Panasonic AG-AF100 series

[Tony Waree]

It seems a lot of you are speculating this camera to record 4:2:2 internally, but bear in mind that the AG-AF100 is branded under the AVCCAM marque, which is Panasonic's equivalent to Sony's NXCAM; and in both cases, they are basically "souped up" versions of AVCHD with LPCM stereo audio 1920x1080 4:2:0 video and 24Mbps bitrate. Panasonic does offer 4:2:2 recording in the AVC-Intra 100 (Mbps) format, but that is only offered on their P2 based camcorders, a different product league. Like Sony's XDCAM EX camcorders that also subsample 4:2:0 for internal recording, HD-SDI out should be 4:2:2 (direct from the sensor) to any capable recorder.

[David C. Williams]

Originally Posted by Thomas Smet

24 mbit AVCHD is also a much better format then 35 Mbit mpeg2 so yes the encoding quality will be better then the EX1 or EX3. AVCHD is a much better encoding technology and 24 mbit AVCHD can look like 50 mbit+ mpeg2.

This is HIGHLY debatable. Personally I think it's rubbish. Mpeg4 is more efficient at low rates vs Mpeg2, but they even out as rates increase.

[Chris Li]

Although most of Panasonic's Lumix lenses for their GH-1 camera are on the slow side, they have a fast, highly reviewed 20mm f/1.7 Aspheric lens equivalent to a 40mm fov in 35mm world. That would offer shallow dof for video shooting.
They also have another fast prime in the pipeline - a 14mm lens, not sure if it's 1.7 or f/2.8 that would be the equivalent of a fast 28mm wide lens.
I own the Lumix 7-14mm wa zoom lens which while not fast at f/4 is very sharp and well corrected even wide open. That's a 14-28 in 35 speak and it's tiny and light weight. I'd love to see this one on the AF100!
I wonder if the lens' IS would work on this camera ? I suppose the auto focus is disabled?

Legacy wide angle lenses are pretty much expensive,useless poor performers in the m4/3 format.
Leica or even Voightlander wa lenses 12-15mm primes are only 24 - 30 when used in m4/3.

just sayin'.

[Thomas Smet]

Originally Posted by David C. Williams

This is HIGHLY debatable. Personally I think it's rubbish. Mpeg4 is more efficient at low rates vs Mpeg2, but they even out as rates increase.

Have you ever compared both formats?

Barry Green wrote an article comparing direct encodings from an EX1 recording both native 35 mbit mpeg2 and AVCHD through an external recorder hooked up via HD-SDI from the camera. The AVCHD had a lot less macro blocks. At one point I would have agreed with you but now that I have seen Barry Green's test and have compared the footage myself I can honestly say that AVCHD at 24 mbits is better then mpeg2 at 35 mbits. Maybe it isn't twice as good as some products would claim but it is better.

AVCHD had a bad reputation because it started out in cheap consumer cameras and the first encoding chips were not very good. The format has really matured in the last two years however and has become just as good and most of the time much better then any mpeg2 format outside of 50 mbit mpeg2.

Yes it is true they even out but 35 mbits is still a bit low for HD.

[Dom Stevenson]

Barry's a fantastic voice in video, who's reviews i read avidly, and i'm not going to dispute his vastly superior knowledge over mine. However his insistence on the merits of anything Panasonic versus other manufacturers makes me take his views with a pinch of salt.

And perhaps he's right, after all he did the tests. But i didn't hear these tests being announced regarding a non-Panasonic camera.

[Steve Connor]

Originally Posted by Thomas Smet

... have compared the footage myself I can honestly say that AVCHD at 24 mbits is better then mpeg2 at 35 mbits. Maybe it isn't twice as good as some products would claim but it is better.

Disagree entirely with this, 35 Mbits is NOT a bit low for HD, Discovery HD allow you 100% shooting on XDCam EX. I understand that some people favour Panasonic over Sony and vice-versa but it gets a bit tiring when the same old arguments persist that fly in the face of what people are actually doing in real world production, and that is where it matters, not in "tests"

[David C. Williams]

I've had my EX3 for 18 months, shot hundreds of hours, and I've never seen any macro blocking. If you personally have any examples, post them. I've read a few people saying this, and never once have they seen it themselves.

[Thomas Smet]

First of all I have no love for Panasonic. I am a Canon guy. I have developed extended mpeg2 editing codecs for Avid Liquid and consider mpeg2 to be one of the best formats we have ever seen. I am used to working with every form of mpeg2 all the way up to 300 mbit/s I frame (which I do not recommend due to file size) so I know how much further mpeg2 can go beyond 35 mbits.

Second Barry didn't use a Panasonic camera in that test at all. He hooked up a AVCHD stand alone encoder through the HD-SDI out on the Sony EX1. You couldn't ask for a more honest test because everything was equal except for the encoder itself. Usually a camera test is subjective due to the fact that so many other parameters of the camera can affect image quality. In his test however it only came down to raw encoding.

Third I'm not saying the 35 mbits from the EX1 looks bad at all. In fact I think it is excellent and we have used them at work along side our Sony F900 and they really held up well. My main point is that AVCHD at 21 -24 mbits/s can and does look at least as good as that and sometimes better. If you don't believe it then show me the samples that say otherwise. The problem is the only way you can prove AVCHD isn't as good is to take a raw HD-SDI source and encode it to compare. You cannot just do a software encode and get the same type of results either. It has to be from the type of hardware encoder you would expect in a camera. Only then do you have a true test of the encoded format.

I do not wish to downplay any camera or format at all because I love all cameras and formats. All I am saying is that AVCHD in this Panasonic camera is not such a bad thing at all. Even if you want to debate if AVCHD can look better it is hard to debate if it doesn't at least look as good as 35 mbit mpeg2.

