So what happened to the resolution test between the AF100 and GH2?

I hear ya Barry and agree 100%. I've been using the AF-100 as our main camera and recently been using a GH2 as our B-roll camera with mini-HDMI out to a monitor, redrock cage, etc... etc... so it's rigged like our A-cam but it still doesn't hold a candle to the AF-100. It's a great little B-cam though
 
I can understand the frustration with the color shifting and I've just come to accept it... The AF100 is clearly like driving with power steering comparred to manual steering on the GH2... They both have great uses though... The GH2 is so small and covert that it does things that I wouldn't w
 
Okay, I've gone about as far as I feel like going with this color-matching, res-matching, gamma-matching nightmare. The results are not what I would be happy with, but they're as close as I'm gonna get them with the workflow that's available to me, so -- if you wanted to configure an AF100 to vaguely match a GH2, you could try to use these settings:

GH2: CINEMA mode, contrast 0, sharpness 0, saturation 0, noise reduction 0.
AF100: DTL -4, V-DTL -1, Coring -4, chroma level +2, chroma phase -3, color temps 0, master ped -5, DRS OFF, Gamma Cine-D, Matrix FLUO.

Not an exact match, the purples aren't quite right, but ... meh.
 
Okay, testing is now complete. Files are being uploaded for those who want to peep.

Simple conclusion -- the OLPF is there and appears to be working. The GH2 looks a little sharper on strict black and white, with more noticeable aliasing and moire patterns especially in some of the color patterns. With its OLPF, the AF100 looks like it is a little softer with less noticeable aliasing.

When recorded out to the Nano, the AF100 is substantially higher color res than the GH2, but the Nano doesn't affect the luma aliasing patterns even at 280 megabits.

I'll post a link when the native MTS files are available.
 
Barry_Green said:
Okay, testing is now complete. Files are being uploaded for those who want to peep.

Simple conclusion -- the OLPF is there and appears to be working. The GH2 looks a little sharper on strict black and white, with more noticeable aliasing and moire patterns especially in some of the color patterns. With its OLPF, the AF100 looks like it is a little softer with less noticeable aliasing.

When recorded out to the Nano, the AF100 is substantially higher color res than the GH2, but the Nano doesn't affect the luma aliasing patterns even at 280 megabits.

I'll post a link when the native MTS files are available.
Click to expand...

Thanks for the testing....now go watch the games!!
 
Barry_Green said:
Okay, I've gone about as far as I feel like going with this color-matching, res-matching, gamma-matching nightmare. The results are not what I would be happy with, but they're as close as I'm gonna get them with the workflow that's available to me, so -- if you wanted to configure an AF100 to vaguely match a GH2, you could try to use these settings:

GH2: CINEMA mode, contrast 0, sharpness 0, saturation 0, noise reduction 0.
AF100: DTL -4, V-DTL -1, Coring -4, chroma level +2, chroma phase -3, color temps 0, master ped -5, DRS OFF, Gamma Cine-D, Matrix FLUO.

Not an exact match, the purples aren't quite right, but ... meh.
Click to expand...

VERY HELPFUL, thanks for that Barry.
 
Barry_Green said:
Okay, testing is now complete. Files are being uploaded for those who want to peep.

Simple conclusion -- the OLPF is there and appears to be working. The GH2 looks a little sharper on strict black and white, with more noticeable aliasing and moire patterns especially in some of the color patterns. With its OLPF, the AF100 looks like it is a little softer with less noticeable aliasing.

When recorded out to the Nano, the AF100 is substantially higher color res than the GH2, but the Nano doesn't affect the luma aliasing patterns even at 280 megabits.

I'll post a link when the native MTS files are available.
Click to expand...

Great one, so I have downloaded the GH2 file.
What it appears that you are saying is that the German test favored the GH2 because it was a BLACK and WHITE test. Thus the resolution appeared to be higher due to a lack of OLPF (and thus greater contrast) and aliasing was minimized because there was no color information.

Looking forward to seeing the AF100 file when you have a chance to upload it.

Thanks for this, Barry.

edit: I have just seen you uploaded the AF100 file while I was writing this post...
 
Last edited:
I copy this. Mucho Thanks!


Barry_Green said:
Okay, I've gone about as far as I feel like going with this color-matching, res-matching, gamma-matching nightmare. The results are not what I would be happy with, but they're as close as I'm gonna get them with the workflow that's available to me, so -- if you wanted to configure an AF100 to vaguely match a GH2, you could try to use these settings:

GH2: CINEMA mode, contrast 0, sharpness 0, saturation 0, noise reduction 0.
AF100: DTL -4, V-DTL -1, Coring -4, chroma level +2, chroma phase -3, color temps 0, master ped -5, DRS OFF, Gamma Cine-D, Matrix FLUO.

