October 15, 2009
Why iFrame is a good idea
First off, as opposed to what the fellow in the Washington Post writes, it's not really a new format. iFrame is just a way of using formats that we've already know and love. As the name suggests, iFrame is just an i-frame only H.264 specification, using AAC audio. An intraframe version of H.264 eh? Sounds a lot like AVC-Intra, right? Exactly. And for exactly the same reasons - edit-ability. Whereas AVC-Intra targets the high end, iFrame targets the low end.
Even when used in intraframe mode, H.264 has some huge advantage over the older intraframe codecs like DV or DVCProHD. For example, significantly better entropy coding, adaptive quantization, and potentially variable bitrates. There are many others. Essentially, it's what happens when you take DV and spend another 10 years working on making it better. That's why Panasonic's AVC-Intra cameras can do DVCProHD quality video at half (or less) the bitrate.
Why does iFrame matter for editing? Anyone who's tried to edit video from one of the modern H.264 cameras without first transcoding to an intraframe format has experienced the huge CPU demands and sluggish performance. Behind the scenes it's even worse. Because interframe H.264 can have very long GOPs, displaying any single frame can rely on dozens or even hundreds of other frames. Because of the complexity of H.264, building these frames is very high-cost. And it's a variable cost. Decoding the first frame in a GOP is relatively trivial, while decoding the middle B-frame can be hugely expensive.
Programs like iMovie mask that from the user in some cases, but at the expensive of high overhead. But, anyone who's imported AVC-HD video into Final Cut Pro or iMovie knows that there's a long "importing" step - behind the scenes, the applications are transcoding your video into an intraframe format, like Apple Intermediate or ProRes. It sort of defeats one of the main purposes of a file-based workflow.
You've also probably noticed the amount of time it takes to export a video in an interframe format. Anyone who's edited HDV in Final Cut Pro has experienced this. With DV, doing an "export to quicktime" is simply a matter of Final Cut Pro rewriting all of the data to disk - it's essentially a file copy. With HDV, Final Cut Pro has to do a complete reencode of the whole timeline, to fit everything into the new GOP structure. Not only is this time consuming, but it's essentially a generation loss.
iFrame solves these issues by giving you an intraframe codec, with modern efficiency, which can be decoded by any of the H.264 decoders that we already know and love.
Having this as an optional setting on cameras is a huge step forward for folks interested in editing video. Hopefully some of the manufacturers of AVC-HD cameras will adopt this format as well. I'll gladly trade a little resolution for instant edit-ability.
November 24, 2008
EBU review of production codecs
This thread on DVInfo is where I saw it, if you want to discuss it further.
November 17, 2008
In search of the holy grail
Andy Ihnatko has a review comparing the Kodak Zi6 and Flip Mino HD. It's very thorough, though I'm not sure I agree with the conclusions.
In truth, comparing image quality on these is almost pointless, as they're all kind of crummy. They both have terrible rolling shutter effects, way too much sharpening and strange tonal reproduction. I think it's more about ease of use, feature sets, etc.
The things I like about the Zi6 are the ability to switch to SD mode, the ability to take stills, and the support for memory cards. The macro switch is rad as well. Things I don't like? The lack of a "recording time left" indicator, the strange battery life and the horrid audio on my example.
Follow the jump for my chart of how things stack up. Ugh! I'm including the Sanyo, because it records in QT-Native files, unlike the other cameras in its price range, which record AVCHD. Otherwise, I'd throw the Canon Vixia series in as well, as a very strong contender.
|Camera||Price||Recording Medium||SD Mode||HD Mode||Mac-Native||Flip out USB||Mic Input||Audio Quality||Optical Zoom||Battery Life|
|Flip Ultra||$129||Internal||X||X||decent||2 hours (AA)|
|Flip Mino||$159||Internal||X||X||decent||4 hours (rechargeable)|
|Flip Mino HD||$229||Internal||X||X||X||decent||2 hours (rechargeable)|
|Kodak Zi6||$199||Internal and SD||X||X||X||X||poor to decent||1 hour? (AA)|
|Sanyo Xacti HD1010||$599||SD||X||X||X||X||decent||X||2 hours (rechargeable)|
November 4, 2008
Dear Sony: Please join us in the 21st century
Seriously - go to Sony's HDV product page and try to figure out some distinguishing features between the Z5U, Z1U and V1U. Is the Z5U better than the Z1U? It's cheaper, so it shouldn't be right? At least they could fill in the 'compare' info...
Sure, if one puts in the effort, you can sort it out, but ugh. What a mess. Luckily for Sony, everyone else is just as bad.
October 29, 2008
DMCA Good? Your head asplode!
Wired has a great blog post looking back at 10 years of DMCA, arguing that while the anti-circumvention parts of the DMCA really suck, as a whole it's been a good thing.
Essentially, they argue that the DMCA, by spelling out takedown procedures and defining the "safe harbor," has allowed for a variety of services that would have otherwise been too risky, namely YouTube. It's a fair point, and an interesting read. The safe harbor provisions definitely give me some confidence in regard to Media Mill, but I'm not sure I would go so far as to say the DMCA is a good thing. And, the DMCA is like a soufflé - it'll only get worse with age. Each new, DRM-encumbered technology serves to remind us of that. So, I'd argue that the DMCA is a badly broken law with a few good provisions. We can do better.
October 27, 2008
Thoughts on Final Cut Studio 3
We've got about six months to go before NAB 2009, so I figure it's time to start making baseless predictions. So, here's my take.
First off, if we don't have FCS3 or some other major video announcement from Apple around the time of NAB (whether they attend the show or not), I think it'd be time for some serious questions about Apple's commitment to the product line. I don't think there's much truth in the occasional rumor about Apple shopping around Final Cut, and I assume that the big FCP shops already know the road map, but it's time to move the platform forward.
There are a few big areas that I see as logical steps for improvement.
GPU Acceleration in Final Cut
Motion already makes heavy use of the GPU for effects, and Final Cut Pro already has the capability for GPU accelerated effects as part of its FXPlug architecture. However, the bulk of the filters in Final Cut are currently running on the CPU, and the realtime effects architecture is achieved based on CPU speed alone.