[David C. Williams]

You can't make a specific claim, then ask others to prove it for you. The onus is on you to prove your claim. You voiced the assertion, you can't ask others to prove a negative.

I'd like to see macro blocking from an EX3? I'd also like to see 24Mbit 4:2:0 AVCHD looking better than 50Mbit 4:2:2 Mpeg2? Or even looking noticeably better on average than 35Mbit 4:2:0 Mpeg2?

XDCAM EX and AVCHD may occasionally look better and worse than each other in different situations, but neither is a clear winner, and you certainly can't say one beats the other. Certainly not 50Mbit Mpeg2.

[Thomas Smet]

Look all I have been trying to say here is that AVCHD at 24 mbits can look just as good as mpeg2 at 35 mbits.

This is not a knock on the EX1 or the EX3 which I consider to be excellent cameras. I also consider the XDCAM EX format a top notch format. I'm not sure how many others ways I need to say that. That still doesn't change the fact that sometimes AVCHD is going to give better results. Check out the NanoFlash as an example of what higher then 35 mbit mpeg2 can look like. Also take a look at Barry's article if you want. Other then that I cannot provide any samples right now and even if I could why would I want to? It isn't my job to defend a format and spend the time and energy to prove what one format can do. I still have the right however to point out what I know about a format. You can choose to believe me or not. I really don't care. What I do care about is people trying to tell other people that AVCHD isn't as good just because it is AVCHD. So yes if I see that I will point out that there have been studies (not by me) that this just isn't true.

Now if you want to post a few links to help me understand how AVCHD isn't as good as mpeg2 I would be glad to participate in a civil "discussion" on the pros and cons of both formats without being told I'm speaking rubbish. Other then that I have given some evidence to back up what I have said.

[Paul Curtis]

Most of my personal problems with AVC have been with the lower bit rate lite versions, the full fat version by comparison isn't so bad. Neither AVC nor XDCAM is that good when dealing with fine detail though.

However at a fundamental level i don't see why we're not getting cameras with less compression now - storage speed and cost is not an issue these days. I'd happily take 50mbs or 100mbs of mpeg. The encoders are around, simple and the support is there.

cheers
paul

[David Heath]

Originally Posted by Thomas Smet

Barry didn't use a Panasonic camera in that test at all. He hooked up a AVCHD stand alone encoder through the HD-SDI out on the Sony EX1. You couldn't ask for a more honest test because everything was equal except for the encoder itself. Usually a camera test is subjective due to the fact that so many other parameters of the camera can affect image quality. In his test however it only came down to raw encoding.

I commented on this at the time, and from memory the comparison only really compared how the codecs behaved with high levels of movement - which obviously is a factor which can stress a codec. But it's not the only factor. Two most obvious ones are how a codec copes with fine detail, and how it copes with even gradients, and to the best of my knowledge Barrys tests didn't look at those factors. That's not a criticism - it is very difficult to do scientifically and objectively, and differing scenes can give widely differing cnclusions. What it does mean is that you can't use Barrys tests to give a blanket "I can honestly say that AVCHD at 24 mbits is better then mpeg2 at 35 mbits".

The other point is that it's impossible to define just how good any codec is at a given bitrate. Coder A will likely give a different result to coder B, even same systems at the same bitrate. More expensive coders tend to give better results than cheaper ones (funny that....!) for real time encoding in camcorders. A coder may compare incredibly well with a clean, noise-free source, but fall apart the moment it sees noise in the video. My experiences with AVC-HD have been limited to the HMC150, and I certainly haven't found the codec in that to equal 35Mbs MPEG2 overall - though it handles motion well.

That's not to say it may not be a different story with a more expensive or more recent camera with a better encoding chip. But AVC-HD won't rival 50Mbs MPEG2 (as used in the new Canon) without a complete redefinition of the standard, if only because the AVC-HD spec doesn't include any 4:2:2 profiles.

Fundamentally, AVC-HD and MPEG2 are built on the same base of technology, it's just that AVC-HD CAN call on extra tools to improve the encoding efficiency - get the same quality for a lower bitrate. The question is whether getting extra recording time on a memory card is worth the extra complexity come the edit, the need for more powerful computing.

[Chris Hurd]

I've had to edit a couple of posts and withdraw others from public view entirely which have devolved into who is getting "defensive" over this stuff -- please keep in mind that this is DV Info Net, and we don't play those games here. Let's keep it technical, and above all professional. There's no need to get emotional or defensive about bit rates.

We don't do "platform wars" on this site. That includes formats, codecs, etc. Thanks in advance,

[Matt Davis]

It's been wonderful to play with really good 'extreme' glass with my T2i, and now the news is that I can keep the glass and swap out the 'back end' for a 'proper' video camera. My T2i experience has been a bit of a rollercoaster, but I'm sold on the glass and the large format sensor. I though it would have to be Epic or Scarlet, but the AF100 ticks all the boxes for me.

[Darren Levine]

We'll i wasn't going to bother posting 2 cents on a camera which is slated for practically next year, but this is just too interesting what Panasonic is doing.

On the argument of bit rate.... h.264 at 24mbit should be great, consider all the hdv cameras recording HDV mpeg2 at ~24mbit. they look relatively great, and h.264 is a more efficient compression. Also, i trust that panasonic wouldn't cripple a camera like that on purpose, they will make sure it looks good.

really fascinated that they chose 4/3rds, which was a dying format. if they truly keep the total pixel count at 1920x1080 we should see some incredible dynamic range and low noise in this camera.

also great that they'll be offering mounts in all brands. but i'll really hate to have to go back to crop factors, but if it's a fantastic camera then that's a sacrifice i'd be happy to make.

ok so 3 or 4 cents... but i won't hold my breath until December, useless until then