Zmatte is dead, long live DVmatte Pro!

For those with long memories of the UK FCP User Group, you may recall my rants about chromakey and DV. How in theory it should not be done, but in reality it can be done with external plug-ins.

Two offerings lead the field in terms of quality and affordability.

I chose Zmatte over DVmatte Pro a long time ago because, while DVmatte Pro was actually better in the long run, it was too slow and fiddly for the extra 20% of quality you got, and I quickly found out that clients didn’t notice.

So I championed Zmatte. Four clicks and you’d be pretty much there.

Then in September 2007, as a member of the Pixel Corps, I tried version 3 of DVmatte Pro.

The interface has been slimmed and yet it’s more powerful.

The speed is astonishing – due to the fact that rather than burdening your Mac’s main processor with arty stuff, it uses your Graphics Processor to do the work, which is exactly what it was designed for. ‘Hardware Accelerated’ is always an exciting term for graphics geeks.

It does the usual magic of using the ‘shredded’ colour information in DV and HDV and supplementing that with extra detail information gleaned from the brightness (luminance) part of the image.

It now does all the graceful things we like in good chromakeys such as edge wrapping. Still not as gracefully as Zmatte, but the quality is superb and the speed…

… DVmatte Pro is bloody quick. You can actually watch a live preview. You can actually watch a live preview of HDV material being keyed on a MacBook Pro. An 8 minute edit of talking heads against green renders ‘better than hardware’ in about 40 minutes on such a machine – from 1080p EX1 footage. That’s fast.
You will need to watch the on-line tutorials as there are some things to understand, but if you’re needing to do chromakey with FCP from DV, HDV or XDCAM-HD, you really should check out DVmatte Pro.


The joys of tapeless workflow

Since making room for an EX1 in my life, something wonderful has happened to my edit work: I have plenty of well named clips in projects that are filed and searchable.

Of course, logging and capturing is the foundation layer of any edit. But sometimes edits happen to tight deadlines under pressure, and the time taken to patiently spool through each tape and refer to the shot logs, selecting the best takes and marking their in/outs ready for batch capture is quite frankly longer than the time to suck in a whole tape, apply DV Start-Stop detection and switch your thumbnails on so you can roughly sort into bins of rushes. And as deadlines shrink and edit budgets get tighter, there’s a temptation to not actually label each clip according to content. Just rely on bin names. And then you find yourself plucking shots from whole rolls of tape rather than individual clips. And that’s when you hit the needle in a haystack syndrome. The trouble is, taking in each clip and giving it a name takes a lot of time that, thanks to the absurd shooting ratio that DV offers (today I am working my way through 3 hours of footage for a 3 minute insert – a shooting ratio of 60:1), it would take days. And for a 3 min insert I have hours.

Of course professional crews don’t overshoot, and take careful notes of what they shot, where, when, who, and so on. But when a client hands you three tapes, what do you do? So to cut to the chase: The Sony EX1 is partnered with XDCAM Transfer – an application/plug-in for FCP users. Its party trick is to highlight, say, a dozen clips out of 60-100 on a card/disk, and enable you to name them all ‘exhibition GVs’, then jump around instantly between clips marking some good, some no good (which can make them disappear from your shot list), pick out some and add ‘sponsor’ to their filename. 15-30 seconds. Onto the next batch of shots. And so on. Select all, import, and lo – neatly labelled shots minus the dross. And I’ve been timing this: for a 16 GB card (an hour’s worth, maybe 100-200 shots for event work), I can log/label and import in half the time it would take to ingest the footage from tape. And this is firmly implanted in my mind today as I ingest 3 hours of PDX-10 footage with an unset clock, so DV Start-Stop detection doesn’t work, and the tapes are full of TC breaks. Yesterday I had 90 mins of footage shot and edited by end of play for a 6 minute item. Batch naming makes the editor’s lot a happy one, especially in time pressured environments.