Not an exact match, the purples aren't quite right, but ... meh.
Click to expand...
 
they both look good. BUT, the Af100 will be the tool on ANY production over the Gh2. But, the Gh2 holds it's own image wise.
The Gh2 REQUIRES external monitoring, and even then, as Barry said, you need to hit record to see what you are getting. The Lcd on the Gh2 is pretty poor. Keep in mind that all frame rates other than 24p are 17 mbit/sec. The Gh2 is quite an amazing little camera, for what it is. But if your going into a production with clients, they will be much more comfortable with you showing up with an AF100. Then run your Gh2 as B and tell them it's for behind the scenes so you don't freak them out!
Thanks Barry for the test. But you should have just enjoyed your day watching the NCAA tournament :-)
 
dustylense said:
they both look good. BUT, the Af100 will be the tool on ANY production over the Gh2. But, the Gh2 holds it's own image wise.
The Gh2 REQUIRES external monitoring, and even then, as Barry said, you need to hit record to see what you are getting. The Lcd on the Gh2 is pretty poor. Keep in mind that all frame rates other than 24p are 17 mbit/sec. The Gh2 is quite an amazing little camera, for what it is. But if your going into a production with clients, they will be much more comfortable with you showing up with an AF100. Then run your Gh2 as B and tell them it's for behind the scenes so you don't freak them out!
Thanks Barry for the test. But you should have just enjoyed your day watching the NCAA tournament :-)
Click to expand...

I look forward to the day my client says, "yeah but my iphone 11 does 3D at 8K resolution at 240fps... and you only have an Epic?"
 
Thanks for taking the time to do these tests, Barry. I think they clearly demonstrate that both cameras can produce excellent results in practice. One thing that stands out to my eyes in the AF100 video is that its fine details show higher contrast horizontally than vertically. As these are MTS files, I take it these videos were recorded in-camera? Do you think the 4:2:2 color depth of the SDI output may improve the vertical contrast?
 
Thanks for the tests! Looking at the footage it's hard for me to judge which one is better in terms of moire and aliasing. I'm seeing large moire patterns on the af100 in the color, but I see small and circular ones on the gh2.. The GH2 looks sharper, but af100 does drastically better on the center blue. However, interestingly, both seem to resolve 1000 lines? is that correct?
 
Thanks for making this test Barry!

Looking at screen grabs from both clips it's obvious that the AF-100 is virtually completely clean, while the GH-2 has a little moire popping up here and there. ( I am surprised at how good the GH-2 image actually is. Not bad for a $1,000 camera )

Looking at the center target from your test, the GH-2 has problems reproducing the bottom right blue quadrant of the target. Here's a graphic of both targets side-by-side:

Middle Targets from Barry's Test : Double click to Open to Full Size
 
Whats interesting about this:

Im viewing it on a Pana Pro plasma via HD-SDI. Now I recently shot a window with the 5DMKII that had this tight pattern and the purpose was to explain that specific issue to a director, needless to say it wasnt where the 5dmkii excels with moire and I got my point across. We viewed that clip on computer screen and panasonic pro plasma and there was little difference.

But these charts Barry provided, show a massive difference. On my screen 30" Apple display I can discern moire on both clips viewing at 100% image size. The AF100 has less. Now when I put them up on the plasma again via HD-SDi. Umm where'd it go? The AF100 shows extremely minor amount and all lines are clearly visible whilst the GH2 still shows a bit more across the chart but its minor stuff. But neither look anything like my cinema display.

In viewing this way - another point - the AF100 resolves every chart perfectly all the way to the edge except the centre blue section where close to the yellow its hard to discern and the orange section, where close to the red its hard to discern. All others I can see perfect lines all the way to the edge. The black and white chart shows aliasing closer to the edges. The GH2 doesnt fair much worse but theres obvious aliasing on the fine lines across the charts. And the centre chart in the blue section, the lines are completely gone and its a solid colour about 30% out from the centre.

This is exactly what Im looking at.

Now Barry, how does this type of chart work? Are the charts 0 - 1000 TV Lines from centre to edge? Can you explain further what it is we are looking at here. Not in terms of results but what the chart is designed to show. Now to me, and Im not a tech guy like yourself Barry, it makes more sense to shoot colour charts for real world results over the black and whites, how many people shoot black and white?
 
