I doubt you would really notice the difference if it was recording 1920x1080. The difference between the two is so tiny that it would really be just a waste to force the encoder to waste bits to encode the extra pixels.

Would you rather have 1440x1080 with a clean compression or 1920x1080 with more compression artifacts and only a super tiny bit of extra detail?

Some other cameras may have a 1920x1080 record mode but it is almost useless because it doesn't really gain any extra detail. Don't get wrapped up in a numbers gain that have no meaning at all.

24p inside of 60i is more so to make sure any NLE out there will be able to work with the material. Most consumers are not interested in removing pulldown and working with raw 24p. While flags would have been nice it isn't needed for the target market of this camera who just want it to shoot nice video and do a straight edit and finish it up.

I wouldn't exactly get too excited over AVC-Intra either. There hasn't really been any images or footage seen from this format yet to show exactly how good it will look. It has all been pretty much merket hype at this point. Intra frame codecs tend to not have a whole lot of compression going on due to the fact that most of the compression comes from interframe. With Intra-frame you throw out most of the compression advantages to the format and left with something not that much different then mpeg2-Intra or mjpeg. Of course AVC-Intra may have a few compression tricks still in there but none of us really know until we start to see how well it looks.

"The direct optical path for a single-sensor design is relatively simpler, and also reduces optical distortion compared to the prism design used in 3-chip cameras. Difficulties with optical alignment in 3-chip camera designs practically limits the upper size of image sensors whereas the simpler optical path in single-sensor designs avoid these problems." http://www.cineform.com/technology/CineForm_RAW.htm

The difference is about 20%. Not tiny. If the difference was tiny then the wasted bits would also be...tiny.

Sorry, but Thomas is right -- and it's not as simple as that. Again, don't obsess over numbers -- you can't claim a "20% difference" as if it matters, because what's important is how it looks, and the average human being can detect any difference between 1440 anamorphic scaled to 1920 square, versus an original 1920 square -- as Thomas correctly points out, you really don't gain anything. Remember the current defacto standard format for HD broadcast masters is HDCAM, which is also 1440 anamorphic. And nobody in their right mind complains about it.

See Douglas Spotted Eagle's excellent input in this thread. The standard says nothing about the quality of the image -- and there has been no such thing as a full-raster 1920 broadcast camera in wide use today -- full 1920 is for display; it's not needed for acquisition. And remember also that these are consumer camcorders. 1440 anamorphic is more than sufficient at the acquisition level for 1920 playback. Don't get hung up on the numbers -- there are other considerations that are far more important, and that's a fundamental concept to grasp.

"Full 1920" is a marketing term, nothing else.

In this case the new Panasonic HDC-SD7 with real full 1920 at 13Mbps
can be not better than Canon HG10 at 15Mbps 1440x1080 ? Less bits per pixel...

No way to know for certain until you have both camcorders in hand, side by side, to compare the images that they're actually producing. And since that's not possible at this moment, there's no way to know for sure. It's utterly pointless to compare the numbers -- that's the whole point here.

I wouldn't exactly say a loss of 518,400* pixels is a "super tiny bit of extra detail", given that DV video is only 345,600 pixels. You'd be correct about compression artifacts given that you're also assuming a fixed compression data rate for 1440 vs full 1920. Canon could up the data rate to 18Mbps to account for full 1920x1080 compression, since it is in the AVCHD specs.

Do you think photographers could see a difference if their dSLRs took pictures at 1440x1080 but upscaled to 1920x1080 for print? You might find them in a mini revolt.

Let's separate what the camcorder is able to resolve from what is possible from the recording format.

Well, I find this statement puzzling. Most are clamoring for full 24P frames written to data, instead of having it interlaced in a 60i stream so we don't have to remove pulldown. Wouldn't that make it easier to edit?

I agree, most consumers are not interested in removing pulldown, considering it wouldn't have pulldown added to begin with. :-S

Consumers deal with 24P all the time. Watch what they do when they rent those Hollywood DVDs from Blockbuster. Those DVDs are raw 24P.

*480 x 1080 = 518,400

I convert almost all my HV20 video to 960x540 for three main reasons, it gives me square pixels which is good for the 3D stuff I'm doing, it saves a ton of disk space and CPU time and I'm hard pressed to tell the difference between 960x540 blown up to 1920x1080 and the original 1440x1080 blown up to 1920x1080! This is sitting in a room watching it on a screen mind you... not up close to the computer monitor examining pixels. Of course I'll never project my stuff on a cinema size screen either, though I do use my brother's HD projector and the footage looks fantastic on a large screen.

So, I'll agree with the popular sentiment that there are more important things than 1920 vs 1440. The quality of the encoder and actually recording a 24p file to the hard drive as opposed to a 24p-in-60i file are much more important IMHO.

Heck, I'd prefer an HV20 that downsamples and records 960x540 at 24p using the same bandwidth as the 1440 stream if I had my way!

Stop thinking of numbers and actually look at video where everything is equal and one is at 1440x1080 and the other is at 1920x1080.

Video cameras are a little bit different then still cameras. Are you saying a 1920x1080 image from the HV20 would look as good as a still at 1920x1080 pixels from a DSLR? How about a digital still from a DSLR at 640x480 pixels compared to DV? There are so many factors that go into image quality that pixel resolution is only a tiny fraction of the overall picture.

Most photographers I know only care about how the image looks and don't sit around all day wondering what a group of pixels will look like. If there was a DSLR that did anamorphic pixels and it looked great nobody would care what it was doing on the inside.

The point myself and Chris are trying to make is to stop thinking of numbers. We are dealing with images and art here and not math and equations.

Take a digital photo you have around and scale or crop it to 1920x1080. Then scale that image down to 1440x1080 and then back up again to 1920x1080 and compare the two. You will see just how little the difference is. Try it instead of thinking of numbers.

Listen to Wes above me. He knows what he is talking about since he is a fellow artist.

I agree with Wes although I do tend to start to notice a little loss of detail when going down to 960x540 depending on the camera. I tested this out a long time ago with footage from the SONY Z1. I converted it to 960x540 and it didn't have hardly any loss of detail at all.

Like Wes I also down convert my HDV material although I down convert to 1280x720p 24p. It saves me a lot of space like it does for Wes but then it is already at a HD standard that is easy to work with.

You forgot about the Hitachi Blu-Ray camcorders that are coming out this October. They have a 5.3 MP CMOS sensor and it records to 1920x1080 using h.264. Technically it should give you a higher detailed image than the Canon HG10, but then again even if the picture quality is better, I don’t think Hitachi’s implementation of h.264 will be as easy to edit as AVCHD. We’ll just have to wait until it comes out.
http://dvinfo.net/conf/showthread.ph...196#post722196

It’s funny how right when Hitachi decides to announce their new camcorders, Panasonic and Canon immediately announces their own just to take some media attention off them. Not saying that I wouldn’t do the same thing.

Now all we need is to have one of these companies release a semi Prosumer model just like the GS400. Perhaps there saving the best for last.

I can not afford to download the original source file, but the performances of the different codecs are suspiciously high. What is the footage like, how complex is the detail, the movement, and the noise, how plain are surfaces?

Also, how much grain does it have? I understand that Mpeg2 had problems with that.

Gizmodo's take on the HG10, with pics:

http://gizmodo.com/gadgets/hd-camcor...der-284706.php

For anybody who was curious about how to focus manually with the HG10, it's managed via the large round selector dial and joystick located on the flip-out LCD viewfinder bezel. Should have more info about that shortly,