Finally saw GH2 footage first hand.. I am not upgrading. Staying with GH1.3

Status
Not open for further replies.
I love it how everyone says the GH2 is every way superior to the GH13 EXCEPT when you 'move the camera or have fine detail or a lot of movement in your scenes'

In other words, the GH2 is superior for taking still pictures, not making movies.
 
Blackout said:
Ok some pretty good high detail native hack clips:

NOTE: my blackout.com server has a high load right now so these may take a while to DL:

High contrast dynamic range scenario:

http://blackout.com/gh13hack/00024-freakytime_marilynshot-blackout_c2010.MTS
Click to expand...




Well, the clip did take a long time to DL - but it was well worth it.

It's nice to see young film makers tackling New Testament biblical texts with a contemporary theme.

This was quite obviously the story of how Salome danced laviciously for king Herod and, when asked what she wanted for a reward, was told by her mother to demand the head of John the baptist.....

Pity the clip ended before the head was duly presented to her on a silver platter :happy:

With Christmas approaching, I can't wait to see your take on the nativity.... :happy:
 
well i dl'd the c2010.mts file and i'm surprised you aren't complaining about the macroblocking in the underexposed areas because it's quite prolific. for example, checkout frames 850-870 ...the quads of the girl wearing the mostly white dress are all blocked up and those jumpy artifacts are easy to see in realtime playback. whenever she alternates her stance, her legs start moving in abnormal ways, lol.

but for the price of this cam, i'm certainly not complaining.
 
i think what blackout is getting at is that its not much more improvement in IQ in his opinion than what we have with our GH13's. I tend to agree too.
 
ajcourtney said:
well i dl'd the c2010.mts file and i'm surprised you aren't complaining about the macroblocking in the underexposed areas because it's quite prolific. for example, checkout frames 850-870 ...the quads of the girl wearing the mostly white dress are all blocked up and those jumpy artifacts are easy to see in realtime playback. whenever she alternates her stance, her legs start moving in abnormal ways, lol.

but for the price of this cam, i'm certainly not complaining.
Click to expand...

I would have cut out before then.... that's not really as good a clip as this one:

http://blackout.com/gh13hack/00028-blackout-gh13-freaktimeb-roll.MTS

Do not DL that until past 7am or the whole file won't be there.....
 
plasmasmp said:
This shot is amazing for testing max resolvable detail. No mud, just a hint of aliasing. Looks really really fantastic.
Click to expand...

I agree you can see every little detail on every little street and window and no breakup of codec even though it is hand held... The GH2 would have avery hard time with that shot.
 
MR Fanny said:
i think what blackout is getting at is that its not much more improvement in IQ in his opinion than what we have with our GH13's. I tend to agree too.
Click to expand...

I'm going a step further and saying it's a downgrade, not an upgrade in the IQ area. There are plenty of shots you just will not be able to do on the GH2 because the codec wont be able to handle it.
 
Blackout said:
...still, in reality you need to see at least a 20 or 25mbit out from GH1 footage to avoid re compression issues.
Click to expand...

Not necessarily. I've been experimenting a lot with encoding lately, and it's possible to re-encode AVCHD footage at 10Mbit H.264 and have it look pretty much EXACTLY like original 40Mbit GH13 footage (though, I've not tried frame by frame comparison, I only compared videos in motion. I suspect that there will be a couple of subtle differences).

Take a look at this re-encode of your video. http://www.mediafire.com/?sxpe0gr3s53bnm7 This one averages at about 1.2Mbps, and doesn't look fantastic, but I thought that it would emphasise the potential of properly encoded H.264 @ 10Mbit, as this is 1080p at 1.2Mbps (!).
 
Matthew P said:
6) Far better 720p in terms of sharpness.
Click to expand...

I use 720P a lot with my HMC40, and plan to mix footage with a GHXX, so I'm interested/concerned about this comment. Why would the GH2 have better 720P than the GH13 - won't the GH13 provide a higher bitrate?
 
Notice that I said 'in terms of sharpness'. Yes, the GH13 has a higher bitrate in 720p mode, but have a look at some GH2 720p samples. It looks like downscaled 1080p, it's so sharp. :) The GH1's 720p is sort of fuzzy, almost as if it's upscaled standard def.
 
Matthew P said:
... The GH1's 720p is sort of fuzzy, almost as if it's upscaled standard def.
Click to expand...
No, no, no...not at all. This was one of the main reasons I gave up my 7D for a GH-1.. The 60/720p is outstanding. As a matter of fact I think it was Plasmaxxxx who did a test shot with 7D 1080p vs GH-1 720p...and the 720p looked sharper (with more detail). It looks more like downscaled 1080p to me. IMO

But then again...that's in comparison with the 7D...
 
