Re: PTool 3.63 and the GH2

Yes, its a hack. Yes, it takes some trial, error, and fiddling, and you have to be choosy about what cards you use.

Round one of the hack did some nice things, but really wasn't what I considered a "necessary" modification. It somewhat cleaned up higher ISO's and made footage a little more color correction tolerant.

With the newest version's ISO liberation, that's it - if you shoot with this camera in anything approaching sub-optimal lighting and DON'T hack it, you're hamstringing yourself.

For your consideration:
Gh2 Color footage 10.000 ASA ISO test on Vimeo

Note: not my footage, but I've seen similar results - outside, at night, in my basement, during concerts - it's ridiculously good.

Remarkable footage. Would be nice to know what lens was used.

Lens used was Voigtländer 25mm/F0.95

see: Gh2 Color footage 10.000 ASA ISO test - Panasonic GH2 on Vimeo (http://vimeo.com/groups/gh2/videos/32204670)

even so, that ~1 stop difference from something like the Summilux DG 25 would get you about the same.

Got a GH2 this weekend, loaded the "Seaquake" intra hack, and started playing around with it last night in moderate incandescent lighting. I shot with the iso at 2500 using the cheap kit lens at f3.5. The quality of the clips was unlike anything I have ever seen from the XF100, XHA1, or 5DII. Excellent resolution and motion representation and very very fine film like grain. I'm not selling my XF100, but the GH2 will now be the go to camera in cases where I have the time to set up the shot or want the highest image quality I can get. Even overexposed shots had a beautiful high key look to them.

Pat

shot with the iso at 2500 using the cheap kit lens at f3.5. The quality of the clips was unlike anything I have ever seen from the XF100, XHA1, or 5DII.

Pat

hi Pat, that's quite a statement. do you really feel the image from the GH2 with a cheap lens delivers better than a 5d mk2 even with a good lens (like an L lens)? I own a 5Dmk2 and a Canon HFg10 and, although I love the image quality, I am almost ready to pull the trigger on the GH2, and posts like yours are providing me with more reasons to do it.

can you compare more specifically your experience with the 5Dmk2 vs the GH2?
thanks

As an event shooter who shoots in low light the hack would be tempting at first glance. But we have to keep in mind it was shot at 42mbps, a F/0.95 lens was used, noise reduction was used in post, there is no comparison footage, and most importantly the hack is not reliable and cannot be depended upon.

One would have to find a card/camera combo that works every time. Is that even possible? Too many stories of bad or corrupted files when using hacks for me to consider it.

Regarding the lens, as the shooter says focusing was very difficult, which is typical with an extremely DOF. The Voightlander is a fine lens, to be sure, and if I had one I would likely enjoy using it, but for event shooting, for me, it is not close to practical. Manual focusing with such a lens would take more time than I have when recording a wedding or reception.

This is a case of pushing things to the extreme, kind of like modding a street car with jet fuel. It will go fast, but it is not practical or even desireable to use as a taxi.

I'm dealing with a wedding with which my editor has lost some original footage, making a re-edit impossible. I am going through hell right now because of the emotional impact this will have on my client, a repeat client who I will lose forever because of lost footage. I mention this because there are enough variables and stresses involved with event shooting that to add the uncertainty of a hack to the equation is not an option for me.

Federico,

With the highest bit rate hacks, the GH2 looks way better than my 5d2 will L lenses. I am using a fast class 10 card, but there are occasional hickups. Also I have found that the GH2 in 24p mode has pulsing (not the same as normal 24p judder) on high-detail elements like grass and tree branches. This is less apparent on indoor shots, and shots where you have narrow or shallow DOF. This pulsing looks just like what I saw on Panasonic's TM700 until I turned in-camera sharpening down. I think Jeff observed this pulsing in some of his recent shots. Turning sharpening all the way down (-2) in the GH2 doesn't get rid of it.

I haven't tried the 42 and 44 mbs hacks, but the higher bit rate hacks occasionally cause card errors of shorts. There are people using the various forms of hacks for pro work, but it is a bit of an adventure, and you would have to be willing to live with or work around these issues.

Some people are getting rid of the pulsing by softening the focus with filters. The tiniest amount of gaussian blur in Vegas helps. Shooting with narrow DOF helps greatly and I never saw problems when shooting this way (nor do I see it in the sample footage discussed here).

Are the hacks good enough for pro work, events, run and gun? I think it depends on how much you are willing to deal with the issues, test, find solutions, and practice.

