Another new HD format. A new codec, carried in a new file wrapper, recorded on a new solid-state media. This is getting a little difficult to keep track of.

Sony XDCAM EX builds in the fairly recently established Sony XDCAM HD products, which are in turn derived from the earlier XDCAM products.

Read the rest of this entry »

This has bothered me for a while, but a bunch of junk mail just came through the letterbox which prompted me to write this down…

Every new widescreen Plasma, LCD or CRT widescreen is marketed as ‘HD Ready’ and people snap them up with the idea that they’ll get a better, sharper picture. But it’s so often a lie. Many of these TVs (the ones that cost less than an arm and a leg) are not full-resolution HD, in fact many are lower resolution than standard definition PAL television.

Where I live we have PAL standard definition TV. It is essentially 768×576 pixels (it’s a little more complicated than that, but that’s where we’ll start). There is PAL 16:9 widescreen too, which is basically 1024×768px (again, more complicated, but that’s the number we’ll work with). All Television in NZ is PAL. There is no NTSC, and no HD broadcasting. However you can get HD DVDs and the newer game consoles support HD.

HD comes in two flavours, 1080 and 720, they are 1920×1080px and 1280×720px. This is the same all over the world. HD resolutions are standardised.

So, we’re working with these numbers, 576, 720 and 1080 – the frame heights of each format. An HD TV (as opposed to an ‘HD Ready’ TV) will be either 720 or 1080 pixels high (and the appropriate width for that height). It’s these extra pixels that give HD television the boost quality – a PAL image will store a maximum of 414,720 individual pixels from which to contstruct the image. HD 720 offers 921,600 pixels (more than 200% that of PAL) and HD1080 has just over 2 million pixels, nearly 5x more image data than PAL.

So where to the ‘HD Ready’ televisions fit in? Well they look pretty, but to save money they have fewer pixel in their screens. A lower resolution screen is cheaper to make. To cope with this, they resize the incomig pictures to fit their screen. There are basically two non-standard 16:9 screen resolutions in these TVs, they are: 1366×768 and 852×480.

Looking at the 768 pixel screen first, to display a PAL image, it must scale it up 133.33%, for a 720 image the the scaling is 106.7% and a 1080 image gets scaled 71.2%.

For the 480 screen it’s 83.3% for PAL (that’s right, lower resolution than standard definition PAL), 50% for 720 (a good number!) and 44.5% for 1080 line image.

What does it all mean? Well the problem is two-fold – first, these non-standard resolutions mean that the incoming video signal (whatever it is) has to be resized and in realtime, this means there has to be image processing circuits in the TV which take all the images, run smoothing and optimising routines over them and then resize the images so they can be displayed on the screen, this processing alone is messing with the picture. The second problem is the scaling factors. With the exception of 720 into 480, they are all very awkward factors. Mathematically it is very difficult to maintain fidelity when scaling 200 pixels down to 89 – lots of detail is lost, colours are ‘averaged’ and sharp edges are blurred.

This scaling is something that happens when viewing 720 HD on a 1080 monitor, but the factor (150%) is more even, and the other way around (66.6%) isn’t too bad either. But amounts like 106.7% and 71.2% are never going to scale very well. To some extent these problems also occur in scaling SD PAL into HD frames, but the conversions are still generally more even, and upscaling is more forgiving than downscaling.

What does it look like? Well it will depend on the TV in question and the methods it employs to do the scaling, but I’ve made some examples with Photoshop that demonstrate the factors with a 100×100px sample image.

A couple of years ago, when Panasonic launched their P2 solid-state product (or more specifically the HVX-200 recorder using the technology) I posted on my site about it, and expressed a lot of skepticism from my perspective as an editor working largely in reality and documentary, about it’s viability for all but a few productions. My criticism was mainly to do with the cost of the media (around NZ$6,000) and size (4GB) of the available media at the time. Shooting at 1080i, a 4GB P2 card holds only 4 minutes of footage.

Yesterday I attended a seminar hosted by Panasonic and presented by Barry Green (of DVXUser.com) about P2 and solid-state workflows. Panasonic’s Rick Haywood pitched Barry’s seminars by saying “[Barry's] seminars are ‘fiercely independent’ and deliberately don’t toe the party line for Panasonic.” While I don’t doubt that Barry is independent, he is still, for want of a better term, a Panasonic Fanboy. The seminar was hardly independent, with many negative remarks about Sony’s products. But it was not without merit of course. Barry is very well informed about the products and workflow, but it felt that he comes at it very much from a drama perspective, but a lot of the work in New Zealand is documentary and reality TV, and often of low budgets, where the capital expenditure on P2 cards is hard to justify and where downloading from cards every half hour or so may be impractical.

The workflow suggested by Barry was to either edit directly from the cards (a handy and time saving technique for news, TVC, or even short film perhaps) or copy to hard drive and edit from that. Then archive the edited material after delivery. This makes me very uncomfortable, as I can’t help but think back on two projects I’ve worked on in the past where drive issues have resulted in lost media very close to deadline. In both cases the project could be recovered by restoring the media from tapes. Something that simply wouldn’t be possible with this workflow. I would be very very uncomfortable working on any major project without a backup of ALL my media available at all times. In a tape-based workflow, this backup is the tapes. In a P2 workflow this backup has to be created at some point.

Another point from Barry that I had some problems with was that with a non-linear acquisition format (P2, or XDCAM, or Editcam, or whatever) is that you can use space efficiently by deleting unwanted footage in the field. This is another thing that, as an editor, I am very uncomfortable with. I can’t begin to count the number of times I’ve found vital moments in what at first glance may appear to be useless shots. The camera op in the field doesn’t always have the perspective that the editor has later on, and by making that decision they are effectively limiting the options of the editor. I read an article years ago about digital photography saying, essentially, that many of the great photos of our time would not exist had digital photography existed, as many may have been deleted, in the field, by the photographers who had not realised the significance of the shot at the time. Often it’s not until you get out of the field and look it all over that you really see what you have.

While I agree that non-linear acquisition is probably the future, I think that Panasonic is still not quite on the right track. P2 cards are very expensive (especially given the base price of the flash memory they are based upon). There is basically still no path from P2 into a tape-based workflow (the likes of discreet Inferno and Flame are still largely tape-centric) which, while perhaps on the way out, for most post facilities the tape-path is still the best supported. I’d like to see Panasonic come out with a deck that would play DVCProHD tapes, hold 4 P2 cards and had an internal drive. This deck, in my imagining, would allow you do play P2 cards into a tape-based workflow as if they were tapes, would allow computer access to them, would allow copying to the drive (and playback from that drive) and would allow bit-for-bit copying to DVCProHD tapes from cards or drive. By offering a multifunction deck like this, the transition to P2 could be eased for many post production houses, and other large producers.

I like P2 in general, and I’m a big fan of Panasonic’s DVCProHD, but I think there a still many problems with the workflow for many people, and these problems are being minimalised and trivialised by Panasonic and it’s P2 advocates.