No! miniDV is NOT a good long term storage format. miniDV is the consumer version of the DV format. Professionals use the thicker, higher quality tape formats of DVCam and DVCPro for long-term archival needs.
Also, no DV formats are designed to be 100% bit safe. A lightweight video codec like DV can handle an off bit here and there without ruining the project. Just a little visual glitch here and there on a single frame, and pro decks do a good job of interpolating the missing data so the problem isn't too obvious. Not really possible with a compressed data bitstream.
Yep, the problem with convergance is that the home electronics start becoming more like computers as well as the computers becoming more like home electronics. A couple of years ago, I got a Dish Network box that ran a version of WinCE, and even had Doom installed (nearly impossible to play with a remote...). Which was all well and good, except that the box would crash randomly, need to download upgrades regularly, and the HD was loud, and eventually failed! The non-MS replacement unit is still working fine, but still.
Meanwhile, all the other electronics just keep on trucking, NEVER doing anything weird. I'm sure I'll still be using my five year old Pioneer amp for years to come, although it'll be depricated from the home theater soon (it only routes S-Video).
A 50-year old TV can watch a broadcast today, and today's TV could have watched a broadcast of 50 years ago. As convergance hits, that cycle will probably drop to 5 years, maybe less.
But, of course, it'll be nice to get 20 channels into the bandwidth where once we could have one.
Indeed. You'd never use WiFi as a multichannel solution, for the reasons you describe.
The article was about doing one local channel. Assuming you got multicast working correctly (no idea what the story with WiFi and multicasting is), multiple viewers can watch the same packets to watch the same stream. This would be more like a local pirate radio solution for video.
As for data rates, I think with the leading codecs today, you could do "good enough" quality (as good as digital cable) for TV display at about 1 Mbps, maybe a little less with some content.
The 19.2 Mbps of bandwidth allocated per channel was the requirement for 1920x1080 60i MPEG-2. WIth modern codecs, we can get a LOT better bang for the bit. In my testing, 1920x1080 60i @ 7Mbps with MPEG-4 Advanced Simple works great. And with the forthcoming H.264/MPEG-4 Part 10, we'll be able to get that lower yet.
For point of reference, when MPEG-2 was originally defined, 7 Mbps was considered about the minimum you'd ever use for standard definition (720x480). The number of bits required per pixel have dropped to about 1/8th in the last decade.
And as long as Moore's law keeps letting us up the MIPS per pixel rapidly, we've still got headroom to design better codecs yet.
No, Apple didn't fail to make you a movie player. They didn't try. I've heard estimates on the order of 100 engineer-years to do a full port of QuickTime to *NIX. The Windows port took over a year more than originally scheduled, and requiried porting huge chunks of the MacOS Toolbox to Windows. Something equivalent would be required for *NIX, with an even bigger moving target. Plus there is LOTS of processor specific optimizations there.
Okay, I can at least make a MPlayer compatible version, by fixing that annoying multitrack and JPEG problem. It's rendering now - figure out a place I can send it to, and it's yours.
Other that.MOV, what format would folks like to have this in? ISMA MPEG-4? MPEG-1?
In futher analysis, this file is probably using Sorenson Video 3, not Photo-JPEG. It's doing something whacky where the first frame is JPEG, but it's appended to the standard video track instead of being the same one.
I'm guessing the file was encoded in Cleaner with the "High Quality First Frame" flag on, and set to JPEG mode. This is a goofy and pointless thing to do with a modern codec, but I've seen it happen. That would also explain why MPlayer playback is failing.
Anyway, this is probaby a Sorenson Video 3.1 Pro 2-pass VBR encode, based on the data rate distribution. Still, they should have used B-frame encoding, and MP3 or AAC-LC audio instead of QDesign. The keyframe insertion threshold was also too sensitive.
So, my 30 MB brag wouldn't work, but I could deliver a lot better audio quality and similar video quality at 100 MB or so.