Unified Solid State Camera Support
Importing video into Final Cut has gotten to be a bit of a mess these days - you've got "log and capture" for tape, "log and transfer" for P2 and AVCHD, plus XDCam and DR60 importers from Sony, plus standalone apps for RED, and a handful of others. It'd be nice to see an importer plugin architecture that allows different importers to plug in to the "log and transfer" framework, so that we could have a unified interface and workflow.
Premiere Pro CS4 does it, so Final Cut should too. I know it's not a "pro" format, but that doesn't mean people won't be creating compelling content with it. Transcoding to AIC really hobbles the support.
Please? It's already Schiavo'd, just end it.
BluRay in DVDStudio Pro
This one I'm a bit less certain of - Steve called BluRay a "bag of hurt" at a recent event, but Encore is doing it. If DVDSP isn't going to get BluRay support, it might be time to officially start phasing it out.
Better Roundtrip Support
Roundtripping is one of those ideas that works great in demos, but never seems to quite live up to expectations in real life. Final Cut -> Motion -> Final Cut tends to work ok, but FCP->Soundtrack->FCP is like russian roulette. Every now and again it'll work OK, but usually it just blows up in your face. It's not sexy or fun, but it sure would be nice if it worked.
Make Color Pretty
Nobody expected FinalTouch to get beautiful in one version, but for Color 2, it's time to get an Apple interface and a refined workflow. Better Quicktime integration would be nice, and again, improved roundtripping.
OpenCL Accelerated Codecs
This one is a bit more "out there," and the WWDC NDA prohibits me from talking about some of it, but OpenCL would provide some great opportunities for hardware accelerated video codecs. Not just for encoding, but also decoding of formats like HDV and AVCHD. Now, whether this will happen is a bit more uncertain - it'd be a pretty sharp departure from Final Cut's reliance on Quicktime. It'd also likely be Snow Leopard only, though perhaps OpenCL can be integrated into an app, the way Core Animation showed up in Motion before appearing as an OS component.
Avid and Adobe are doing it now, so it only seems logical that Apple will add some sort of script-sync feature. Is this something people use in real life, or is it just nice for demos? I'm not sure.
Real Final Cut Server Integration
It'd only be logical to start building in Final Cut Server integration, to really start putting a bit more pressure on Avid in that space.
Other Crazy Stuff
If they were going to make it Snow Leopard only, there's lots of other interesting stuff that could be done - grand central scheduling, proper soup-to-nuts 64bit, etc. Maybe for Final Cut Studio 4.
Hey, whatever happened to the Shake replacement?
October 23, 2008
I can't take it no more, so I'm going on a shopping spree, yeah
So, in Apple's quarterly earnings call, Steve Jobs made repeated references to the "opportunities" out there for a company with $25 billion in cash.
So, what could they buy? This little economic hullabaloo has beaten up a number of companies, so lets look at some market caps:
AMD - $2 billion
NVIDIA - $3.8 billion
Avid - $600 million
Sandisk - $2.1 billion
Sony - $22 billion
Creative - $155 million
Thomson (as in Grass Valley) - $400 million
Bogen - $17 million
Sanyo - $2.8 billion
Autodesk - $5 billion
Now, not all of those are great buys. AMD would be a bit like buying a lead ballon (and not a Mythbusters lead balloon). And some of these would be impossible given anticompetitive issues.
If I had that kind of pocket money though, NVIDIA, Avid, Sandisk, Sanyo and Autodesk would make a pretty nice bundle of acquisitions. $15 billion total, leaving $10 billion in cash still floating around. Nifty!
October 2, 2008
Scenarist captions hurt my brain
If you were going to pick a format for closed captions in the year 2008, would you a) choose and easy to use, xml-based, human readable, machine parse-able format, or would you b) choose a format that uses 7bit hex values with 1bit parity, in two byte chunks with machine level control codes (clear buffer, move cursor, etc) and make it a proprietary format that costs $170 for the spec?
If you selected b), you might be Apple!
Ugh. Scenarist captions. Brain hurt.
I'll do a bigger post about how to properly do captions with Compressor before too long. As soon as I get done writing a DFXP->SCC parser. Woot.
August 13, 2008
Olympic Political ads
Is it just me, or are all of McCain's ads standard def, and all of Obama's high def?
Should we read into that?
July 22, 2008
Lessons in Bad Dialog Design - From Apple!
I just purchased a season pass for Dr Horrible from iTunes, using a machine other than my desktop.
iTunes prompted me with this dialog box:
Wow, someone should really write a set of interface guidelines for stuff like this.
Here's a tip - if the bulk of the dialog box has to explain what will happen when the user clicks each of the buttons, you probably haven't done a good job. What it should be, and what 99% of other Apple applications would do, is "Would you like to make this computer automatically download all of your iTunes Season Passes?" with a button saying "Yes" and a button saying "No."
Best Demo Sources
CEPro has a slideshow of 35 great CD/DVD/BluRay discs for demoing your home theater system, selected by "industry professionals."
Yeah, I'm procrastinating this morning.
March 25, 2008
Angry Man Tilts at at Windmills
StudioDaily is hosting a rant from Scott Simmons about all those darned kids and their Final Cut Pro. He's upset that us young punks don't respect the old ways of doing things. While there is surely some truth in that, I don't think it's necessarily a bad thing.
It's a common criticism in many technical fields - essentially, it's gotten too easy/affordable to do something that was once a protected trade. The influx of new users brings folks of all skill levels, and with all kinds of attitudes about the way the world should work. Final Cut makes it "too easy" to be an editor, so existing editors that have had to work hard for their skills feel threatened.
It's worth a read, if only to get a rise out of people.
September 18, 2007
A look at consumer video
I've been thinking about the consumer video market a lot recently. I figured I'd take some time to sum up my thoughts about where we're at and where we're going.
First off, lets just all agree that for consumers, tape is dead. While HDV and DV will hang around for a while, coughing up blood, consumers should never again see a linear format for video. The optical formats (mini-DVD) may hang around for a bit, but they too should die off before too long.
Why? Because I think most folks who have shot home movies agree that the linear nature of tape is a huge disincentive to repeat viewing. Because tapes give you 60 minutes of continuous video, you're inclined to shoot far more than makes sense, which makes it that much less fun to go back and find the good parts. Similarly, bringing it into a computer is too hard and takes too long to make it worthwhile for most consumers.
The DVD formats are a bit better than tape, but come with their own negatives in terms of form factor and robustness.
