-
Step away from the Apple Keyer now! (And build a better one of your own.)
-
[Walter Soyka] “They did 98% of what they needed to do, but not the last 2% to really make it work. If that last 2% never comes, it almost makes the 98% effort a waste.”
Thanks for the kind words, Walter.
I think you’re being way too generous to Apple. As I pointed out below, the single objective measure of whether a keyer is fit for purpose is not whether it can cut a hole in your backing but whether it can resolve edge detail. On that score it’s essentially a 100% fail.
It also worth remembering, as I think you are doing here, our discussion about the defective lightwrap – which at the very least represents exceedingly poor quality control.
More than the details of what’s gone wrong, which are in themselves intriguing, I’m interested in what it says about Apple’s approach to pro video.
Simon Ubsdell
http://www.tokyo-uk.com -
[Walter Soyka] “Have you tried the PHYX Keyer? Lots of sizzle there, too, but I haven’t sampled the steak. Maybe it has both.”
I haven’t tried it, but I’ve certainly heard good things.
[Walter Soyka] “do you feel the low pricing of FCPX/M5 has sucked the oxygen out of the room, or has it created a larger user base that can make up lost margin in volume to fill holes like this poor keyer?”
I’d say that the user base for FCP X is simply vast – I’d guess it’s on a scale that’s never been seen before for a professional level NLE.
I’ve had users that include dentists and fire departments and hosts of others that you wouldn’t immediately have thought of as the target market.
Conversely, I’d have said that there is almost no market worth looking at for building something at the higher end like an advanced keyer. I have no doubt that Jeremy is right and the Apple keyer is all most users want if they are staying inside the app – and for those that are going outside there are all the usual suspects that will deliver great results.
Certainly there has not been a keyer released specifically for FCP X and I doubt there will be.
I’m not sure if that answers your question which seems like a very good one.
Simon Ubsdell
http://www.tokyo-uk.com -
[Jeremy Garchow] “I also think it’s the fastest rough cut keyer available. For that, I praise it. It is really easy to get a look roughed out, and content approved/recut, before moving on to the more detailed work in some other application. In short, it’s not worth nothing, but really close to nothing.”
Fair. The keyer is fast, and I was really stoked about it with the bald interviewee. But you’d think the NLE purpose-built for internet cat videos would have a solid keyer, what with all that hair!
(I kid!)
I just looked it up, and this is what 2013 Walter had to say about the FCPX keyer [link]:
[Walter Soyka] “I’m finding that the FCPX keyer is just insanely fast, and has pretty decent auto-settings — but I’m still struggling a bit refining the edges on hair to the point where I am happy with them. (This may be operator error, and I’m going to give it another go.)”
I will try pulling a key in Mamba when I get back to the office. I do expect that will be both very fast and very good.
Walter Soyka
Principal & Designer at Keen Live
Motion Graphics, Widescreen Events, Presentation Design, and Consulting
RenderBreak Blog – What I’m thinking when my workstation’s thinking
Creative Cow Forum Host: Live & Stage Events -
[Simon Ubsdell] “I think you’re being way too generous to Apple. As I pointed out below, the single objective measure of whether a keyer is fit for purpose is not whether it can cut a hole in your backing but whether it can resolve edge detail. On that score it’s essentially a 100% fail.”
Being overly generous to Apple is not something I’ve been accused of much lately!
I was referring to development effort. I think we are really saying the same thing. I suspect the amount of work they would have to do to fix the keyer is relatively trivial compared to the amount of work they did to get it to the state it’s in today.
[Simon Ubsdell] “It also worth remembering, as I think you are doing here, our discussion about the defective lightwrap – which at the very least represents exceedingly poor quality control. More than the details of what’s gone wrong, which are in themselves intriguing, I’m interested in what it says about Apple’s approach to pro video.”
Maybe they don’t know it’s a problem (i.e., the QC team doesn’t know what to look for, or the beta program is too small and not doing the kind of work that would show this flaw).
Maybe they do know it’s a problem, but don’t care (i.e., it doesn’t affect a large enough portion of the user base to be worth the effort to fix).
Is it ignorance or apathy? I don’t know and I don’t care!
*rimshot*
Walter Soyka
Principal & Designer at Keen Live
Motion Graphics, Widescreen Events, Presentation Design, and Consulting
RenderBreak Blog – What I’m thinking when my workstation’s thinking
Creative Cow Forum Host: Live & Stage Events -
[Walter Soyka] “I was referring to development effort. I think we are really saying the same thing. I suspect the amount of work they would have to do to fix the keyer is relatively trivial compared to the amount of work they did to get it to the state it’s in today.”