Well, a reencode from the current file wouldn't help much, since Photo-JPEG is a lightweight codec for playback, and the QDesign would be simply terrible source, since it is so artifacted. They should at least have used the Pro QDesign encoder for what is presumably going to be such a popular file. That allows data rates up to 128 Kbps, instead of this lousy 48, and has better bang for the bit to boot.
If anyone has access to the source, I'd be happy to recompress it pro bono to something better, smaller, and more compatible. MPEG-1 comes to mind, which plays everywhere.
They money they saved by not hiring a compressionist to do this is being lost 10x over in bandwidth costs, I'm sure! This file should have been maybe 30 MB.
JPEG for video? The file is about 10x larger than it would need to be to provide te same quality with, say Sorenson Video 3.
But QDesign Music 2 Basic audio? This is the lame, free, 48 Kbps maximum version of a four year old audio codec. They should have used MPEG-4 AAC-LC, or if they wanted to keep compatibility with QuickTime 4, MP3 audio. Would have sounded MUCH better, without all that nasty phasing.
As it is, they would have been better off just using MPEG-1. That'll play everywhere, and would have provided the same quality at a much smaller file size.
Actually 2:1 is a pretty typical compression ratio for all kinds of digitized media. Still images, audio, you name it. Video can get a little bit better, because of interframe correlation.
But for any given encoder, there is always at least one source that will result in output at least one bit more than the input.
Any given content will have a Shannon Limit, which is the maximum theoretical compressibility. The more entropy/randomness in the content, the higher the Shannon Limit will be. A completly random series of numbers will have a a Shannon Limit of the size of the file - it's impossible to compress*.
*and yes, it's cheating to know what the random number algorithm is and working back to the seed number. When I say "truly random" it means "truly random."
Assuming you're running a 4:3 aspect ratio monitor, your images on the screen are distorted at 1280x1024. 1280x960 is the proper square pixel resolution for a 4:3 monitor.
If you did that, you also wouldn't have those black bars.
Indeed. Skins support seems to mean there will be a choice of dozens of whacky looking interfaces with crappy usability. I don't mind them being an option, but I want the software to inherit the native UI of the system it's running on!
MPEG4IP on RedHat 8.0 does the whole Blue thing, which means it is doing the right thing today.
While it doesn't play all the bazillion formats of MPlayer, the CIsco-sponsored MPEG4IP has quite a nice little UI, compared to all the others. And it's just one package to compile, none of these source code + codec packs + skin + font downloads.
MPEG4IP only does MPEG-4, but since that is almost certain to be come the standard "open" video format, I'd expect it to become more and more useful throughout 2003. We're testing Cringely's NerdTV against it, and it works just fine.
There is a common tendency to think one's political opponents are motivated by evil. While a lot of Bush's actions are to me inexplicable, experience has forced me to assume he's merely wrong, not malicious.
Let's have a little relativism here. I didn't vote for Bush, nor do I care for him as a president. But before you start with the whole "so corrupt and broken" could you give a comparaison? Perhaps a past US government? The current one is certainly much less corrupt than any one of the 19th century. Compared to other modern countries?
And which kind of corruption are you thinking about? License Raj? Widespread bribary?
Things certainly should be better, but our current government is better than those that at least 95% of humans have lived under.
Speaking as a respectibly left wing Democrat, I'd vote for Jobs over Nader in a heartbeat. Nader has simply shown zero ability for that kind of a job. When he says there is no difference between Democrats and Republicans, he's either lying, or an idiot. Who is president MATTERS, no matter what he says.
Leadership requires the right mix of idealism and pragmatism, and Nader badly fails that test. If he actually WON the presidency, he'd be disasterous at it. And since even he knows that he isn't going to win, running mainly makes him just the Perot-of-the-left, working as a spoiler to get Bush reelected.
Slashdot Mirror
DV isn't exactly JPEG or Motion JPEG, although it's close. It is DCT based, though. MJPEG requires higher data rates to provide equivalent quality.