That brings us to solid state. Most cell phone video is truly terrible right now, but another few months and we'll start seeing more VGA-quality recording. For a lot of people, it'll be good enough to capture Jimmy blowing out the candles on his birthday cake, etc, just like for many folks, the stills produced by a cell phone camera are "good enough" for day-to-day events.
The next step up from cell phones is still cameras with video capabilities. These vary widely right now - some cameras don't do video at all, some do video with no audio, and some do proper all singing, all dancing video. The Casio Exilim line, with its "youtube mode" is a good example. Quality takes a step up for a cell phone, but the form factor is oriented towards stills (obviously) and audio, if present at all, is bound to be subpar.
Moving beyond still cameras (incidentally, the Exilim retails for well under $250) you get into a murky world of old-line camcorder manufacturers and new companies you've never heard of.
To start out with, you've got the Pure Digital Flip Video Ultra. For under $200, you get a camera that will record between 30 and 60 minutes of video (depending on the model) with a not-all-that-horrible microphone, 640x480 30fps video and a rather neat built in USB plug. Aimed squarely at YouTube users, the Flip may be a good choice to give to the kids for recording their hijinks, but the lack of a still-frame mode means you're going to be carrying the Flip in addition to your digital still camera. Who wants that?
If you're looking for a decent still camera with good video capabilities and a nice form factor, you really need to investigate the Sanyo Xacti line. Whether you'd like a waterproof SD camera, or a full 1080i monster, they've got you covered. The Xacti line differentiates itself by providing a proper lens, and on many models, a microphone input. That's a feature that really sets the Xacti line apart for folks who want to do slightly more professional recordings. Quality isn't quite at the level of an HDV camera, but it's not bad either. If I had to spend my own money, I'd probably get their HD2, which shoots 720p video and 7.1 megapixel stills. I love the form factor - it feels great in your hand, and the "one button for video, one button for stills" layout makes it incredibly easy to use.
Downsides? The video isn't optically stabilized. The digital stabilizer is slightly better than absolute crap. No problem on wide shots, but doing a 10x zoom handheld is going to be pretty jumpy, especially after that second cup of coffee.
Expect to pay around $350 for the lesser Xacti models, on up through $600 for the HD2 and $800 for the HD1000.
From there, things start to get really serious, as we start to find the old guard of camcorder makers. Sony enters with a variety of Memory Stick and HardDrive based models. I'm not a fan of harddisk recorders - you've got moving parts, you're limited to the built in capacity and battery life takes a hit. The HDR-CX7 meets all the right criteria though - direct to memory stick recording, AVCHD recording, and all the features you'd expect on a consumer camcorder (no mic in though!). 6.1 megapixel stills round out the feature set, and you can find it for around $900.
That said, the real bargain in this space is the Panasonic HDC-SD5. You get 3 1920x1080 imagers, optical image stabilization, and direct-to-SD card recording. The fact that it's 3CCD really sets this model apart. It records in the same AVC-HD as the Sony, so you can expect compression quality that is only slightly inferior to similar HDV cameras like the Canon HV20. The Panasonic will run you a bit under $900.
Downsides? Form factor. You won't slip either the Panasonic or the Sony into your pocket. You'll look like more a tourist with your fancy camcorder. It's not huge by any means, but it can't be an after-thought in packing either. For me, the lack of proper audio inputs is a disappointment as well. I think that the higher-end consumer cameras really need to get features like audio inputs if they want to differentiate themselves in the minds of savvy consumers.
So that's where we're at right now. Most of these cameras include basic editing software in the box, but keep in mind that iMovie '08 should manage and edit the video produced by any of these models. Nifty!
September 11, 2007
The brilliance of iPhoneSimFree
So, say you're a company that intends to market a software product whose primary audience is the hacker set. You know that your product will be instantly reverse engineered, recoded and rereleased. But you'd still like to make a buck. How do you do it?
Generate huge buzz (and it doesn't get much bigger than "software iPhone unlock!"), stall for 6 weeks, while leaking out more and more proof that you really did what you said you did, and then require prepayment before releasing the software. Churn out a few thousand presales at $50+ per copy. Release the software to much joy. Gather up the money for 24 hours.
Then the hackers release their version, and you go on an extended, well earned vacation.
June 12, 2007
WWDC Bummer Roundup
So yesterday was a SteveNote out in San Fran. Here's the gist - Leopard still looks pretty cool, but it won't blow your mind. There still isn't a way to develop for the iPhone (though they're very proud of the fact that you can make webpages for it), and that's about it.
MacRumors has confirmation that the iPhone won't ship with Flash support, which is quite a bummer. One hopes they'll still move this direction, particularly if there is user outcry. The web is a lot more useful when you can load Flash content.
MacRumors is also saying that ZFS won't ship in Leopard, and certainly not as "the filesystem." Sun's Jonathan Schwartz said last week that ZFS was going to be the future of OSX. One hopes that this isn't a case of a major project being "Steved" ...
Hopefully more good comes out of WWDC for the developers that are out there. For those of us watching from the sidelines though, it's been a bit of a disappointment.
May 18, 2007
Edit: yeah, I'm being emo. Deal with it. Off to italy for a while, check the travel blog.
May 4, 2007
Why I hate Flash (and why CS3 is a blessing and a curse)
Since Carla asked ...
I don't hate Flash the concept, or Flash the delivery platform, I just hate authoring Flash. Flash is hands down the best way to deliver video on the web, and I'm really exciting about Apollo and AMP.
I just got Adobe Creative Suite 3, which includes Flash CS3. There's just so much about Flash (the application) that drives me insane - the way the workspace seems to always need more desktop real-estate than you've got, the way palettes fade out, and a million other minor nit-picky things that make it unpleasant to work with.
That said, that's not the reason for my comment yesterday. That was out of the frustration of learning that using their new FLVPlaybackCaptioning component (which lets you really easily add captioning to any FLVPlayback component) requires you to port the application to Actionscript 3. Realistically, it's probably a good thing to port to AS3, because it's actually a decent language, but I had hoped with something as important as closed captioning, they'd allow easy integration into existing applications.
Beyond that, they've changed the way FLVPlayback works such that it no longer loads FLV files from a URL if that URL includes a query string (download.php?video=1234). That makes dynamic video playback a bit of a pain - you need to move to a mod_rewrite solution (download.php/video/1234/.flv) on the server side to trick Flash into loading the video. Furthermore, they haven't fixed the crashing-bug related to loading large (>200meg) FLV videos on Windows using FLVPlayback.
