GH2 sensor goes into impressive Panasonic G6, adds 1080/60p and focus peaking!

[Sean Cunningham]

The encoding on those YT streams is absolute garbage.  They must throttle bandwidth and encoding quality depending on the account type because they're not even at the same quality as several YT "partners" that I semi-regularly check out there.   That or the encoding done by the uploader was very bad as well.  

 

I honestly don't understand why people still use it in this capacity.  I'll add duplicate streams of stuff I put on VIMEO, acknowledging the ubiquity of YT but I don't actually share those links with people.  

 

In the case of both sites whatever you upload gets further compressed (heavily) so your upload needs to be of exceedingly high quality if you expect anything decent by the time a stream goes live.

[Ernesto Mantaras]

The encoding on those YT streams is absolute garbage.  They must throttle bandwidth and encoding quality depending on the account type because they're not even at the same quality as several YT "partners" that I semi-regularly check out there.   That or the encoding done by the uploader was very bad as well.  

 

I honestly don't understand why people still use it in this capacity.  I'll add duplicate streams of stuff I put on VIMEO, acknowledging the ubiquity of YT but I don't actually share those links with people.  

 

In the case of both sites whatever you upload gets further compressed (heavily) so your upload needs to be of exceedingly high quality if you expect anything decent by the time a stream goes live.

 

That is absolutely correct, but just to make it double clear: while YouTube compression can bring a bazzillion types of artifacts it will not render your perfect full HD resolution, carefully encoded video into something this soft: http://www.youtube.com/watch?v=3oijsASQQ6k&feature=share&list=UUxSt5kmZG3dNXsBJqtzIKmQ (or, in other words, it cannot turn GH2 worthy footage into that low resolution video)

And if there are differences made between partners and other regular channels, then those aren't to blame either, because you can find perfectly detailed videos like this one in the same PanasonicLumixVideo channel: http://www.youtube.com/watch?v=MT-2VF3-fWo

[nahua]

That is absolutely correct, but just to make it double clear: while YouTube compression can bring a bazzillion types of artifacts it will not render your perfect full HD resolution, carefully encoded video into something this soft: (or, in other words, it cannot turn GH2 worthy footage into that low resolution video)

And if there are differences made between partners and other regular channels, then those aren't to blame either, because you can find perfectly detailed videos like this one in the same PanasonicLumixVideo channel: http://www.youtube.com/watch?v=MT-2VF3-fWo

Judging by the video link I think this is shot in 60P Auto mode.  But it does look like the codec may be an issue.  I think the detail is there, but the codec (AGAIN) is holding it back.  This is strange considering the AVCHD on the GH3 has excellent detail.  Or, they recorded 60P with mp4 with is considerably worse than AVCHD.  We won't know until we get the actual files.  Just keep up hope, Panasonic hasn't let us down just yet.

[espertype]

Panasonic are listening that we all say the gh2 sensor has some special magic about it!

 

That sounds weird. If the sensor's "character" (is that what you mean by "special magic"?) is so desirable to warrant it being used again, then why is its multi-aspect-ratio potential – being a characteristic of this sensor that is appreciated by many just like its readout speed and low-light performance – left unused at the same time?

I understand that you are probably only referring to the look the sensor produces and not to what format or image circle it covers. But both are features belonging to the same component and if one is dropped (or, better to say, disabled), that is sort of disappointing and lowers the appeal of the remaining features.

 

It seems as if they want to keep this feature out of the mid-range and entry-level cameras so vehemently that even G cameras of a later generation than the GH camera originally using it aren't allowed to have this feature. What a tease – and a good reason to not praise the use of the GH2 sensor.

 

I have no problem with new camera models using a slightly dated, but proven to be good sensor. I like the G6 overall and might end up buying it to step up from my GF2 and get a little more serious about video (unless there is a GX2 announcement coming in the meantime that is more convincing). But locking out one of the sensor's nice features in the firmware leaves a bad taste.

 

The only good thing I can see is that it could be a hint for a future GH model having a (working) multi-aspect ratio feature again, hence why they continue to keep this feature out of the GF, GX and G lines.

[P337]

@Andrew may be in the best situation to assess this since he just finished the DPReviews' video section of the GH3 where he compared it to the GH2 but I leave this question open to anyones speculations.

 

Since no one who actually had hands on this camera did any proper video tests (DPReview should have send EOSHD lol) we are forced to estimate quality until those tests are available but if we assume it has a sensor just as good as the GH2/G5 and a processor as good as the GH3 which do you expect the video quality to be closer too?  GH3 or GH2? 

 

Specifically my worries are going back to the noise levels of the GH2 , seeing the banding return and all this stuff:

 

A major area of improvement over the GH2 is with under exposed areas of an image. It’s important to be clear what this means. This doesn’t just cover shadows in a night shot at ISO 3200. We’re talking gardens in the shade at ISO 200. Almost any shot has an area of shadow where the light is dimmer than it is compared to the rest of the scene. Anything that is exposed to the left. Usually you don’t notice much noise in these under exposed areas when you crush the shadows, but if you’ve underexposed to protect a large highlight area like a window from blowing out and you want to correct the underexposed interior, here the GH3 does much better than the GH2. Better colour, smoother gradation, more contrast, far less noise. It is just easier to work with in post. -http://www.eoshd.com/content/9920/gh3-review

 

Let my put this question another way; do you think the issues of the GH2 compared to the GH3 mostly lie in the sensor or the processor of the camera?