This has been a very informative test for me, and I am very grateful for it Barry. I am seeing more smear, and less refined edges on the circles (especially red/green) with the GH2, but that would be about what I would expect with the lower bit rate and the movement of the camera. Just wondered from a methodology point of view how the movement was controlled so that was equivalent on the two cameras, as there appears to be more movement on the GH2?

Perhaps I am not reading this all correctly. Would you mind explaining what it is you are looking at here in these charts, so that I am sure to understand how they are supposed to be read?

I definitely see the added contrast (and sharpness) of the GH2, because of it's missing OLPF, but the aliasing actually feels pretty similar to me otherwise. The pale yellow definitely benefits from this on the AF100, but otherwise they feel very close to me.

As I stated before I now understand why the German test showed greater resolution. Because it was in Black and White. What I don't understand is how the Af100 seemed to have more aliasing. This does not seem to be the case here on your test. What could explain this?
 
Okay, a few things -- first, don't look at the res numbers for absolute information. This chart isn't framed for accurate res numbers. It's only for comparative between these two particular cameras from this particular framing at this particular moment in time. I like the Wringer for many reasons as compared to a strictly black and white chart, but this particular version of this chart shouldn't be used to rely on for resolution numbers. DSC has a revised version that I even helped consult on a little bit, which will have pinpoint accurate numbers when framed accurately. I don't have that version yet.

So -- this isn't for comparing lines of res. It is for saying "which is sharper" and "which aliases more".

As for "which is sharper" -- that can be determined simply enough by finding still frames (there are sections in each recording where the tripod was absolutely stable). To me it's simple enough to just read the word "resolution" on each side -- it's a little clearer on the GH2 version, which is an open-and-shut case that the GH2 is a little sharper in res.

As for aliasing -- that's why the charts move. Aliasing reveals itself in motion, not in static shots. I mean, yes, moire patterns reveal themselves in static shots, but aliasing itself might sneak by undetected in a static shot. Add just a little bit of motion to the shot and the aliasing and moire patterns start to stick out very visibly. So the motion in the shots isn't meant to be controlled or exactly matched or revealing in sharpness or whatever, it's just a little bit of jiggling designed to make the aliasing more visible.

Example -- the percolating, bubbling circles in the upper right circle pattern, in the upper-left red/blue quadrant. The difference in aliasing is pretty huge there. There's a little hint of it in the AF100, but much more clearly visible and sharply defined on the GH2 shot. The GH2 has actual circles of bubbling lava and a very obvious set of reflected circles on the vertical and horizontal axes.

In summary -- the OLPF appears to be in place and is taming the aliasing quite a bit, at the expense of a little bit of sharpness. And in these findings, it appears to confirm what Rick Burnett found in his testing, that both cameras showed aliasing and moire on b&w charts, but in real-world footage the GH2 showed visible aliasing and even moire in some shots, but the AF100 didn't.
 
Last edited:
Very clear explanations. Very interesting reading of a chart I have never seen before too. I think this pretty much explains the German test and it makes sense. It also dispels the myth that the GH2 is a "better" camera. What is surprising is how well it holds up to the AF100 though given it's inexpensive nature. Basically, I think this is a real gift for all those folks looking for an affordable A/B cam scenario. As I have made no secret of what I view as a large change in the working methodology of many videographers today (multicam), this is very valuable. Not only have you provided AF100 settings to match the two cameras, but we see more clearly where in certain instances one camera is a more appropriate choice than the other.

In a way it's a bit like looking at two swiss army knives. One has a really sharp blade, but has only a few tools. The other has a much larger fuller set, but is also a bit bulkier and not always necessary. The appropriate choice is based on what adventure you are heading off to for the day, and when you have a chance to work with an assistant or a team, you take both.
 
Exactly how I look at it, and the reason that I have both. There's rarely a case where both are equally suited for any particular job, like I said before -- moped and tractor. They're very different, and each is far more suitable to one particular use than it is to the other. I love the GH2 for what it is. I just don't pretend that it's more than it actually is. And, even if the images were exactly the same between them, or even if the image from the GH2 could be demonstrated to be "better", it wouldn't change the way I feel about the AF100 -- it's simply the more appropriate tool for the job in most professional shooting scenarios that I find myself in. Which is why it costs more. The GH2, on the other hand, gives you quite a lot of performance at a much lower price. A motorcycle is very fast and gets good gas mileage, and is cheaper -- like a GH2 -- but it's lousy at hauling luggage and when it starts raining it all goes downhill. A car costs more, and does a lot more -- especially if it's raining or you need to take more than one passenger. The more-capable product costs more and does more. It all makes sense... :thumbsup:
 
Last edited:
You must log in or register to reply here.
Back
Top