Anyways, to sum it up, I love the delivery side of Flash, hate the authoring side. That's how it goes though.
April 19, 2007
I'll post some more product-specific stuff as I get time, but I wanted to post my general thoughts on this year's show, now that I'm safely out of Las Vegas.
The show seemed even more massive than in the past, with the north hall almost entirely taken over by video folks. It's really a hard show to get anything useful out of, unless you go in with specific needs. It's fun to wander around to the 16 different jib companies, but unless you're going to put the time in to really evaluate the differences between each, it's a bit silly.
I'm also feeling a bit turned off by the whole scene. The show consists of far to many middle aged men looking for a free trip to vegas on the company's dime, rather than folks actually there to evaluate technology. It's just too much of a "good old boys" club. Sitting in the Scopebox booth, I had more than a few of them do a "you buys enjoying vegas seen any good wink wink shows? you should really go to XYZ strip club" ... bleh. Not only is it obnoxious, it's slightly creepy.
The flip side of that is that the same sort of middle aged men with suspenders tend to be representatives in the booths of the major vendors. And they tend to think that nobody under 30 could possibly have any reason to purchase their product. Reps from Sony and Grass Valley literally turned and walked away from me when it was my chance to ask a question. Panasonic was nice however, and that will weigh on my mind as I'm evaluating who to purchase three HD studio cameras from.
On the plus side, folks from outside of the US are consistently friendly and helpful. Props to Telestream in particular.
March 20, 2007
Twitter is weird
So, twitter is pretty weird hu? Yeah ...
January 25, 2007
I'm so Web2.0 that I bleed rounded corners
Sorry for the slow updates lately. We've started our semester here at the University of Minnesota, so much of my time is being spent getting the young folk excited about production. Woo.
Anyways, the video industry is in a post-CES, pre-NAB lull right now. Everyone is off drinking at Sundance,and frankly I don't really care about Sundance...
December 30, 2006
AACS cracking - here's the deal
Earlier this week, news broke that someone had "cracked" the AACS DRM system used by both HD-DVD and BluRay. At this point, there hasn't been independent verification of any of this, but here's the deal as best as I understand.
A guy named muslix64 on the Doom9 forums figured out a way to extract title keys from HD-DVD discs, very likely using a vulnerability in Power DVD 6.5. He then wrote a decryption tool based on the publicly available AACS specifications. He released the software, including source, but did not release any title keys. He made hints that there is a fundamental flaw in the way title keys are handled, and that player revocation is unimportant. With player revocation, a flawed player (such as Power DVD, if indeed it is vulnerable) can be disabled from playing future discs.
It's interesting, if indeed it's true. There is no evidence that the AACS encryption itself is flawed, and indeed, that seems unlikely as it's essentially an implementation of AES. However, we know that using encryption for DRM on untrusted hardware is likely to have some vulnerability, if you're willing to dig deep enough. That is apparently what's happened here.
What does it mean? Not much right now. Muslix64 has disappeared, and nobody has been able to discover the relevant title keys at this point. If someone does recreate a title key extraction method, it would be an interesting alternative to the world of DVD cracking. Instead of having a DMCA-violating circumvention device in the form of DeCSS or any of the later decryption programs, you could instead have a totally legitimate decoder tool paired with some title keys. The legality of the title keys would be questionable - the court case would be very interesting. In any case, it's likely that you'd get your HD-DVD, check online to get the title key, paste it into your ripper and be done.
Give it 6 weeks and check again. This could be a blip, this could be the death of BluRay and HD-DVD. We'll see.
December 25, 2006
The truth of depth of field
A number of sites have linked to this article about depth of field, which looks at the commonly held belief that zooming in creates a shallower depth of field.
The article isn't super clear, but the essential fact is that zooming in doesn't actually change depth of field, it just makes the stuff in the background bigger. The stuff is no more blurry, it's just easier to see the blurriness.
He's absolutely right of course, but it's really splitting hairs. In the end, it doesn't really matter.
December 21, 2006
HDV is dead, long live HDV
There have been a few articles in the past week stating that the end of HDV is near. This article in particular goes into depth about the emerging intraframe formats which are vying for the low-end professional market.
I agree that AVC-Intra (note: AVC could be either inter- or intra-frame, don't assume!) and JPEG2000 are good options for compression going forward. They're both good steps forward, but I don't think they're HDV competitors. Let me explain.
At this point there are no "professional" HDV cameras on the market. The closest you get are the XDCamHD products from Sony, which are more or less HDV wrapped in MXF going onto an optical disc, with the ability to bump up the bitrate a little bit. All the other HDV cameras on the market are, in my opinion at least, consumer or pro-sumer level. JVC might argue a bit, but they're JVC so who cares?
The products being discussed in the Nordahl article are most closely related to the XDcamHD products. Neither AVC-Intra nor JPEG2000 are particularly well suited to tape based storage, at least miniDV style cassettes. The Panasonic and Grass Valley (respectively) cameras instead make use of different direct-to-disk recording options. AVC-Intra is just an i-frame-only version of H264, which itself is just a further development of the technology from MPEG-2. JPEG2000 uses wavelet compression and could be pretty impressive. I've never used it in production. Both are solid choices for higher-bitrate recording.
I don't think HDV is going anywhere soon. Getting away from LongGOP compression is a good idea in the long run, but for the low end of the market I think it'll have to wait until flash memory becomes much larger and much cheaper. I'm a firm believer that it will be difficult to penetrate the low end of the market without the ability to easily swap media in the field, without lugging along a laptop. That means being able to carry a pocket full of flash cards, preferably of some variety that can be purchased at a local Target or Best Buy when you're shooting in the field. P2 is a start down that path, but I think widespread adoption is still a ways off. Products like the Firestore are just bridging the gap until we can have proper direct-to-memory capture.
From a technical standpoint, I think we're just starting to see what HDV can do. For example, read Steve Mullen's article on smart GOP splicing. If you can avoid the generational issues of reencoding HDV, the remaining issues are based on processing speed. I wouldn't be surprised to see realtime HDV output over firewire in the next version of Final Cut.