All I can say for sure is that low light shallow DOF shots at home a few days ago were the best I have ever seen from any of my cameras (XF100, 5D2 with L glass, and XH-A1)

Pat

Consider that while the ISO deregulation and reduced noise at increased bitrates allows you to shoot with a fast lens in minimal light, it also lets you use a slow lens (~f4-5.6) in reduced indoor light.

I do agree that hacking your camera and then immediately using it on an ultra-long-take paying job is likely not the best idea. With a stageplay shoot coming up, I'm pre-testing by just setting up the camera at home with a spanning hack and letting it run for 2-3 straight hours. So far, there's been no problem.

Kevin, when I'm done editing my currentl backlog, I am likely to try it to see the results. If things were to work out and prove stable enough, it would be an incredible advantage to have.

GH2 test vs 5D2 at wedding.

GH2 at ISO 3200, Aperture 2.8 14mm lens. Wide shot
5D2 at ISO 1600, 50mm lens. Medium shot.
Audio from original microphone build in with GH2 Level 1.
Footage is direct from original cameras (no hack) no color correction.


GH2 Test on Vimeo

Thanks for posting. What was the lens used on the Canon?

Thanks for posting. What was the lens used on the Canon?

Prime lens 50mm on Canon 5D.

1.4 or 1.8?

1.4 or 1.8?

My lens is 1.2L :)

GH2 test vs 5D2 at wedding.

GH2 at ISO 3200, Aperture 2.8 14mm lens. Wide shot
5D2 at ISO 1600, 50mm lens. Medium shot.
Audio from original microphone build in with GH2 Level 1.
Footage is direct from original cameras (no hack) no color correction.


GH2 Test on Vimeo (http://vimeo.com/32472017)

So you didn't hack the gh2? and still got those shots at 3200ISO?

So you didn't hack the gh2? and still got those shots at 3200ISO?

Yes. I just bought the camera 2 hour before I went to shot the wedding last Saturday. The raw you see in this clip is original from GH2 setting at 3200 ISO with 16MB/s. And I did hack my GH2 yesterday. Now it can record 44MB/s but not good for using Sandisk 32GB class 10 30MB/s (give you the message) and then I down to 35MB/s still using the same card, press record for over an hour. So far everything is perfect :)

Re: PTool 3.63 and the GH2

Yes, its a hack. Yes, it takes some trial, error, and fiddling, and you have to be choosy about what cards you use.

Round one of the hack did some nice things, but really wasn't what I considered a "necessary" modification. It somewhat cleaned up higher ISO's and made footage a little more color correction tolerant.

With the newest version's ISO liberation, that's it - if you shoot with this camera in anything approaching sub-optimal lighting and DON'T hack it, you're hamstringing yourself.

For your consideration:
Gh2 Color footage 10.000 ASA ISO test on Vimeo (http://vimeo.com/32204670)

Note: not my footage, but I've seen similar results - outside, at night, in my basement, during concerts - it's ridiculously good.

do you have a settings file that I can look at and test with? Im trying this out over the winter break

I am always quite happy to use ISO3200 on the GH2 as the high ISOs on the GH2 are remarkably free of noise but the shot needs to be properly exposed. If you use high ISOs & it is well lit with adequate aperture then it will look fine but if it is underexposed & you try to lift it in post then it will look terrible.

How would that work on shots using a slow telephoto like the 100-300 F4-5.6 ?
Could I push the ISO, shoot fast shutter and open up to 5.6 on the long end of the zoom
I'm thinking surfing stills and video in bright morning light.

Federico,

With the highest bit rate hacks, the GH2 looks way better than my 5d2 will L lenses. I am using a fast class 10 card, but there are occasional hickups. Also I have found that the GH2 in 24p mode has pulsing (not the same as normal 24p judder) on high-detail elements like grass and tree branches. This is less apparent on indoor shots, and shots where you have narrow or shallow DOF. This pulsing looks just like what I saw on Panasonic's TM700 until I turned in-camera sharpening down. I think Jeff observed this pulsing in some of his recent shots. Turning sharpening all the way down (-2) in the GH2 doesn't get rid of it.

I haven't tried the 42 and 44 mbs hacks, but the higher bit rate hacks occasionally cause card errors of shorts. There are people using the various forms of hacks for pro work, but it is a bit of an adventure, and you would have to be willing to live with or work around these issues.

Some people are getting rid of the pulsing by softening the focus with filters. The tiniest amount of gaussian blur in Vegas helps. Shooting with narrow DOF helps greatly and I never saw problems when shooting this way (nor do I see it in the sample footage discussed here).

