Panasonic Removable Lens Four Thirds HD AVCHD Camera

I probably have noticed since I don't look at a lot of low end AVCHD footage. It appears to be a 24p vs 30p issue:

Originally Posted by David Newman on a different forum
The camera is a 24p signal on 60i, that is pretty standard. However, there is a minor chroma crosstalk every 4-th frame (not a luma crosstalk with interlaced as this user shows.) This is because the 4:2:0 encoding of HDV, can't have separate chromanance information of each field, so when you get a 24p frame that stradles a 30p frame slot, there is some chroma crosstalk. However if you remove pulldown correctly (which all the CineForm tools do) you will not notice this. Luma is separatable and that is what your eye tracks, so the image looks fine. This is primarily a keying issue, and that I recommend live capture form the HDMI port at 4:2:2, so there are no issues will pulldown chroma crosstalk, or MPEG compression.
We don't know if the Adam Wilt example was 24p or 30p. David Newman's HDV example would tend to indicate it is a 24p over 60i issue. However, that would affect the GH-1 in 1080p mode, which with this camera, I think we all want to have available.

More investigation. But it seems they left quite a bit on the table if this is a problem. Maybe Canon limited the 5DM2 to 1080p30 to avoid this. I wonder if the HDMI's 4:2:2 chroma subsampling avoids this.

More:

" CineForm Aspect HD does support automatic extraction of 24p from HV20's 60i stream. However, Ridlen is correct that the 60i encoding of 24p is not completely reversible, whereas the in 24F in the XH-A1 and XL-H1 is it. The reason the lack of completely reversibility is the 4:2:0 60i encoding which has only chroma value for pixels over two scan lines -- these scan lines may contain data from different frames, cause chroma cross-talk. The luminance is perfectly correct. Luma is encoded at 60i, and chroma at 30p, you can extract 24p from luma but not fully from chroma. If the pulldown is extracted correctly, the chroma cross-talk will only appear on every fourth frame. As our eyes are far more sensitive to luma, it is very hard to see this in motion. The only time this can be an issue is when keying, for that I recommend using HMDI and shooting live via the Black Magic Intensity card (using the new 1.5 drivers.) All CineForm tools support real-time pulldown removal using that card from the HV20. When you do a live capture the 24p in 60i is encoded as 4:2:2, with has separate chroma values for each scanlines, now the pulldown is completely reversable.

David Newman
Yet more:

It is not a CineForm problem -- please guys, you should know better than that. Unfortunately this is one of those cases when I gave too much information for most users to ingest. CineForm can completely reverse the pulldown live form the HDMI feed, with zero chroma issues, it is perfect -- try it for yourself. However, if you capture from tape, over Firewire or HMDI the chroma's pulldown can't be fully reversed -- but 99% of you will never see the issue if you are not keying. So in both sitatuations CineForm is capturing over HMDI, in one case (live) it is prefect, in the other (tape) everything 4th frame has a 50% chroma cross-talk with the next. So this not a CineForm screw-up, this is the result of MPEG2 4:2:0 encoding.

4:2:0 encoding of 60i, produces a chroma signal of 30p (not 60i which requires 4:2:2 color sampling.)


luma field 0 1 2 3 4 5 6 7 8 9 0 1 2 3
in 24p frames 01 234 56 789 01

chroma field (4:2:0) 0 0 1 1 2 2 3 3 4 4 5 5 6 6
in 24p frame 00 11(2) [23] (3)44 55

For three fields to construct a chroma "frame" 112 and 344 we can ignore the (2) and (3) to get full chroma -- no cross talk -- the CineForm pulldown handles these perfectly. However the middle chroma frame is contructed from two fields from different chroma source frames, an therefore choma crosstalk. This can't be fully reversed.

But like I said, most of you will not see this, it is very mild unless you are keying. For keying, capture live via HDMI and the issue completely go away. __________________
David Newman
CTO, CineForm

And a link: http://cineform.blogspot.com/2007/07/canon-hv20-24p-or-not.html
 
Last edited:
Ok...I now convinced the problem is created like this:

The Digic 4 cameras are advertised to play to TVs, which requires a 60i output. They were aware of the 24p over 60i 4:2:0 artifacts and chose 1080p30 as a result.

Panasonic, knowing the desirability of 24p for over 6 years (the DVX100 was introduced at NAB 2003) and maybe hearing the complaints from 5DM2 users, included it. However, the Panasonic side was less attuned to the 24p over 60i issue; their consumer cameras are not video pro favorites and the HMC150 supports native 1080p24 and 720p24 in AVCHD. The brochure says "In the Native mode, AG-HMC150 record only active frames.". Smarter people than me will have to say what that means.

It may be that the GH-1 people got that wrong. If not, maybe there is a native mode no one reported (the video capability reporting from PMA was very weak). Or maybe it doesn't work. The 17Mb/s max bandwidth would indicate that attention wasn't placed on the best recording quality. Unfortunately, more bandwidth was a freebee that was passed up. Maybe it can be fixed in firmware, maybe not. Lisa, do you know?