Here's my predictions for recording formats in 2007 and 2008. For cameras under $10,000, HDV will remain the dominant force. The HVX-200 and successors will continue to embrace P2, but I think it'll be 2009 or 2010 before P2 is really practical in all situations.
For cameras from $20,000 - $50,000, you'll see a few formats. AVC-Intra will replace DVCProHD, as NLEs add support for that format. AVC-Intra has major benefits over DVCProHD with no downsides. XDCamHD will stick with the current setup through 2007, but in 2008 I'm expecting to see a higher bitrate recording system (XDCamHD2 or some such) which will add a non-GOP recording format. It'd be nice if it was JPEG2000, but I'm guessing it'll be SONY2000 or something stupid and proprietary like that. JVC will continue to push ProHD long past its sell-by date, and the HD100 will become a faded memory.
Above $50,000, I think what we've got today is pretty much where we'll stay, with the exception of DVCProHD being replaced by AVC-Intra. HDCam, HDCam-SR, D5 and the other "big tape" formats are with us for the long haul at this point. Various direct-to-crazy-raid turnkey solutions will probably begin to creep into this market space as well, but it'll be a slow process.
And me? I'll just go back to shooting Hi8.
November 8, 2006
... "Why I don't care about AVCHD but maybe you should" ...
Panasonic has released a couple of new AVCHD-based camcorders, the HDC-SD1 and the HDC-DX1. The SD1 shoots to SD card, the DX1 shoots to 8mm DVDs.
Both of these cameras record in AVCHD, a format which appears to be gaining some traction in the consumer space. What is it? Essentially it's an H264-based recording format targeted at folks who don't have any intention of serious post production. Because H264 is not only long-GOP but also bi-directionally predicted, cutting it in an NLE is relatively problematic. But then, how many soccer-moms or nascar-dads are cutting their video of little Jimmy's baseball game? Not that many.
Anyways, I haven't paid much attention to these devices because I'm not convinced they're worth the trouble. For most people, digital cameras are quickly replacing separate camcorder devices, because the 640x480x30fps video that most digicams shoot is "good enough." For those who need a little more, something like the Sanyo Xacti HD1A is probably a more convenient device than a traditional camcorder form factor. I'll have one of those in for review in a few days, so we'll see.
In any case, I think AVCHD is an interesting format which I don't really care about. But perhaps you do.
October 25, 2006
Politics in the Age of YouTube
Seeing as I have a Political Science degree rotting on my shelf, I figured I'd make a big-picture post about my view on the impact of YouTube during this election cycle.
By providing the ability to distribute video without concerns over bandwidth and storage space, YouTube has had two very distinct impacts on the way the "netroots" groups operate during this cycle. While "netroots" is a rather stupid term, in this case I'm using it to refer to the politically aware, internet savvy folks who troll sites ranging from DailyKos to (gulp) FreeRepublic.
I want to look at two cases to show by example. The first is the case in which a candidate is able to audition ads to the entire country, without making an ad buy. They can then use the ad to raise funds. In this way, folks can give money directly in response to what they believe is effective advertising, rather than giving money in the hopes that the candidate applies it wisely. Whether or not this is a good thing for the political media landscape is yet to be decided.
An example is this ad, from the Ohio 2nd (the fighting second!). The Wuslin campaign has made very effective use of YouTube to get their ads out and to solicit feedback. Other groups and candidates are making similar use of YouTube. Unlike in the past, this isn't a case of individuals recording ads from TV and posting them, these ads are being posted by the campaigns themselves. The global availability of these ads, combined with candidate websites and political blogs, has made every race a national race.
The second way in which YouTube has become a force in the election is the ability to quickly and easily share video of politicians being stupid. Before this election, a politician would say something stupid, like "We've never been stay the course," and the website hosting the video would get hammered and knocked offline. But now, when someone says "Macaca," or talks about "the Google" the video can live on and on without the hosting provider pulling the plug. The impact of this shouldn't be discounted.
Will YouTube decide the elections? Definitely not - just like blogs didn't decide the election in 2004. But what it will do is increase interest in candidates who might otherwise not get significant exposure. Thoughts?
September 28, 2006
If only it wasn't Sony
I mean, it's totally useless, but darn it sure does look nice.
Useless for me at least. Perhaps you'll find it to be a world full of happy. It's M2T and AVI only, and likely won't support anything like Canon's 24F or JVC's 24p. So, FCP users are out in the cold.
As I say, I wouldn't give it a second thought if it wasn't such a darn nice looking product. Sigh.
Oh, it's also FAT32. It's time for FAT32 to die. Yes, I know, it's the only truly cross-platform disk format. Frankly, I'd much rather a device that gave me the choice to run the disk in NTFS or HFS+. Generally, when I'm out shooting, I know whether I'll be hooking the drive up to a Mac or a PC. Since these are all using Linux under the hood, it shouldn't be too difficult - Linux can more or less write to both NTFS and HFS these days. Getting rid of the 2gb file limit would be a dream.
Done with that then...
September 6, 2006
Finally, the dreaded August lull is over
Well, things are definitely back in swing here at the University of Minnesota. Students are back on campus, wandering around with doe-eyed faces, while the rest of us complain about how much easier it was to park last week.
In the rest of the world as well, things are heating up. August is traditionally a very slow news month in this industry, and this August was no exception. Hence the lack of posts here. There was literally nothing to talk about.
August is gone now, and September is here. And with it comes news!
First off, Apple released new iMacs and Mac Minis today. The iMac gets the upgraded Merom processor, with 64bit support. They've also added a monster 24" iMac to the line, which even includes Firewire 800. The Mac Mini gets an all-dual lineup, at the same price points.
While this by itself would be good news, even better is what it foretells for next week's Special Event. Obviously there's something Hollywoodish going on with the "It's Showtime" invitation. The fact that they released the new macs today instead of using them as filler for next Tuesday is a good sign that Apple is pretty excited about whatever they have to show off.
Even more interesting is a trademark filing Apple recently made about their iMovie trademark. They've added a few new classes to the trademark, indicating that they're likely going to move the iMovie trademark into some new markets. The new classes deal with online transactions, electronic sales and other items. While this may be purely coincidental, it sure seems that the name iMovie is going to end up being associated with their new movie distribution push.
Furthermore, IBC starts this weekend. Anyone want to fly me to Amsterdam? We can expect more announcements from across the industry.