Are the hacks good enough for pro work, events, run and gun? I think it depends on how much you are willing to deal with the issues, test, find solutions, and practice.

All I can say for sure is that low light shallow DOF shots at home a few days ago were the best I have ever seen from any of my cameras (XF100, 5D2 with L glass, and XH-A1)

Pat


I have to update what I said about the pulsing issue. I discovered that most of this goes away if you make sure that IR (intelligent resolution?) is off and sharpening is set to -2 in whatever film mode or preset you are using. With these appropriate settings, 24p from a hacked GH2 is beautiful.

Nick Driftwood and the hack community have a nice showreel that may give you an idea what the GH2 can do:

The Personal View 'Driftwood INTRA' Showreel 1 (ungraded version) - Panasonic GH2 on Vimeo

Pat

Maybe I have really low standards or am just going blind in my old age. I played around a little with the hacks this weekend. I only experimented with the ones for AVCHD in 1080/60i since that is what I shoot in and don't care for the 24p.

I loaded the popular 42mbps hack for the 60i. I did a few tests shooting my computer screen with dark areas around it in low light. I could see no improvement with the 42mbs hack over the one I created to bump the 60i up to 24mbs. I tried the Driftwood SPANmyBITCHup settings as someone told me it was the cleanest that still spanned on long files properly. It was a pretty high bit rate and big file size. Didn't look any better to me than my basic 24mbps hack I made. The file sizes were about 3 times the 24mbps rate and I didn't see any improvement in handling low light artifacting. Maybe it would be visible in other types of shooting but for me the biggest thing I want is improvement in the quality of low light footage.

For now the 24mbps in 60i is what I keep going back to and I still get 1.5 hours on a 16gb card.

I did shoot some video at the 10K ISO with my 20mm/1.7 and overall it was acceptable if I was put in a situation where I needed it. I don't have access to NeatVideo but it was still decent compared to what I expected. My third GH2 arrives anytime now and along with it will be the 14mm/2.5 and the 45mm/1.8 lens. That should be good enough to be decent in low light for my work and only need about 1600-2000 ISO on bad shooting days.

Phillip, if you need lowlight footage look at the 25mm F/1.4. It really is a huge upgrade over the 20mm f/1.7.

With the 12mm F/2.0, the 45 F/1.8 and the 25 F/2.0 your basically done shopping for lenses, minus your zoom lenses.

Philip -

What kind of scene were you shooting when you tested the hacks/higher bitrates?

I haven't yet personally tried any of the hacks on my GH2. I'm waiting to see what the new official firmware that's been announced for December has to offer. (I hope it won't be much longer now.)

But the type of scene matters highly when evaluating compression artifacts.

A static scene, or one with an uncomplicated background won't generate much difference between frames, and therefore a low bitrate will be sufficient to capture all the detail.

But highly complex, detailed scenes (leafy trees, for example) or fast motion, and especially both combined, use up all the available bandwidth digitally describing the changes between frames, causing all kinds of artifacts from blocking, loss of detail, uneven motion, etc. Its reducing those types of motion artifacts which attracts people to the hacks.

People who primarily shoot interviews or static scenes probably won't notice or need the benefit of the hacks.

Also keep in mind that each generation of compression/recompression introduces new artifacts. Going from AVCHD to ProRes, then exporting the final product as H264 or MPEG2 can reveal some of the weaknesses, although certainly the all-digital, nonlinear workflows we have today hold up far better than the generation losses suffered to projects for those of us who started in the 3/4" tape era. :-)

Yea, that 25/1.4 looked real tempting when I was ordering everything but unfortunately my Canon A1 cams didn't quite sell for enough to allow me to spring for the extra money. I had to stay within those numbers. Now when I shoot a few jobs with the setup I have I may find I don't need that second 14-140 kit lens and could sell it and one of the primes I did get and get the 25/1.4. I know that 1.4 lens is sweet but I still have to ask people using them how they old focus. All my buddies shooting Canon's brag about their 50mm 1.4 lenses and how they can see in the dark with their 5D bodies. And I always ask how they hold focus if someone moves a step the wrong way. Just seems it would be tough to hold focus. Now granted we GH2 users are a little superior cause we have that nifty autofocus feature to help us along. :-).