But at what cost to performance? Could we still scrub a thumbnail or get realtime playback?
I truly think that’s what Apple is going for here, a promise of near “real time”. At least for now.
-
[Walter Soyka] “Maybe they don’t know it’s a problem (i.e., the QC team doesn’t know what to look for, or the beta program is too small and not doing the kind of work that would show this flaw).
Maybe they do know it’s a problem, but don’t care (i.e., it doesn’t affect a large enough portion of the user base to be worth the effort to fix).
Is it ignorance or apathy? I don’t know and I don’t care! “
I know it’s going to look like I’m labouring this point excessively but I do care 😉
And I care because of what it might say about the culture at Apple – which in turn might affect one’s opinion of their entire pro video enterprise.
Let’s look at the scenario here.
Your name is Joe and you’re the developer who’s been tasked with building the light wrap component, and If you know anything at all about light wrap (and even if you don’t) this is probably no more than the work of a few hours if you’re being slow.
To recap, you take the matte output of the keyer (already calculated) and you blur it.
You then subtract the un-blurred matte from the blurred version.
You use the resulting edge matte to composite a blurred version of the background back over the keyed foreground.
You add a couple of sliders to adjust the blend amount and the blur depth for the matte, and you add a couple of compositing options for the blend.
Because this isn’t your first time doing this kind of compositing job (please tell me this isn’t your first time!!), you’ll have made yourself a note to check that you’ve handled the edge pixels correctly in your blurring operations – because it’s an easy enough mistake to make, but you don’t make it twice!
(Let’s just remind ourselves that there’s almost nothing to building a light wrap beyond the blurring operations.)
Then, if you know anything about developing a compositing application, you test what you’ve done across your ample collection of test footage – which you will make absolutely certain contains the commonest green/blue screen scenario, namely a talking head where the lower half of the body is cropped at the bottom edge of the frame.
Because there’s almost nothing to this operation, you’ll at least remember to check that you’ve set the blurs correctly at this point. If nothing else you’ll want to verify that you’ve set the range for the blur depth to your satisfaction.
At this point your supervisor will come over and ask you how it’s looking – and remember to ask you if you’ve set your edge pixels to blur correctly. Hey Joe, let’s see how it looks at an extreme setting, he’ll say, because he’s also familiar with scenarios where this has been overlooked. (I mean I take it that supervisor guy knows a little bit about compositing even if Joe is not that hot on it. There is someone at Apple who knows about compositing, right?)
But this didn’t happen. It got signed off just like that with the most obvious error in place – in fact, the only error that you really needed to look out for.
Now if Joe is like any other developer out in the real world, he will go home and in the wee small hours of the night he’ll be awake wondering if there’s a way he could have made his light wrap better. And he’ll probably think, you know what, I bet I forgot to check my edge pixels and he’ll come in early the next morning to fix it – if he hasn’t rushed across to town to fix it in the middle of the night.
But he didn’t.
Then the product gets released, and still if he was a typical developer Joe would be worrying away at whether he could have done a better job and he’ll spend his downtime worrying away at the keyer looking for possible improvements.
Or at the very least he’ll have enough pride and/or interest in his own work to keep playing with the keyer to see whether it works as well as he and the rest of the team had hoped. Because the sad fact of the matter is that most developers can never let anything go – even long after it’s released. And because most developers expect that their products will continue to grow beyond the moment of their first release – updates and upgrades which will make their creation even better than the original conception.
But in Joe’s case this didn’t happen. Ever.
And here’s the really depressing thing for me. Not only did Joe not care enough to look at his own work after it was finished. No-one else working on the keyer ever bothered to check if it all worked perfectly.
Not in the period before its release – and not once in the years after. Not to this very day.
No-one at Apple pro Apps was interested enough in their own keying product to spend their own time playing with it and discover this elementary mistake.
Really??
No-one at Apple actually uses this stuff?
Yes.
That scenario I can believe.
I can believe that pro video is simply not in their lifeblood the way it is for the majority of people who use their pro video products. “Just about good enough but only barely” is as far as they take it and no further.
Whereas the evidence suggests that in the real world of pro video product development, the best you can do can always be made better. Making the best even better is what gets you out of bed in the morning and keeps you up at night.
At Apple there is clearly a culture that says make it flash, don’t spend too long on it and once it’s done don’t waste any more time thinking about it.