 

Aside from that wouldn't an increase of bit-rate solve everything?  Since they both use h.264 in (I assume) IPB, does the GH2 at 50Mbps equal the GH3 at 50Mbps?  And isn't there an Intra (ALL-I) hack for the GH2 as well?

[andy lee]

GH2 hacked is the best non raw camera on the market - the level of detail that sensor creates is stunning

Mine are kicking out 170mb/s and I cannot stress how good the image is

so yes the GH2 sensor does indeed have a magic

it is in a differnt league to the GF and G series cameras

now we have a G6 with this sensor too....

[andy lee]

That sounds weird. If the sensor's "character" (is that what you mean by "special magic"?) is so desirable to warrant it being used again, then why is its multi-aspect-ratio potential – being a characteristic of this sensor that is appreciated by many just like its readout speed and low-light performance – left unused at the same time?

I understand that you are probably only referring to the look the sensor produces and not to what format or image circle it covers. But both are features belonging to the same component and if one is dropped (or, better to say, disabled), that is sort of disappointing and lowers the appeal of the remaining features.

 

It seems as if they want to keep this feature out of the mid-range and entry-level cameras so vehemently that even G cameras of a later generation than the GH camera originally using it aren't allowed to have this feature. What a tease – and a good reason to not praise the use of the GH2 sensor.

 

I have no problem with new camera models using a slightly dated, but proven to be good sensor. I like the G6 overall and might end up buying it to step up from my GF2 and get a little more serious about video (unless there is a GX2 announcement coming in the meantime that is more convincing). But locking out one of the sensor's nice features in the firmware leaves a bad taste.

 

The only good thing I can see is that it could be a hint for a future GH model having a (working) multi-aspect ratio feature again, hence why they continue to keep this feature out of the GF, GX and G lines.

if you are coming from a GF2 it is no way near as good an image as a hacked GH2 -

I just hacked a GF3 to 100mb/s with a Driftwood hack as a crash camera to use on car rigs and the image is only 85% as good as GH2

its soft and not as much detail.

[Andrew Reid]

I've never seen a single YouTube 1080p stream that looked anywhere near the original camera file, so for me all the YouTube uploads are invalid. Vimeo streams too are only 7Mbit. Much better to wait until people with a clue get hold of the camera :)

[VanWeddings]

panasonic is taking the right approach. recycle previous generation sensor for a lower model, but add useful and valuable features. if canon were releasing the g6, it would be the g5 with wifi or something stupid.

[ntblowz]

Btw with G6 you can start and stop video recording from you phone/tablet, with 30fps streaming video too, pretty useful for controlling camera from a distance like Jib or crane, not sure if focusing is still available during video recording.

 

GH3 to me look like a half finished product that rushed to the market. no peaking, half ass WIFI feature..

[Sean Cunningham]

That is absolutely correct, but just to make it double clear: while YouTube compression can bring a bazzillion types of artifacts it will not render your perfect full HD resolution, carefully encoded video into something this soft...

 

By this logic all hacks/patches must render near identical detail...

 

 

...and let's not forget, there's no details given regarding camera settings, lenses, the method used in editing, or anything at all to give a sense whether the uploader knew at all what they were doing.  

 

Then there's the issue of even if this is the same chip as the GH2 it doesn't mean they used everything else.  It doesn't mean the baseline, stock compression is even as good as the under-achieving stock settings on the GH2 or whether it can be opened up yet.

 

Basically, judging this camera based on these clips isn't intelligent.  Period.  It might suck.  It might offer even more hackable potential than the GH2.  Nothing in these YT clips are a valid indicator of anything.

[Ernesto Mantaras]

I've never seen a single YouTube 1080p stream that looked anywhere near the original camera file, so for me all the YouTube uploads are invalid. Vimeo streams too are only 7Mbit. Much better to wait until people with a clue get hold of the camera :)

I have to insist, Andrew, and I don't mean to be negative but I don't want there to be false hope either. YouTube won't give you the original file quality, but it will certainly not destroy it as much to make this big a difference between two videos that, according to our expectations, should look very similar in terms of detail under the same conditions.

The detail present in the first video can't compete with the one you see in the second one, and considering they've both been equally destroyed by YouTube compression they should both look just as good or bad, and instead there's a clear winner in the GH2 video.

G6 video:
http://www.youtube.com/watch?v=3oijsASQQ6k
 

GH2 video:
http://www.youtube.com/watch?v=rcKuGoMN-VQ

 

Again, I hope I'm wrong on this, but what if Panasonic didn't want to take sales away from the GH3 and crippled the image quality in the G6? Because considering how popular the GH2 is, and if it has been improved in every way by this new compact model, then many people will think about it more than twice before spending almost double as much on the GH3. I know I would. The GH2 competes with the GH3 in video quality, but now it has been discontinued. Would Panasonic allow the G6 to compete against the GH3 as well?