Matthew P said:
Not necessarily. I've been experimenting a lot with encoding lately, and it's possible to re-encode AVCHD footage at 10Mbit H.264 and have it look pretty much EXACTLY like original 40Mbit GH13 footage.
Click to expand...
While I have achieved similar results with the full-featured implementation of the MainConcept AVC encoder, there were a few of aces up its sleeve:

1. A six-core, 3Ghz, 10GB RAM Intel workstation
2. Expert tweaking of encoder I, P, and B-frame parameters
3. Multiple passes through the video to assess bit allocation before encoding each frame

The third factor provides the biggest encoding advantage over a real-time AVCHD codec. The power to spend as much time as needed to literally predict the future cannot be overestimated.
 
Matthew P said:
Not necessarily. I've been experimenting a lot with encoding lately, and it's possible to re-encode AVCHD footage at 10Mbit H.264 and have it look pretty much EXACTLY like original 40Mbit GH13 footage (though, I've not tried frame by frame comparison, I only compared videos in motion. I suspect that there will be a couple of subtle differences).

Take a look at this re-encode of your video. http://www.mediafire.com/?sxpe0gr3s53bnm7 This one averages at about 1.2Mbps, and doesn't look fantastic, but I thought that it would emphasise the potential of properly encoded H.264 @ 10Mbit, as this is 1080p at 1.2Mbps (!).
Click to expand...

Why are you re-encoding H.264 to H.264? I always transcode to an intraframe codec before editing.
 
lpowell said:
While I have achieved similar results with the full-featured implementation of the MainConcept AVC encoder, there were a few of aces up its sleeve:

1. A six-core, 3Ghz, 10GB RAM Intel workstation
2. Expert tweaking of encoder I, P, and B-frame parameters
3. Multiple passes through the video to assess bit allocation before encoding each frame

The third factor provides the biggest encoding advantage over a real-time AVCHD codec. The power to spend as much time as needed to literally predict the future cannot be overestimated.
Click to expand...

Hmm, there are a couple of free GUIs out there that pretty much do all the hard configuring for you. StaxRip is what I use, along with the 'placebo' setting for encodes. My Q6600 with 2GB of RAM manages about 1 frame a second. Worth it though considering the final quality of the encode. It's fine if time isn't too much of an issue, as a 100 min film would take about 40 hours to encode. To me this doesn't seem that long because I did some 3D work once that took 160 hours to render (though I never want to make such an inefficient 3D project again!).

Also, I thought multi-passing was to predict file size only, and not improve the final quality? Like a common myth... so I've heard.

Ben_B said:
Why are you re-encoding H.264 to H.264? I always transcode to an intraframe codec before editing.
Click to expand...

I mean as a final output to the web, not for editing.
 
I have found NOT going to an intermediate codec to produce superior results on my hacked gh1s than going through cineform. Any steps of further compression, even cineform, introduce slight changes and slight noise. Sony vegas pro uncompresses into 32 bit color space and then into whatever format of your choice... so 1 less step of convertion... but, you need a beefy machine... or neo booster 2. I haven't done a 2 hour movie in sony vegas pro yet. A half an hour project is as far as I have gone and then, with multiple layers, filters, and yada yada going on, yeah the machine started to chunk a little but it was still workable and the final output was fantastic.

If I do a full movie I'll probably go the cineform route or edit in edius, but for now, on my projects, I love sony veges pro 9e.
 
Last edited:
I have found not going to an intermediate codec to produce superior results on my hacked gh1s than going through cineform. Any steps of further compression, even cineform, introduce slight changes and slight noise. Sony vegas pro uncompresses into 2 bit color space and then into whatever format of your choice... so 1 less step of convertion... but, you need a beefy machine... or neo booster 2.
Click to expand...
I don't know if it is just me, but a lot of the very clean fooatge that I have seen in the GH1/13 footage section, has been through Edius. It almost looks sometimes as though the footage is playing straight from the card and that's viewing Vimeo versions. Either the output/encoding from the timeline is responsible, or you're right about native editing. Wished I could get this for Mac!


Blackout, thanks for putting forward your opinions on what you've seen. I'm in no rush to jump on the GH2 at the moment, but I do hope that Panasonic do something ( are able to? ) about this.
 
Status
Not open for further replies.
Back
Top