Sony PMW-200 Brings HD 4:2:2 Workflow to XDCAM Camcorder Line

Originally Posted by John M. Kim

The cost of SxS Pro cards is certainly a bitter pill to swallow. Luckily I was able to find used 32GB cards on the auction site for ~ $325 each. Expensive, but much cheaper than new.

Not by much. B&H has a new 32GB Sony SxS card for $356.

[Jack Zhang]

Don't forget the 200 can use XQD cards in 50mbps.

[David Dwyer]

Originally Posted by Jack Zhang

Don't forget the 200 can use XQD cards in 50mbps.

Didn't think that was possible?

[Serge Kouper]

QUOTE:
Which lens remote is suitable for the PMW-200 as I have made a error with purchasing the below item

524CFi Intelligent Zoom Remote Control For ENG Lenses 524CFI - Lens | Manfrotto

Is there any way I can get that working with some converter cable?

Manfrotto MVR901ECEX and MVR901EPEX,
I ordered the 901EPEX and was told there was some delays (CVP UK) so I went for the ECEX instead. I have not received it yet, however I've been able to put my hands on them (at the BVE London 2013) and the feeling is quite similar. They seem ok but not "heavy duty", I prefer a lil bit the EPEX, a bit thicker and maybe easier to control. The ECEX is nice too (this is the one I'm going to get). According to the Rep on the Manfrotto stand, elecronically and mechanically they are absolutely identical. Be sure you take the EX model since you want to plug it in the 8 pin socket.

[Alister Chapman]

Yes, XQD cards can be used for UDF and thus 50Mb/s.

[Joe Cain]

Is the Sony PMW 200 capable of exporting 10 bit 422 to an external HD-SDI Recorder? Is it recording 8 bit 422 to the internal SxS cards?

Sincerely,
Joe

[Doug Jensen]

Joe, the answer is yes, to both questions.

Mastering the PMW-200-160-150-100 Camcorders

[Terence Morris]

Originally Posted by Ed Arszyla

Not by much. B&H has a new 32GB Sony SxS card for $356.

The more I consider this, the more I think investing in an external recorder is the way to go. Quite a bullet to bite, but most economical per Mb and opens the door to 10bit if ever needed. The samurai is about the same outlay as a couple of 64Gb SxS.

[Glen Vandermolen]

Originally Posted by Terence Morris

The more I consider this, the more I think investing in an external recorder is the way to go. Quite a bullet to bite, but most economical per Mb and opens the door to 10bit if ever needed. The samurai is about the same outlay as a couple of 64Gb SxS.

The reason the PMW200 has the XDCAM 422 codec is so you won't need an external recorder. It's fully broadcast capable as is. The less stuff hanging off my cameras, the better.

[Alister Chapman]

You have to think of SxS cards as a long term investment, just like the camera. They are not consumables, they are part of the system. Compared to tape, over the life of the cards SxS is very cost effective. I'm using cards purchased 5 years ago and they are just as reliable as the day purchased.

If you move to a pro broadcast system then expect to have to pay pro prices for the accessories that go with it. What your getting with SxS is reliability, very fast transfer speeds and peace of mind that it will work as advertised.

Yes you can hang an external recorder off the back of your PMW-200. Buy some SSD's and use that. But you will also need to figure out how to power it or take a second battery kit. You'll have to get some form of mounting system and there will be at least one cable between the camera and recorder. It adds weight and unless you are prepared to spend a decent amount on the mounting system it's going to get in the way and annoy you to the point where you won't use it. If you are going to use an external recorder you may as well save some money and just get an EX1 or EX1R.

[Terence Morris]

I very much appreciate your input, Alister. It has helped clarify my thinking looking at this from the broader perspective. I think on balance I shall be following your advice for the time being. I do mainly gun and run stuff at the moment anyway. Outboard acquisition is a nice option down the road should it suit a particular project. 10bit colour does seem very rich and detailed, at least in the examples I have seen, and would be much more forgiving of post-production tinkering of course.

[Alister Chapman]

There really is not much difference between 10 bit and 8 bit acquisition with most cameras. The noise levels of smaller sensor cameras are bigger than the smallest 10 bit samples, so much of the benefit of having 10 bit is lost. What does bring benefits is lower compression gained through using higher bit rates. Compression adds noise and most compression systems break the image up into small blocks called macro blocks. In highly compressed systems, when you start to grade the image in post production these blocks start to become visible and more often than not it is these blocks that lead to banding in an image, not whether it's 8 bit or 10 bit. 50Mb/s 422 XDCAM is a very good codec and well matched to the PMW-200. Yes, an external recorder can be used to reduce the compression ratio, ProRes HQ at 220Mb/s will reduce the artefacts and give a very small benefit in post, but the benefit really is very small. With low noise cameras (typically cameras with sensors bigger than 2/3") then 10 bit becomes more significant.

