've just taken some shots at my Grandson's 3rd birthday, and the footage in a room completely dark except for three birthday candles, is quite amazing (for the price of the camera and the inneptitude of the operator ;)

VK's hack has some really valuable features - the removal of the 30 minute European time limit, for one. There is also no doubt that increased bit-rates help the party go with a swing - but, after those things?.......

Some might find a benefit in endless tinkering with the settings - but, to be honest, when all this stuff is rendered I'd be hard pushed to tell the difference bewteen most of the consumers cams now on offer.

(and if I can dare to say it - I still want a Nex 7 -;) )

Anyway, that's my last word on the subject - peace and prosperity for you in 2012, and footage that makes you happy.

Regards

F
 
Framerate said:
(and if I can dare to say it - I still want a Nex 7 -;) )

Anyway, that's my last word on the subject - peace and prosperity for you in 2012, and footage that makes you happy.

Regards

F
Click to expand...

Everyone's opinion is welcomed, even if it turns into debate. The issue with the thread you started was that it was seen as flame baiting, not just stating opinion. All's well and good.

But, for reference, I've been at the DSLR stuff first out of the gate. That means when the D90 dropped I was at Best Buy getting it before it was even released, and when the GH1 started selling in Japan I had one on a plane with friends the next day headed back to LA.

I'm returning because I saw something in the GH2 Hack that intrigued me enough to sway from RED. I'm not a hack afficianado by any means, but I do hope to offer experiences with the camera putting it on as many serious gigs as possible and reporting on it.

It's always good in my book, no harm done. I'm not exactly innocent, either, so.

Good 2012 and Shoot It Up.
 
Kholi said:
Good 2012 and Shoot It Up.
Click to expand...


Easier said than done, Sir!

I only do video as a hobby - Steam Fairs, family occasions, that sort of thing.

When I had my little HG10 all was serene - I'd produce the camera as we walked around, shoot some footage, pop it away, and all was well.

However, as I added to my GH1/2 colection - rig, handles, a external recorder, external monitor, etc, the family became ever more reluctant to be seen in my company ;)

The final straw was the Matte Box - which may have broken the proverbial camel's back, but most definitely not it's spirit!

Simmering discontent has become open rebellion - "You're *not* coming around with us with *that* thing...."

Oh, I've tried to explain what the matte box is for - tried to tell them about the necessity of using ND filters to keep the shutter speed at 50/1000 - but they just don't care! ;)

Bottom line? - they think I'm walking around like Stephen Speilberg, and they think it's highly peculiar.

Perhaps it is ;).....
 
You must log in or register to reply here.
Back
Top