Sony NXCAM -- Announcement Coming November 18th

[Jeff Kellam]

David HeathAh, but I did say my experiences were specifically with AVC-HD as output by the HMC150 - it's quite conceivable the coder in the HMR10 is better than that in the HMC150, even if the same codec, same bitrate, etc. It's also conceivable that AVC-HD performs well with a "clean" input, badly in the presence of noise.

Its concievable the coders are different but extremely unlikely. The codecs are a mathematical based model standard. They should behave exactly the same if they are the same standard.

I think it's safe to say the simpler and less complex an overall image is the easier it is for any codec to compress the image.

Hovever, small pixel level image elements are not an issue for MPG-2 or 4 codecs.

The HMC is certainly capable of producing lots of very fine grain noise in comparison to other cameras which produce a larger appearing noise grain. The AVCHD codec uses a smallest luma predicition block size of 4X4 (MPG-2 is 8X8). During the luma intra mode predicion decision to determine luma direction, the 9 modes of the 16X16 overall block motion are evaluated by the codec and unlikely modes are filtered out (it's no wonder CPUs have a hard time decoding). Non-contiguous pixels or smaller than 4X4 blocks of pixels are filtered and have no effect whatsoever on the codec in the direction prediction.

The bottom line is noise is too small, and being fully random rather than directional, is filtered from the intra direction predicion. So there is no effect on the codecs performance (especially MPG-2) and the codecs take advantage of this.

Have a good read of this paper on intra prediction (the basis of compression) and you can easily understand how the H264 codec works and why it's not some vodoo. But it is far advanced from MPG-2.
http://www.waset.org/journals/waset/v13/v13-9.pdf

[Hans Ledel]

Originally Posted by Paulo Teixeira

Why shouldn't it be?

.

Don´t ask me.
I don´t know anything about this cam since I have not seen what this cam can produce

That is why I asked you.
I was under the impression that you had some inside info since you made that statement, that it was "much much better".


Cheers

Hans

[David Heath]

Originally Posted by Jeff Kellam

Its concievable the coders are different but extremely unlikely. The codecs are a mathematical based model standard. They should behave exactly the same if they are the same standard.

Not so - the codec specifies a number of features which any given coder *MAY* employ, but doesn't insist that any given feature is used. (For a full list of all the possible features, see H.264/MPEG-4 AVC - Wikipedia, the free encyclopedia , "Features") Leaving out any feature(s) obviously makes the coder cheaper and simpler - but less efficient at lowering bitrate (for the same quality) than MPEG2.

See the EBU report at http://www.ebu.ch/en/technical/trev/...3-schaefer.pdf . Ignoring the technicalities, the key section relevant here is on page 10 - "Implementation Reports", and in particular:

The H.264/AVC standard only specifies the decoder, ........... Therefore, the rate-distortion performance and complexity of the encoder is up to the manufacturers.

In other words, all H264 coders are not equal, and it is unrealistic to expect the coder in a cheap camcorder to equal that of a broadcast encoder costing many times more. As the EBU report makes clear, the oft quoted figure that "AVC-HD is about 2x as efficient as MPEG2" is only true when all the features are enabled.

That is not currently going to be the case in a real time encoder in an inexpensive camera.

I think it's safe to say the simpler and less complex an overall image is the easier it is for any codec to compress the image.

Yes, but different codecs can struggle with different things. High levels of motion, fine detail, smooth coloured gradients, noise - the list goes on. And one codec may handle one of those factors well, another badly. That's why it becomes so difficult to make blanket comparisons between them, and that's ignoring differences between individual coders of the same codec!

[Barry Green]

Originally Posted by Cristian Adrian Olariu

Sorry Randy but you are right only on the first part. HMC151 is the same to the HMC150 because it is the PAL version of the HMC150 but it doesn't have a pal/ntsc switch. I noticed there are HMC152 and HMC153, the numbers represent a specific zone of the market.

No, Randy had it right the first time. The 151 is NTSC/PAL switchable, and sold only in Europe. The 152 is PAL-only, sold in southeast Asia and Australia/NZ.

[Tom Hardwick]

You sure Barry? I'm in PAL land and the 151 brochure in front of me makes no mention of NTSC capabilities.

[David Heath]

Originally Posted by Barry Green

Are you 100% sure you didn't have DRS on? Depending on the contrast difference of that right-angle edge, that's the kind of thing DRS can do, .........

I'm virtually certain that what I saw was codec related for a variety of reasons. For example, I managed to arrange to step through frame by frame and it was odd - you could see a "grain" type effect (more like JPEG artifacting), but which didn't change frame by frame until it suddenly jumped to a different pattern. The number of frames between "jumps" was what I'd expect the GOP-length to be.

Since this was happening on pictures with little motion, hence the conclusion that it was related to the way I-frames were being compressed. The codec was being "broken" not by high movement or gradients, but by fine detail. Hence my wondering how you felt it compared with 35Mbs XDCAM on still, detailed images.

As said before, it's quite possible it was a combination of the codec and a camera front end feature.

Originally Posted by Tom Hardwick

You sure Barry? I'm in PAL land and the 151 brochure in front of me makes no mention of NTSC capabilities.

It will do 60Hz and 50Hz frame rates, both 720p and 1080i standards, though strictly speaking neither PAL or NTSC. (It won't do SD.)

http://www.resource.holdan.eu/specs/..._AG-HMC151.pdf Page 11, under "Recording Format".

[Jeff Kellam]

Originally Posted by David Heath

In other words, all H264 coders are not equal, and it is unrealistic to expect the coder in a cheap camcorder to equal that of a broadcast encoder costing many times more. As the EBU report makes clear, the oft quoted figure that "AVC-HD is about 2x as efficient as MPEG2" is only true when all the features are enabled.

That is not currently going to be the case in a real time encoder in an inexpensive camera.

I totally agree on this point. A lesser camera with lesser processing power can use a manipulation of the codec within the standard to reduce the required processing power and compression effectiveness.

I would hope Panasonic would use the same AVCHD compression in two products advertised in the same Pro line and produced in the same time period with the same intended market.

Also when you step through frame by frame and see differences in sensor noise; it's not the noise affects the codec, it's the codec affects the noise, there is a difference.

[David Heath]

Originally Posted by Jeff Kellam

Also when you step through frame by frame and see differences in sensor noise.......

That's not what I saw - stepping through there was a "pattern" which remained constant frame to frame, then suddenly became a different "pattern" which remained constant for a number more frames, then changed again. My memory is that the number of frames between "jumps" was what would be expected for GOP-length.

The "pattern" looked like compression on a still JPEG, especially around edges, and was separate to "normal" noise. (Which did change frame by frame.)

[Robert Lee Colon]

Amazing quality!!!
Sony | Product Catalog - Sub-Category Landing Page

the specs
Sony Product Detail Page - HXRNX5U

b and h listing
Sony | HXR-NX5U NXCAM Professional Camcorder | HXR-NX5U | B&H

[Jyrki Hokkanen]

One observation - bee and dog sequences show auto focus problems.
I have found Sony's focusing algorithm surprisingly weak with furry animals, pumping constantly even when they are in the middle of the frame. This has been the weakest point in V1/FX7, which otherwise is a great camera for wildlife filming in difficult conditions. With animals, there is often no time to focus manually.

[Alaa Ghuniem]

http://www.sony.jp/professional/News...b20100107.html

[Chris Hurd]

Now that we've opened a dedicated forum for these camcorders, please post new threads or reply to existing ones that are relevant to your desired topic. Thanks all,