No! miniDV is NOT a good long term storage format. miniDV is the consumer version of the DV format. Professionals use the thicker, higher quality tape formats of DVCam and DVCPro for long-term archival needs.
Also, no DV formats are designed to be 100% bit safe. A lightweight video codec like DV can handle an off bit here and there without ruining the project. Just a little visual glitch here and there on a single frame, and pro decks do a good job of interpolating the missing data so the problem isn't too obvious. Not really possible with a compressed data bitstream.
It's Moore's law. Every year is the best year ever for consumers for computers.
All praise Saint Gordon!
Yep, the problem with convergance is that the home electronics start becoming more like computers as well as the computers becoming more like home electronics. A couple of years ago, I got a Dish Network box that ran a version of WinCE, and even had Doom installed (nearly impossible to play with a remote...). Which was all well and good, except that the box would crash randomly, need to download upgrades regularly, and the HD was loud, and eventually failed! The non-MS replacement unit is still working fine, but still.
Meanwhile, all the other electronics just keep on trucking, NEVER doing anything weird. I'm sure I'll still be using my five year old Pioneer amp for years to come, although it'll be depricated from the home theater soon (it only routes S-Video).
A 50-year old TV can watch a broadcast today, and today's TV could have watched a broadcast of 50 years ago. As convergance hits, that cycle will probably drop to 5 years, maybe less.
But, of course, it'll be nice to get 20 channels into the bandwidth where once we could have one.
Indeed. You'd never use WiFi as a multichannel solution, for the reasons you describe.
The article was about doing one local channel. Assuming you got multicast working correctly (no idea what the story with WiFi and multicasting is), multiple viewers can watch the same packets to watch the same stream. This would be more like a local pirate radio solution for video.
As for data rates, I think with the leading codecs today, you could do "good enough" quality (as good as digital cable) for TV display at about 1 Mbps, maybe a little less with some content.
And to follow up on the implicit 1),
The 19.2 Mbps of bandwidth allocated per channel was the requirement for 1920x1080 60i MPEG-2. WIth modern codecs, we can get a LOT better bang for the bit. In my testing, 1920x1080 60i @ 7Mbps with MPEG-4 Advanced Simple works great. And with the forthcoming H.264/MPEG-4 Part 10, we'll be able to get that lower yet.
For point of reference, when MPEG-2 was originally defined, 7 Mbps was considered about the minimum you'd ever use for standard definition (720x480). The number of bits required per pixel have dropped to about 1/8th in the last decade.
And as long as Moore's law keeps letting us up the MIPS per pixel rapidly, we've still got headroom to design better codecs yet.
Yeah, QuickTime's AVI support hasn't every been THAT good. The large file thing was fixed in 6.1, which is only out for MacOS X at this point.
And cp is an encoder how?
No, Apple didn't fail to make you a movie player. They didn't try. I've heard estimates on the order of 100 engineer-years to do a full port of QuickTime to *NIX. The Windows port took over a year more than originally scheduled, and requiried porting huge chunks of the MacOS Toolbox to Windows. Something equivalent would be required for *NIX, with an even bigger moving target. Plus there is LOTS of processor specific optimizations there.
And how would that help Apple?
The 4GB file limit was fixed in Windows back with NTSF and NT workstation. You're probably runnine ME, or have a FAT32 formatted drive.
Actually, QuickTime the file format is completely documented and open.
Apple's implementation is propritary, as are some of the codecs. But as a file format, it is radically better than AVI for doing media authoring.
An open source implementation would be good forever.
Drop me an email privately, and I'll hook you up in a few hours.
Okay, I can at least make a MPlayer compatible version, by fixing that annoying multitrack and JPEG problem. It's rendering now - figure out a place I can send it to, and it's yours.
.MOV, what format would folks like to have this in? ISMA MPEG-4? MPEG-1?
Other that
In futher analysis, this file is probably using Sorenson Video 3, not Photo-JPEG. It's doing something whacky where the first frame is JPEG, but it's appended to the standard video track instead of being the same one.