Just to come back round to your point about beta-testing – beta-testing is pretty useless for picking up stuff like this unless you’re very lucky. You simply have to have systems in place to pick up basic technical flaws – even if it’s just your co-workers taking sufficient interest in what you’re doing.
So I don’t blame Apple’s scanty beta testing program – I blame the lack of pride in their job of the people doing the work, sorry to say.
Simon Ubsdell
http://www.tokyo-uk.com -
I agree that aspects of fcpx that feel hasty.
You make a good case. I have no idea what it takes to develop something like fcpx. As a whole, and not drilled down to one effect, I don’t think it’s necessarily easy.
I don’t know the hypothetical development environment. Any one of us could spin it another way where Joe is waiting on some sort of update to apply better blur strategies. Or maybe Joe knows something else is coming and decides that it’s not worth fixing it since it’s all going to change anyway.
Maybe not enough people have reported that the keyer sucks and there’s no light warp edge crop.
I don’t really know. Apple could fix the keyer, I’m sure they have the capability. Why they haven’t is a mystery.
I’ve submitted bugs to feedback and have actually received a response and asked for proof. I’ve sent proof and those bugs were squashed in the next version. So, Joe is busy doing something, he just doesn’t seem to be super concerned with the keyer.
-
[Jeremy Garchow] “I have no idea what it takes to develop something like fcpx. As a whole, and not drilled down to one effect, I don’t think it’s necessarily easy. “
There is a danger of talking about building an NLE as if it were the summit of all human endeavour. There are more complicated things that people have managed to achieve in far smaller time frames and with far more limited financial resources!
[Jeremy Garchow] “Any one of us could spin it another way where Joe is waiting on some sort of update to apply better blur strategies.”
I think it would be pretty hard to claim something like that – this is probably one of the most basic image processing functions there is and any and all issues surrounding it are well known. But I could be wrong.
[Jeremy Garchow] “Apple could fix the keyer, I’m sure they have the capability. Why they haven’t is a mystery. “
Let’s ask the question a different way. Supposing you had left a flash frame in one of your projects and it had gone live, would you not do everything you could to fix it? Even if no-one else knew it was there.
Isn’t that what makes you a professional? Your pride in your work and your determination to make sure it is as entirely free from errors as you can make it.
Either Apple can’t spot the flash frame (very bad) or they know that the flash frame is there and they just don’t care (very, very bad).
Simon Ubsdell
http://www.tokyo-uk.com -
[Simon Ubsdell] “There is a danger of talking about building an NLE as if it were the summit of all human endeavour. There are more complicated things that people have managed to achieve in far smaller time frames and with far more limited financial resources!”
I hear you. I also think that it’s not a trivial exercise.
Yes, you’d think with all the time that Apple has had, fcpx would be further along. Although it has come a long way from that fairly dismal first release. Again, I could probably attribute this to Apple knowing what was coming in the form of a MacPro and getting things to jive on dual GPUs (a new endeavor for Apple who once shunned the need of dual GPUs). I’ve mentioned it before, but I’m not trying to make excuses here. I try to take Apple as the sum of its parts. They aren’t just a pro content creation software company, and so I have to take the good with the bad. The straightforward with the confusing.
I do think that most of what we are taking about here boils down to performance and what FCPX can do on even the slimmest of Apple hardware. At some point, certain pieces are going to suffer?
[Simon Ubsdell] “I think it would be pretty hard to claim something like that – this is probably one of the most basic image processing functions there is and any and all issues surrounding it are well known. But I could be wrong.”
Since little ole me can build a light wrap in fcpx itself using nothing but mattes and transfer modes, I know that building a light wrap isn’t hard or CPU expensive. I don’t think you’re wrong. I also think the reason for a busted lightwrap is not because Apple is incapable, but I can see how it would feel that way.
[Simon Ubsdell] “Let’s ask the question a different way. Supposing you had left a flash frame in one of your projects and it had gone live, would you not do everything you could to fix it? Even if no-one else knew it was there.”
Of course I’d fix it, and it would keep me up at night.
Maybe Apple doesn’t care becuase you, Bret, and I are the only people who do. Everyone else seems pleased.
So, what does that say about us??? 😉
-
[Jeremy Garchow] “Maybe Apple doesn’t care because you, Bret, and I are the only people who do. Everyone else seems pleased.
So, what does that say about us??? ;)”That your professional.
Herb Sevush
Zebra Productions
—————————
nothin’ attached to nothin’
“Deciding the spine is the process of editing” F. Bieberkopf
Reply to this Discussion! Login or Sign Up