June 12, 2006
HDV error correction
I've been helping out on a shoot using our XL-H1s over the past few weeks, and one thing I've noticed is that using Sony PR tape stock, we're seeing roughly one dropout per tape. This isn't so bad, and we're recording to harddisk as well, but it has gotten me thinking about HDV dropouts.
In DV, a dropout might mean you lost a few macroblocks scattered throughout the frame (the macroblocks are even laid to tape in a non-sequential order so that you don't lose a whole chunk of image to a dropout - clever!). In HDV, you lose 15 entire frames. I'm very curious to know whether this is due to technical necessity, or a choice on the part of engineers. It seems to me that if you handled an HDV dropout the same way you handle a DV dropout, the effect would somewhat similar. However, since HDV is interframe within the GOP, you'd end up with a dropout artifact that potentially travelled around the screen for 15 frames. Have they chosen to just not display the whole chunk of frames because it would look worse to display the artifact?
A dropout on an HDV recording will also knock out more picture information (since there's more picture data per byte) than a similar dropout in DV, but I still can't grasp why we need to lose a whole 15 frames. Even if the dropout hit on the I-Frame of a GOP, you can still often do a passable job reconstruction a frame using P-frames, especially if you could look back to the previous GOP.
So, I want better error correction in HDV. Make it so.
April 26, 2006
Here's what I considered the possibilities for what RED is all about:
1) It's a total scam, just for fun. (unlikely)
2) They're serious, but will miss their goal time.
3) They're serious, but will under-deliver.
4) They're doing this in the hopes of being bought out dotcom-style.
One need only pause for a second to get a sense of the scale of their problem. At 4k, 12bit, 4:4:4, 60p, your data rate is in excess of 2gigabytes (BYTES!) per second. A duallink, 4gig fiberchannel connection tops out around 1gigabyte/sec. The fastest RAID array I saw at the show, which had 42 drives in it, could sustain 750 megabytes/sec.
So, they'd need an interface that could push that much, and an array that could write that much. Even Infiniband can't do it in a single link connection (I don't know if infiniband does dual link).
They also claim that the sensor will be upgradable to higher resolutions in the future. That would imply that the backplane and everything else internally can handle even HIGHER datarates. I'm just not convinced - though I always leave open the possibility that I'll be proven wrong...
April 19, 2006
Dear On2: Please join us here in 2006
Let me preface this by saying that On2 has been a consistently friendly and helpful company when contacted with questions. That said, their FlixExporter Quicktime plugin, for generating flash videos within Quicktime, is really 1990s. The biggest problem is that it isn't multiprocessor aware - it'll only peg one chip at 100%.
Second, the interface has a major heaping of ugly.
Third, and perhaps not On2's fault, the exporter cannot be used to distribute compression across compressor nodes. This may be a limitation imposed on all export components, so perhaps I should be angry at Apple.
Anyways, seeing as I've just been wrapped up in Flash video all day, I figured I'd make a post about it ...
February 23, 2006
Focus Enhancements Makes Me Sad
Focus Enhancements really disappoints me. It's a company with a really good product, in the Firestore line. Maybe not the most polished interface, but it gets the job done in a usable form factor. But somehow the company is still managing to destroy itself.
We recently ordered 3 80gig FS-4Pro units. We got shipped PAL units the first time, and 40gig units the second time. Still waiting to see what happens the third time. I've had similarly disappointing experiences in the past.
I really get the feeling that it's a company suffering due to a lack of direction and focus (ironically). They make semiconductors, digital signs, asset management tools, portable harddisk recorders, etc. The company is now at risk of delisting from the NASDAQ. It just makes me very sad, because it's a company that deserves to prosper. They've got the right product at the right time, but somehow it's still being botched.
Enough ranting, I just feel like they're blowing it and it makes me sad...
February 7, 2006
The Continuing Tape Conspiracy?
I know that every field has its own set of neuroses and urban legends. Video in particular seems to have quite a few, many stemming from the transition from analog to digital video. Worrying about horizontal blanking on SDI signals so the wide variety of mysticism surrounding square waves. Videotape is another area in which a few bits of anecdotal evidence have spawned whole new ways of behaving.
More crazy conspiracy theories after the jump.
The newest cult surrounds HDV tape. Pretty much everything about the way HDV goes to tape is identical to DV. All that's different is the content of the data. Because of the higher compression ratio of HDV, there's more picture information packed into any given chunk of tape, so a dropout is more likely to cause problems. Similarly, because of the 15-frame GOP structure of HDV, the problem has the potential of spanning multiple frames.
Somehow that fact has been transformed into this idea that a single dropout will destroy the entire 15 frames, and could in fact lead to the end of the world. Therefore, it's critical that you drop $15 on the special dropout-resistant HDV tapes. I'm not convinced.
Similarly, there's the whole issue of tape lubricants (another whole rant), never mixing brands of tape in a deck, etc. I think, in the indie world at least, this sort of took off around the time of the DVX-100, which was a fairly drop-out prone camera. It's a fine rule to have, but I don't think it needs to be religion.
Anyways, I just felt like ranting. From the perspective of students, and many indie filmmakers, cost beats all other concerns. Plus, tape is dead, right?
February 2, 2006
I'll have a lot more to say about this camera in a few weeks. We've now officially ordered three of them for the Studios, along with Firestore disk recorders and some assorted support equipment. But, I just wanted to post a few things that I found particularly interesting when playing with the loaner.
Here's my laundry list of things that I hadn't found in other online reviews:
- HD-SDI output is really quite good. I ran it in to our Videotek VTM-440 HD/SD-SDI scope and the signal looked excellent, in both HD and SD modes. The internal color bar output is accurate as best I could see.
- The tactile feel of the camera is excellent. Very positive feel on all the buttons and knobs, and it feels very well put together.
- Like all XLs, it's rather front heavy. Having the integrated shoulder mount is nice, but it's not all that useful because of the weight distribution.
- The composite out gives you a downconverted SD signal, even when the camera is in HD mode. This is nice for feeding an SD field monitor. Canon support said this wasn't the case. Additionally, you can chose to output an exact mirror of what's on the viewfinder, including alerts and everything.
- The neutral density roll-in on the lens is a nice, physical motion. I appreciate that.
- The focus and zoom rings on the lens have the typical lag that comes with non-mechanical rings. Not terrible, but if there was an option of getting a manual lens, I'd go for it.