As always what is far more important is how you expose the shot, how you frame it. An accurately exposed shot will grade much better than a badly exposed one. Using the most appropriate gamma curve for the scene can help maximise what's captured.

[David Heath]

Terence, I've said quite a lot about 10 bit versus 8 bit in another thread which you may find relevant - http://www.dvinfo.net/forum/panasoni...m-tempted.html

Originally Posted by Alister Chapman

.........most compression systems break the image up into small blocks called macro blocks. In highly compressed systems, when you start to grade the image in post production these blocks start to become visible and more often than not it is these blocks that lead to banding in an image, not whether it's 8 bit or 10 bit.

Exactly right, and from the thread linked above it may be worth posting here again my suggestion for a simple 5 minute exercise with Photoshop which nicely shows exactly what Alister says.

Bit depth issues *MAY* give rise to banding, but banding is far more likely to be down to general compression issues - and I think that's exactly what you've found when you talk about your experiences with DVCProHD. (Which, as you say, is also 8 bit.)

I think most people can intuitively see why bitdepth may give rise to banding, but why compression should do it is probably not as obvious. If anyone wants proof, (and doesn't like theory! :-) ) then there's a very quick test you can do in Photoshop in less than 5 minutes. Make a new (blank) canvas, then select a narrow vertical strip on the left hand side and fill it with black. Select the gradient tool, and draw a line from the black area to the (white) right hand side of the canvas - the image should be a nice smooth black-white gradient left to right. Now try saving it as a JPEG - firstly with low compression, then with the highest level (quality=0).

Anybody still unconvinced that banding may be caused by over compression alone!? :-) If you want the (very basic) reason why, read on, if you don't like theory, skip the next paragraph!..................

[Terence Morris]

Thanks, both Alister and David, for this useful input to my education. It's particularly interesting regarding banding, which I came up against a while back with somewhat less professional equipment. I was rendering to monochrome and playing with contrast curves etc, in post, but the results were basically unusable. I come from a still photography background and my interpretation at the time was purely in terms of having a limited bit depth to start with (as in raw vs. jpg). But there are clearly other factors involved in compression. In any context, it is sound advice regarding getting optimal exposure, as you mention Alister. I look forward reading further into this.

[David Heath]

Originally Posted by Terence Morris

I come from a still photography background and my interpretation at the time was purely in terms of having a limited bit depth to start with (as in raw vs. jpg).

Raw v JPG is a bit of an apples and pears comparison, as the former is RAW data (so not a viewable image as such) whilst the latter is a "finished" product, but we'll gloss over that.....

There is another simple Photoshop exercise you can do if you're interested, which this time will simulate banding caused by limited bit depth, and how noise influences the resulting banding.

Start off as before. Form a new (blank) canvas, then select a narrow vertical strip on the left hand side and fill it with black. Select the gradient tool, and draw a line from the black area to the (white) right hand side of the canvas - the image should be a nice smooth black-white gradient left to right.

But now use the "levels" tool ("Image>adjustments>levels" on my version) and in the lower box ("Output levels") drag the right hand slider to the left until the bottom right hand box has the value 31. Click OK. Then go back into "levels" and drag the right hand slider in the upper ("input levels") to the left until you again get 31 in the corresponding upper right box, and click OK. You're back to a sawtooth - but heavily banded. This is banding caused by insufficient bitdepth - 5 bit, so it's hardly surprising!!

But now the interesting bit. (Well, interesting maybe if you're into geeky things... ! :-) ) Step back to the previous stage, and go to "Filter>noise>add noise". (I found a figure of about 2% worked well.) Now do as before - go back into "levels" and drag the right hand slider in the upper ("input levels") to the left until you again get 31 in the corresponding upper right box, and click OK.

And see the difference the noise has made!?! Still effectively 5 bit video, but the "dithering" effect of the noise has almost completely masked the banding. (Not that I'd ever advocate 5 bit video! :-) ) A couple of posts back, Alister Chapman said " The noise levels of smaller sensor cameras are bigger than the smallest 10 bit samples, so much of the benefit of having 10 bit is lost." Hopefully that Photoshop demo experimentally proves the principle of exactly what Alister was saying to be correct.

OK, you may argue that 10 bit can't be any worse than 8 bit, can it? With a 1/3" camera it may not give much of an improvement - but surely it can't do any harm?

To an extent - true. But generally for an acquisition codec bitrate is a precious commodity, within limits. Raise the bitrate, and you need more memory for a given recorded time, and higher bitrate may mean higher performance memory needs to be used. And 10 bit working will mean a 25% increase in bitrate over 8 bit with equivalent compression. I'd argue that for cameras in this class, for a fixed bitrate, that data may be better off used to lower the overall compression rather than give 10 bit - remember the previous Photoshop demo, and how it showed that banding could be equally caused by too much compression.

Of course, that may not be what marketing people wanting to sell you 10 bit recording solutions want you to hear.......... :-)