For now I have the 20/1.7, 14/2.5 and the 45/1.8 plus the 14-42 and two 14-140 kit lenses. Add into that my old manual focus lenses with the adapters from my T2i and I also have a 28/2.8, a 58/2.0 and a 70-210/3.5. That's a pretty decent setup to run between 3 cams....but you are right that 25/1.4 would probably seal the deal. I just need a few jobs first to see how this arsenal works. I read alot of threads around here narrowing it down to that setup trying to get the most bang for my buck.

For now I'm just going to enjoy the great footage I can get with what I have even if I have to bump the ISO to 1600 or 3200...it's sooooo much better than what I am used to seeing from my Canon A1 cams.

Philip -

What kind of scene were you shooting when you tested the hacks/higher bitrates?

I haven't yet personally tried any of the hacks on my GH2. I'm waiting to see what the new official firmware that's been announced for December has to offer. (I hope it won't be much longer now.)

But the type of scene matters highly when evaluating compression artifacts.

A static scene, or one with an uncomplicated background won't generate much difference between frames, and therefore a low bitrate will be sufficient to capture all the detail.

But highly complex, detailed scenes (leafy trees, for example) or fast motion, and especially both combined, use up all the available bandwidth digitally describing the changes between frames, causing all kinds of artifacts from blocking, loss of detail, uneven motion, etc. Its reducing those types of motion artifacts which attracts people to the hacks.

People who primarily shoot interviews or static scenes probably won't notice or need the benefit of the hacks.

Also keep in mind that each generation of compression/recompression introduces new artifacts. Going from AVCHD to ProRes, then exporting the final product as H264 or MPEG2 can reveal some of the weaknesses, although certainly the all-digital, nonlinear workflows we have today hold up far better than the generation losses suffered to projects for those of us who started in the 3/4" tape era. :-)

Bob, my testing was very unscientific. I just pointed the cams at the bottom of my monitor with the room lights low at night. I exposed pretty well for the monitor screen and then checked the black plastic around the monitor and also the surrounding light objects cluttering up my desk of which there was plenty. I then stacked them on the timeline and toggled between the different clips at different bit rates. Nothing fancy or moving. I did try and set the cams up with the same aperture/shutter/ISO to keep it as common as possible since I had different lenses on the cams. I found a huge improvement from the stock 17mbps to the 24mbps I bumped it up to. I also bumped it up to 32 and 42mbps and really didn't see a noticable improvement in the lighter colored objects or things around my screen. I know that is not where the higher rates are really exploited but I figured it should be some kind of a difference. The 24 looked as good as any.

One challenge for me is I am not looking to fill up my 16gb cards real fast. The last hack I tried would have filled up my 16gb card in no time. So far I have found the 24mbps hack to be a nice compromise between quality and storage capacity. I may try the 32 mbps hack again outside in my yard one night and see what I get. That would be a better test. I was thinking for the weddings I shoot the 24 would be good for most stuff and if I needed to I could shoot 32 at the reception to help it out a little in quality. I could just setup FSH at 32 and FH at 24. Shoot at FH most of the day and flip to FSH when it gets dark. If the 32 is significantly improved I may go that route. Guess I need a better, more scientific test. Guess I wouldn't be good at Mythbusters.....just not enough science there.

I have still not wrapped my brain around the whole Intra stuff and the settings. I know Intra is supposed to be better but everytime I try and grasp it I just get confused. :-(

If you shoot weddings primarily, you'll ditch the 14-140mm pretty quick. I had mine maybe a week after the first time I used it, and then I couldn't sell it fast enough. Same with the 14-42, pretty much worthless for virtually all wedding work, except outdoors. I've gone through about a dozen lenses, and most all of them are long gone at this point.

Philip,

You probably already know this, but the bit rates in the hacks are often set lower for 60i than they are for 24p. So you might be looking at an even lower bit rate than you think. I have not liked the hacked 60i. I think the 24p really shows off the improvements offered by the higher bit rates and intra compression. Maybe the new firmware will give us 30p.

Pat

I would consider 24p if there was an overcrank mode to 48fps so I could do some slo motion. I don't do much as a rule but there are times I want to use it for a few clips and 24p sucks big ones for slo mo. That's the main reason I don't do 24p....plus I don't really care for the aesthetic of 24p on shots that have movement. That's my personal feeling but I know others really really love it. If I could overcrank it like you can in 720/60 I would consider it.

The new FW does have 30p at 24mbps wrapped in 60i. I've tried it out already and still prefer my hacked 24mbps setup I created in PTools but that's me and my tests were very unscientific. I may have to go run more tests. Yea, that means I can test and not work...always more fun that way. :-)