Sony Z7 – a view from the trenches

If you’ll forgive the location-dropping, I’ve just returned from a shoot Marrakech where I managed to compare the Z1 and the Z7 side by side in a variety of shooting conditions. 12 hours before, I was shooting with a PD170. It’s been an eye-opener.

The Z1 is a venerable beast of burden, a sort of Ford Mondeo of adequacy which hides some truly great features behind a facade of bland competence. Not at its best in low light, it needs +3dB of gain in theatrical lighting. The lens needs help through electronic sharpening. The audio quality is adequate rather than startling. But beneath this bushel is an amazing camera. Quick auto-focus (even if it prefers to auto-focus on the background), snappy zoom (even though it’s servo-only and has a rubber-glove feel), and absolutely unburstable audio limiter; if you record events or accept a line feed, this is a wonderful feature. It is a rubber sheet, in that you really don’t want to imagine that you’ll need it, but you’re thankful it’s there when you do).

The Z1’s audio limiters and extra gain, tweakable white balance, hard OIS and HDV downconverts have made it a safe bet for event videography where there are no take-twos, no re-lights and often no idea what’s going to happen next. It’s battle-proven and whilst not perfect, it does the job well.

But let’s introduce the Z7.

The zoom is slower. But it has a true manual setting and you can whip it if you wish. The autofocus is lethargic, but the screen and manual focus is so good there’s no point in autofocus (I’d like to say that one just wanks it back for perfect focus). And then there’s the ‘see in the dark’ PD170-style low light performance. I’ll mention that the PD170 I used the day before had horrible audio problems: you could hear the lens servos interfering with the audio, the screen and viewfinder were difficult (unreliable?) to focus with – thus requiring Push-Auto to focus on the fly. Come back, EX1 – all is forgiven!

Yet the Z7 just kept going. I was reminded of my early PD150 experiences, shooting parties at night. On the Z1, I’d be shooting 25 fps at 18 dB Gain at wide angle, trying to get every last photon I could onto tape. The Z7 was happy to keep recording by torchlight, and the lens allowed me up to 60% zoom before cutting out the last stop of light.

The Z7 images of my borrowed unit needed some tweaking, but they exhibited the same lift in shadow detail as my tuned Z1 settings, and the highlights were good too – I think I could get them to DVX100 levels of filmic roll-off given a little time alone with a good CRT. In fact, the lens requires none of the sharpening that the Z1 demands, and produces natural looking images that I can compare favourably with DSR-450 footage.

The Z1 has got some great Event Videographer tweaks to it which I am yet to master on the Z7: tweaking the White Balance, sorting the manual audio levels with limiter for the same unburstable control, balanced shot transition, HARD mode Optical Image Stabilisation (truly amazing results for handheld work). I’ve not found the equivalents for the Z7. Yet.

As I watched Z7 and Z1 images go into FCP, I wanted to sell my Z1s. With the Compact Flash recording, I wondered about selling my EX1 too. The CF option is very compelling, especially for me as I write, watching 23 HDV tapes ingest. The Z7 is very very good. If only it inherited a few more things from the Z1.

The biggest draw of the Z7 in this day and age is the progressive modes. HD, the web, computer playback, it’s all progressive. Only CRT uses interlace. To convert Interlace to Progressive, you lose 25%-50% of your vertical resolution, making images soft. I know. I’ve been fighting this with my Z1s. The Z7 shoots native progressive images. That yields big sharp images from camera to screen. And that’s what sells HD.

The ergonomics of the Z7 are very good. But I’m concerned that by adding so much depth to the camera controls, people will not twiddle it or alter the controls to make it a better camera. Yes, the zoom is slow, the autofocus is ponderous. Work around it. The image is good, the screen is helpful, the sensitivity is wonderful.

The Z1 could do Standard Definition better than DV by going HDV and downconverting through hoops of fire.

The Z7 finally delivers HDV that deserves a true HD badge whilst fully serving the rest of us who need good Standard Definition video in a progressive world. And it almost sees in the dark.

The PD170 is dead. Long live the Z7.