- The viewfinder is a bit lower resolution than I might hope for. Luckily, there are conveniently placed viewfinder-zoom and peaking controls, so make focusing a bit easier. Being able to set the viewfinder to black and white is helpful as well.
- I appreciate that you're never forced to go into the menus for basic control stuff. Everything that you could need on a recurring basis has its own knob or button.
- There is a bit of chromatic aberration, especially at the edges of bright objects, but I think it's the sort of thing that mostly gets noticed in blown up still frames. It wasn't distracting when watching any of the footage I shot.
Overall, it's a remarkable camera. It does 99% of things right, and the things that aren't right are non-critical in my opinion. I'm glad we're getting three, and I think it's going to take a lot of work for someone else to top it. Now Canon, build us a deck!
I'm not posting any sample footage yet, as everything I shot was rather boring. I'm also not sure what the point would be unless I posted full 1920x1080 video, and that gets messy.
January 25, 2006
Where hath all the CRTs gone? (Rant)
It's starting to get really hard to get a CRT video monitor. This makes me sad. I've always been a big fan of the Sony PVM series, but they are not long for this world (limited stock is still available). The replacement is Sony's LUMA LCD line. Now, go and read the Luma FAQ, especially the question "Is the quality of the LUMA line as good as the PVM?" - their answer? "No."
So why discontinue the PVMs? Arg.
Apparently they're still making the 16:9 BVM series monitors, though even that is rather ambiguous. I've heard from at least one person that Sony is dropping those as well. This confusion led me to call Sony. Big mistake. Read on to learn how my head exploded.
So, first up, if you call Sony's Sales number, you're greeted with a friendly computer that tells you that all the information in the world can be found on Sony's website. And then hangs up on you.
Now, this is a lie. And, frankly I don't think it's right for people make computers do their lying for them. At least give me a minimum wage call center worker who can lie to me. Everyone on the planet knows that Sony Professional is the worst website in the world. Beyond just being a terrible design, it's a terrible design that contains no real information. Compare that to Sony Europe which has a terrible design which hides tons of useful information.
Anyways, so having been hung up on by the computer at the Sales line, I called the corporate offices. There, a friendly chap pointed me to another number. The person there pointed me to another number. Where, surprise, I was pointed to another number.
And that's where I met Max. Max is one of these new-fangled voice activated phone interfaces. I really don't agree with them having names, because I don't want to be friends with Max. But Max really wanted to help me... buy a DVD player.
Max: "What product are you wondering about?"
Max: "I heard AHT3059XM"
Max: (getting angry) "I really need to know what product you're calling about!"
Me: 0! 0!
Max: (angrier) "PLEASE TELL ME WHAT PRODUCT YOU'RE CALLING ABOUT!"
Me: 0! 00! 0!
Max: "I can't help you! Goodbye!"
Bloody passive aggressive slacker AI program hung up on me. And sadly, that's the most human interaction I've had with anyone from Sony.
And that's when my head exploded.
January 19, 2006
Why are there no good review sites of this stuff?
Part of the inspiration for starting this site was the utter lack of decent review sites for video gear. At best you'll find reviews of cameras, and even then they tend to be "dude in his bedroom shooting out his window" type of stuff.
Case in point, I'm trying to find a decently priced HMI kit. B&H has plenty to chose from. Why isn't there a site I can go to to read reviews? Hear about quality? At best I'm left searching forums looking for answers. Someone should get some proper VC together and start an all-encompasing production gear review site. Anyone want to volunteer?
January 11, 2006
Post MacWorld Roundup
So, another MacWorld has come and gone. I must say, it was a pretty good one. The big news is probably best covered by some of the more Mac-centric sites, but I wanted to post a few things that caught my attention.
First off, the new iLife suite looks pretty sweet (ha ha ha). The most significant new feature in my opinion is the Garage Band podcasting features, integrated with the iWeb software for posting a podcast. While folks at the University of Minnesota have UThink which provides an easy way to post podcasts, other folks aren't so lucky. I think it'll be a big hit.
The other iLife updates are less substantial from my point of view.
For me, the single most exciting bit was the announcement that the Pro Apps (Final Cut, Aperture, Logic) will all be universal binaries by the end of March. This has been one of my big concerns - from what I've heard, the move from OS9 to OSX was rather painful for Final Cut, and I was worried that the port to x86 would be similarly difficult. After talking with some reps from the Pro Apps group a month ago, that fear was heightened by their "we'll be on x86 someday" attitude.
Anybody who still has an older version of Final Cut will also be able to take advantage of the transition to Intel to upgrade to Final Cut Studio at a reduced rate as well. It remains to be seen whether the upgrade will apply to academic discount versions, but we can hope.
It's exciting to know that the next generation of the platform is now in the wild. It's an exciting time. Once the iMacs start arriving in the hands of end users, there will be much more to write about.
January 6, 2006
Rumor Rumor on the Wall (42" or 50"?)
For true Macintosh aficionados, the next 3 days are just about the most exciting time of the year. Only 3 days, 18 hours and 8 minutes until the keynote address at Macworld San Francisco! There are some interesting rumors this year. I'm expecting Intel iBooks and Mac Minis, some new form of living room integration and updates to iLife and iWork. If we got all that, it'd be very exciting. Anything more and I may just end up in a pool of my own drool...
January 5, 2006
"We don't want HDV because it's not really High Def"
How many times have you heard that? I'm not sure where all the FUD has come from regarding HDV but it's starting to get on my nerves. Is HDV the perfect format? Definitely not! Does it get 1080i into a 25mbps datastream with surprisingly good results? Sure does!
I wasn't really around when the transition to DV happened - were the same conversations happening? Did people say "We don't want DV because it's not really Standard Def"?
For people who aren't clear: HDV can support up to 1080i, 60 fields per second, 8 bits per pixel, 4:2:0 sampling. Is it as good as HDCamSR? Nope. But DV isn't as good as DigiBeta either. That doesn't make it any less valid a format.
If anything, I think HDV makes it even more clear that the import issue is not what technology you use, but rather how you use it. Home movies shot with HDV will still look like home movies.
January 4, 2006
Is the Canon XL-H1 the Holy Grail?
So, in my post talking about what I hope to see at NAB, I mentioned that I'm hoping to see second generation HDV devices. That raises the question of what a second gen HDV device will actually be - are we there already?