We still don't know for sure if there is HDMI output or if there is, is it upconverted 24p over 60i 4:2:0 or uncompressed 4:2:2 (still 24p over 60i I'm afraid, but without the 4:2:0 artifacts.).

P.S.
It appears DVInfo.net blocked users who mentioned the issue which may account for it not being more widely known.
 
Last edited:
Thanks...but I'm not really happy. We know that the Canon's data compression is falling where we would expect it terms of the observed image quality. Better than anything but the best professional video products. And that they moved over a potential artifact (can you imagine the howl from the Canon shooters if they didn't?) However, the GH-1 seems to be the mid range of consumer camcorders and only records 1080p in a way that guarantees artifacts. (Cineform Neo Scene is $129).

The solution seems to be a portable recording device for over $3k. At least, also assuming the GH-1 does a full sensor scan, that recording device will have great stuff to record...IF...it outputs live HDMI.
 
As most of the new ultra wide zooms I see have no aperture ring so if you want really wide might only be able to go with newer micro or four thirds lenses and an adapter?

And? Where is the problem? If you need to go really wide you have the Olympus Zuiko Digital 7-14 f4, one of the best wide angle zoom lens available in the market (maybe only the new Nikon 14-24 is better than this). The new 9-14 is also nice for a fraction of the 7-14 price. And the new Panasonic MicroFT 7-14 also looks good.
 
.IF...it outputs live HDMI.

This is so important that I'm beginning to think that live HDMI just isn't gonna happen....

It would be killer, though. Having seen the SR11 captured stuff.

Using the NanoFlash or even building a miniRig that can handle this is totally feasible. Where-as the NanoFlash gets you completely portable/handheld, the Mini-RIG or PCi-EXpress to Laptop solution would be cheaper.

And, here, it's all riding on a big if: live HDMI or no?
 
If you are shooting video to the camera's AVCHD, the two Panasonic zoom covering a range of 7mm to 140mm (14-280mm 35mm equivalent, I do wish we used FOV rather than focal length) are probably more than good enough.

If you are recording HDMI to Cineform or nanoFlash XDR 160Mb/s XDCAM I-frame, you way want to look at primes.

http://www.cameralabs.com/reviews/OlympusE714mm/

http://www.popphoto.com/cameralenses/5427/olympus-zuiko-digital-ed-7-14mm-f4-af-lens-test.html

782008155541.jpg


http://www.popphoto.com/cameralenses/4597/lens-test-panasonic-leica-d-summilux-25mm-f14-af.html

9122007135424.gif
 
And? Where is the problem? If you need to go really wide you have the Olympus Zuiko Digital 7-14 f4, one of the best wide angle zoom lens available in the market (maybe only the new Nikon 14-24 is better than this). The new 9-14 is also nice for a fraction of the 7-14 price. And the new Panasonic MicroFT 7-14 also looks good.

The problem would be people with newer, none Olympus/Panasonic lenses wont be able to use them on the gh1 if they have no aperture ring. Using other manufacturers lenses is suppose to be one of the benefits of using this camera. Best people know they might not be 100% compatible with their current lens collection.
 
Using other manufacturers lenses is suppose to be one of the benefits of using this camera. Best people know they might not be 100% compatible with their current lens collection.

It has to be seen yet how other manufacturers lenses will perform with this camera. Sean Reid of www.reidreviews.com has tested some Leica M lenses with the G1 just to find that, contrary to his initial expectations and enthusiasm, the results were dissapointing and the Panasonic especifically designed for MicroFT lenses performed better than the M's. Of course, the problem wasn't in the inherent optical quality of the Leicas but in the design of the G1 mount and 4/3 sensor. The point here could be that this differences in performance may be taken into consideration when working with the G1 as a photographic tool but they could be not so critical with the GH1 as a videocamera (in any case, the sensor implemented in the GH1 is different from the the G1's).
 
The GH1 will take most lens systems out there with an adapter. The adapters do not cause vignetting and they do not change the crop factor. The mFT to FT adapter is sold by panasonic and there are many FT to other lens system adapters-- but people are starting to make direct adapters. You can even use C-mount lenses, though in 4:3 some of the c-mount lenses have some vingetting because they don't have coverage-- but most of them, when shooting video 16:9 don't have it, or if they do have a small amount. Other than C lenses, everything works well.

I find it amazing the way dvinfo.net is blocking and banning people who express certain things. Its astoundingly intolerant of them-- as if its some sort of political correctness based on technological ignorance (if you aren't ignorant, you're banned.)

I come from a technology background, so I'm used to working with engineers. Engineers understand the technology and when they have disputes they go to the source and determine the facts.

I am becoming aware of the fact that the video world is not this way. Film and video people are apparently generally creatives rather than technologists. And so there's a lot of myths that run around as fact. Worse, marketers play on these myths and perpetuate them, and then people who have even less of a grasp of technology perpetuate them further, and insist they are the case, even against people who actually do know what they are talking about.