And just to make it clear, the GH2 unhacked delivers great images by itself. The hack only boosts that up (and by far!) but detail is there to begin with, you can see it right out of the box. On the available G6 sample videos there's no such thing.

[Ernesto Mantaras]

By this logic all hacks/patches must render near identical detail...

 

Why? YouTube compression will destroy the quality of a video, but it doesn't do it to the extent that, say, a camera that supposedly shoots just like the GH2 ends up looking like an FZ200. If that was the case, then all the GH2 footage uploaded to YouTube should look like that, and it doesn't. In fact, that DrewNET video I posted was the reason I became crazy about the GH2 after learning a lot from it from EOSHD, and convinced me to take the plunge two days before I bought a Canon. You can't deny there's a lot of detail there in spite of YouTube trying to obliterate it.

[Sean Cunningham]

 You can't deny there's a lot of detail there in spite of YouTube trying to obliterate it.

 

And you can't deny that in-camera compression radically effects the perceived detail in an original recording.

 

And what if it was destroyed before YT had a chance to do its own damage?

 

Convince me you know that the baseline compression, matrix, all the goodies are set the same as they are for the GH2 and that Panasonic hasn't pulled a typical corporate move and de-tuned an existing setup designed to go into a lower-spec camera than the original platform.

 

Convince me that the uploader rendered to a professional codec before then compressing to MP4 with a high quality encoder.

 

Convince me that they didn't render to MP4 straight out of their editor.

 

All of that is incidental, however, to the fact of YT delivering bad quality content from unsophisticated or non-technical accounts.  Even playing back the HD streams with the window scaled down didn't serve to filter the ugliness of their compression.   And, sorry, but compression and bandwidth does affect detail and sharpness.  Otherwise stock GH2 would look as sharp as Moon Trial 5.  And it does not.

[Ernesto Mantaras]

And you can't deny that in-camera compression radically effects the perceived detail in an original recording.

 

And what if it was destroyed before YT had a chance to do its own damage?

 

Convince me you know that the baseline compression, matrix, all the goodies are set the same as they are for the GH2 and that Panasonic hasn't pulled a typical corporate move and de-tuned an existing setup designed to go into a lower-spec camera than the original platform.

 

Convince me that the uploader rendered to a professional codec before then compressing to MP4 with a high quality encoder.

 

Convince me that they didn't render to MP4 straight out of their editor.

 

All of that is incidental, however, to the fact of YT delivering bad quality content from unsophisticated or non-technical accounts.  Even playing back the HD streams with the window scaled down didn't serve to filter the ugliness of their compression.   And, sorry, but compression and bandwidth does affect detail and sharpness.  Otherwise stock GH2 would look as sharp as Moon Trial 5.  And it does not.

 

But of course Moon Trial 5 looks amazing! There is a very noticeable difference between that patch and the stock GH2 footage, and in all cases YT compression will affect detail and sharpness, I never said it didn't.
Perhaps we're not understanding each other correctly (maybe the fact that I'm a non-native English speaker gets in the way).

I AM saying that the footage was destroyed (or rather just plain bad) before being uploaded to YouTube.

I AM stating that I BELIEVE (I don't know for a fact, I'm inducing) that Panasonic did in fact made the quality of the G6 video mode not be as good as the GH2 to protect the GH3.

I AM considering the possibility that the uploader took all the wrong steps to end up with a shitty MP4 that misrepresents the G6 video quality, but considering all the sources I've seen so far (from several different uploaders) are just as soft, I am more inclined to believe in the "crippling" theory.

I think we're pretty much in the same line, but we're not quite tuned.

[andy lee]

I've never seen a single YouTube 1080p stream that looked anywhere near the original camera file, so for me all the YouTube uploads are invalid. Vimeo streams too are only 7Mbit. Much better to wait until people with a clue get hold of the camera :)

I'm with Andrew on this - too many people are trying to compare cameras - lenses etc off clips on youtube and vimeo

You cant !!!  there just is not enough detail there

The only way to test gear is in your own hands doing A/Bs against stuff you know.

A hacked GH2 file is incredably detailed - you just cannot see any of that on YT or Vimeo.

 

Anyhow the bottom line is the G6 is just another tool to get the job done

and its got a sensor I like alot!

[Brellivids]

If the video quality is lacking it would be a clever move from Panasonic to push people who were hoping to get little improved cheap GH2 to consider buying the more expensive GH 3 instead. 

[Andrew Reid]

You see how much confusion bad marketing causes?

 

Stills shooters would not put up a JPEG compressed to hell as representative of the camera's outright image quality, so I have no idea why these YouTube uploaders feel it is good practice to put broken video samples up and not provide the original files.

[TJB]

Interesting but not exciting. Still happy shooting the GH2. Noticed a GH3 at my local camera store on sale - 150 Euros off recommended retail price! Give me peaking and I'll think about it.

[liondor1969]

Panasonic -:)
I paid 1200 for gh3, Plese peaking in this kamera -:)