We're certainly a lot closer. A year ago, we had the Sony FX1 and Z1U and that was pretty much it. We've now got the Sony A1U and HC1, the JVC GY-HD100, the Canon XL-H1 and the Panasonic HVX-200. Additional choice is never a bad thing. I want to use this post to go through a basic summary of my thoughts on the various cameras out there. This is not intended as an indepth review - those will come later. Just wanted to give a brief overview of the landscape as it exists today.
First, I think that the Sony A1U is a great camera for the market segment that it addresses. The CLA-TV Studios currently send out Sony PDX-10s are our primary prosumer camera. When we move to HD, the A1U seems like a clear successor. Similar form factor, XLR audio and a pretty decent image. I think that it gets largely ignored by camera-snobs for having a single chip. While it'd certainly be nice if it was a 3ccd camera, you can't complain about the price. If you had about $2500 to spend on a camera right now, it's almost a no-brainer.
The JVC GY-HD100 disappoints me. It had such great potential when it was announced last April - A nice form factor, interchangeable lenses, proper progressive chips - the whole package! Unfortunately, it's a very flawed camera. The quality of the images in general doesn't impress me, and I have difficulty accepting the reasoning behind the ProHD format. Why start splintering a format so early in the game? I'm especially upset that, despite breaking their compatibility with every other piece of gear, JVC still isn't using the full bandwidth available on the HDV tape when doing 24p. If the whole reasoning behind ProHD is to avoid having to do a pulldown to 60i/30p, why waste 20% of your available bandwidth? Give us a less severe compression, or a more redundant recording structure - something! Additionally, I wonder about the reasons behind the slow adoption of HD100 support by Apple in Final Cut Pro (and Avid too?). Is it a purely technical situation, or are they hoping the format will be stillborn?
The Panasonic HVX-200 is in a bit of a different class from these HDV cameras. If I were an indie-filmmaker with some decent cash looking to shoot a feature, this is probably the camera I'd go with. The benefits of DVCProHD are enormous, both in terms of quality and flexibility, and the drawbacks of P2 can be avoided if you've got control over the whole shooting process. Many people were scared off by the idea of expensive memory cards and the limited recording length. I firmly believe that with a little bit of planning, these problems can be minimized. When the Firestore FS-100 ships (my Focus Enhancements rant will get its own post), things will be even better. While I'll reserve final judgement until I see some more footage, the flexibility offered by this camera is unmatched. The footage that Kaku posted really blew me away. I'm a big fan of the DVX-100 though, so I may be a bit biased.
Finally then, we've got the Canon XL-H1, and we can revisit the question posed in the title of this post. First off, I have to admit that I've never been a big fan of the XL series. I don't like the form factor, I don't particularly like the image tonality, the audio preamps sound terrible, and the XL1 and XL1S especially just never seemed like great value for money. Because of this, I was even more surprised to have the XL-H1 be so impressive. We get 1080i, 1080"p", 720"p", plus standard def. It is important to note the quotes there, because these are not progressive scan CCDs. However, unlike the Sony camera's, all indications are that Canon generates their 24p (24f in their lingo) timing with a proper cadence. This means that when you go to post, you can cut a 24fps timeline without the crazy motion problems of Sony's Cineframe. While they are doing some onboard deinterlacing, it appears that they're getting around 80% of the vertical resolution of a true 1080p image, which is still pretty good - better than the HD100's 720p chips.
What really sold me on the XL-H1 though was the HD-SDI output. To me, this says a number of things about Canon's mentality regarding the camera. First off, it shows you what a manufacturer can do when they don't have a high-priced "pro" business to maintain - Sony wouldn't dare put a similar output on such an inexpensive camera. If someone can come up with a mildly priced HD-SDI DTE box (no, the suitecase sized boxes don't count), this camera will be the absolute king. Maybe they'll throw Triax on some future "XL-H1S" and give us a proper CCU - then we can really watch Sony squirm.
So to answer the question - The Canon's as good as we get right now. They do so much right that I can't imagine anyone beating it soon. I hope that the HVX-200 lives up to expectations as well. Between the two of them I think most people's needs will be covered.
Of course, both of these cameras will run you nearly $10,000 (once you buy P2 media for the Panasonic). If you're an indie filmmaker with only $5000 to spend, you're in a pretty tough spot right now. Do you shoot with the JVC in order to have 24p? Do you shoot with the Sony Z1U and convert to 24p in post? Do you buy a DVX-100B and shoot standard def? Anyone have suggestions?
So, we're in to the New Year now. CES is about to start, Macworld San Francisco is next week, and we're a mere four months away from NAB. This is an exciting time! So, what should we expect?
Macworld San Francisco may end up being the most exciting of the shows in many regards. If we see the first Intel macs next week, it will kick off the transition and make for some exciting times. If there's no Intel news, it will sure disappoint a lot of folks.
This post is really intended to start some discussion about what we expect at NAB. I think that 4K will be the big deal this year. Last year, it was clear that the transition to HD, from a vendor perspective at least, was basically over. There were already a few people beginning to play with realtime playback of 4K material on lower end gear. I expect that this year we'll see 4K being out in full force. There's lots of buzz on the net about the new RED camera. I'm personally unsure whether to believe it or not. The DVInfo Forums have quite a bit of discussion about it, with messages that claim to be from the various folks involved with the project. It sure seems fishy. But if it's for real, it'd be a pretty big deal.
Similarly, Thinksecret is now reporting that Apple will announce Final Cut Pro 6, as well as a new product called Final Cut Extreme. FCP6 is a pretty safe bet - NAB-announced upgrades have been the trend for a few years now. Final Cut Extreme is a bit less clear. First off, how will we abbreviate it? FCE is already taken. FCEx? Second, one of the powerful things about FCP is that it grows with your needs pretty well at this point - from DV to HD and beyond. Adding a third "high end" product would muddy the waters a bit.
I also really hope that we start seeing second generation HDV equipment. There is a huge need for dedicated HDV decks that are more than just stripped down camera tape transports. Give me something like a DSR-45 that can play back HDV and DV, including 720p24 and 1020f24 and the rest! Even better, give me a DSR-1500-style deck that does all of the above, plus has HD-SDI in/out and AES/EBU. Then I'll be in heaven.
Please offer your thoughts on what you think will be coming in the first half of this year!