An example of this is Steve Mullen. Somewhere, someone told him that CMOS sensors could only capture video at a shutter speed of 1/60th of a second. And he believes it to this day. (I bet they were shooting 60fps at the time.)

There's no reason to believe that the GH1 will be limited to a 1/60th shutter speed. I even went and looked at data sheets for various CMOS sensors to confirm that they support a variety of clock rates for reading data of and even variable clocks. And this means the shutters can be as long as you want, dependent on your frame rate and the time it takes to read the data off of the sensor. So at 24fps the GH1 should easily be able to do a 1/30th shutter, and I wouldn't be surprised if it didn't let you do a 1/24th shutter as well.

Another persistent myth is that bitrate is a determinant of quality. This one comes from the days of MPEG1 and MPEG2. Both standards had common codecs where the compression ratio was essentially fixed. DV has a fixed "compression" ratio as well. Thus in those days, the quality of your image depended on your tape speed, and effectively, your bit rate. This is not the case with MPEG4. You can take an uncompressed file and encode it in H.264 at 50Mbps and at 25Mbps, and depending on your settings, the 25Mbps can look better than the 50Mbps.

Finally, the biggest "myth" is the idea of the "Scarlet" as something to compare cameras against. This is the most absurd as well. The Scarlet is not out, it is not a real camera, its just vaporware. Plus, given that RED has already reneged on their promises for this camera once, its quite silly to pin hopes on it and compare every other camera to this fantasy ideal. Further, they charge $500 for a single 16GB CF card, and they also charge $500 for the CF slots for the RED One. I have no idea how much the Scarlet is going to ultimately cost when you put all the necessary modules together to have a real camera, but it is not going to be near the GH1.

Frankly, I think RED get so much attention because people who can't afford any camera can use it as the ultimate hot rod to lust over and discuss.
 
Last edited:
We don't know if a good image can be captured, but if it can, I think we can get bokeh.

Voigtlaender lenses work fine on the G1, so will work fine on the GH1. There are a bunch of G1 owners out there busy exploring different lenses. They shoot at 4k, so higher resolution than we will be using the camera for, and they are happy with the adapted lenses.

Or put another way, if the lens is of a given quality, then you don't lose quality on the G1.

From what I understand, Voigtlander lenses are high quality.

I've been thinking about the Navitar DO-5095 and DO-2595 machine vision lenses as primes. They are both f0.95 and the pair can be bought for about $1,100. On the GH1 these are effectively 50mm and 100mm. They are C mount, and I believe should give full coverage, but I'm not sure of that yet. Seems like a killer deal if they are.
 
Last edited:
Finally, the biggest "myth" is the idea of the "Scarlet" as something to compare cameras against. This is the most absurd as well. The Scarlet is not out, it is not a real camera, its just vaporware.

Well to be fair, the GH1 isn't out either. And by the time it rolls (June, July, August?) out I expect to have more details about Scarlet. Its been established that Scarlet is not vaporware, it's just a matter of when it will be released. And people are looking at it as the ideal because of the image quality that's been produced by Red One.

Here's the first Scarlet footage we've seen.

http://www.reduser.net/forum/showthread.php?t=27537

For vaporware I'd say it looks fantastic. If the GH1 can get me these vaporware results I'd be pretty ecstatic, but I'm not sure that's possible at 17Mbps. :)
 
Red is definitely not vaporware. It is the only spatially oversampled video camera out there that I am aware of other than the DSLRs. IF you can get the image off the camera, the GH-1 looks to be the best.

Leica lenses are nice, but $10k for an F0.95 50mm isn't in my budget.
 
Quoted price so far, Fixed lens Scarlet 2/3" with everything needed to shoot out of the box, $3750, Basic fixed "Brain only" $3000, 2/3" Cinema brain $2500 other bits extra.
We've seen early footage and prototypes over at Reduser.
We know the mini primes exist at least in prototype.
So no Liza, Scarlet ain't vapourware, and there's no need to diss it just to sell the GH1, different cameras for different markets, both rock in their own context.
Dave
 
Against a 2/3rds sensor, the biggest think the GH-1 is going to give up is in processing and storage. It stores a wavelet compressed non demosaiced 280Mb/s RAW image. The GH-1 processes that in camera and if the HDMI works, uncompressed 4:2:2 8-bit output.

Within a year or so there will be multiple 4/3rd, APS-C, and FF35 cameras with full manual controls and supporting dual link HDMI, and maybe some outputting Cineform RAW. The SI2k does that, but is using a 2k sensor, so it isn't oversampled. However, you it has been proven that you can shoot a best picture Oscar winning movie with it.
 
It's vapor until the first camera ships. And then it'll be beta.

This from a current RED user.

Kholi, you remind me of a very big discussion I had here years ago about the "vaporware" term... Good, old times! Remember that, Barry? :)
 
Back
Top