I'm guessing the file was encoded in Cleaner with the "High Quality First Frame" flag on, and set to JPEG mode. This is a goofy and pointless thing to do with a modern codec, but I've seen it happen. That would also explain why MPlayer playback is failing.
Anyway, this is probaby a Sorenson Video 3.1 Pro 2-pass VBR encode, based on the data rate distribution. Still, they should have used B-frame encoding, and MP3 or AAC-LC audio instead of QDesign. The keyframe insertion threshold was also too sensitive.
So, my 30 MB brag wouldn't work, but I could deliver a lot better audio quality and similar video quality at 100 MB or so.
Well, a reencode from the current file wouldn't help much, since Photo-JPEG is a lightweight codec for playback, and the QDesign would be simply terrible source, since it is so artifacted. They should at least have used the Pro QDesign encoder for what is presumably going to be such a popular file. That allows data rates up to 128 Kbps, instead of this lousy 48, and has better bang for the bit to boot.
If anyone has access to the source, I'd be happy to recompress it pro bono to something better, smaller, and more compatible. MPEG-1 comes to mind, which plays everywhere.
They money they saved by not hiring a compressionist to do this is being lost 10x over in bandwidth costs, I'm sure! This file should have been maybe 30 MB.
Okay, who encoded this thing?
JPEG for video? The file is about 10x larger than it would need to be to provide te same quality with, say Sorenson Video 3.
But QDesign Music 2 Basic audio? This is the lame, free, 48 Kbps maximum version of a four year old audio codec. They should have used MPEG-4 AAC-LC, or if they wanted to keep compatibility with QuickTime 4, MP3 audio. Would have sounded MUCH better, without all that nasty phasing.
As it is, they would have been better off just using MPEG-1. That'll play everywhere, and would have provided the same quality at a much smaller file size.
Actually 2:1 is a pretty typical compression ratio for all kinds of digitized media. Still images, audio, you name it. Video can get a little bit better, because of interframe correlation.
But for any given encoder, there is always at least one source that will result in output at least one bit more than the input.
Any given content will have a Shannon Limit, which is the maximum theoretical compressibility. The more entropy/randomness in the content, the higher the Shannon Limit will be. A completly random series of numbers will have a a Shannon Limit of the size of the file - it's impossible to compress*.
*and yes, it's cheating to know what the random number algorithm is and working back to the seed number. When I say "truly random" it means "truly random."
Assuming you're running a 4:3 aspect ratio monitor, your images on the screen are distorted at 1280x1024. 1280x960 is the proper square pixel resolution for a 4:3 monitor.
If you did that, you also wouldn't have those black bars.
Indeed. Skins support seems to mean there will be a choice of dozens of whacky looking interfaces with crappy usability. I don't mind them being an option, but I want the software to inherit the native UI of the system it's running on!
MPEG4IP on RedHat 8.0 does the whole Blue thing, which means it is doing the right thing today.
Folks,
While it doesn't play all the bazillion formats of MPlayer, the CIsco-sponsored MPEG4IP has quite a nice little UI, compared to all the others. And it's just one package to compile, none of these source code + codec packs + skin + font downloads.
MPEG4IP only does MPEG-4, but since that is almost certain to be come the standard "open" video format, I'd expect it to become more and more useful throughout 2003. We're testing Cringely's NerdTV against it, and it works just fine.
http://mpeg4ip.sf.net
Yeah, this would be my fear as well. Although running a big public company well implies he's learned at least SOMETHING about compromise.
If he decided to run (which he probably won't - who would he let run Apple?) whether or not he could pull this off would quicky become apparent.
There is a common tendency to think one's political opponents are motivated by evil. While a lot of Bush's actions are to me inexplicable, experience has forced me to assume he's merely wrong, not malicious.
This is good, because wrong we can work with.
Ah, you've mistaken cynicism with insight.
Let's have a little relativism here. I didn't vote for Bush, nor do I care for him as a president. But before you start with the whole "so corrupt and broken" could you give a comparaison? Perhaps a past US government? The current one is certainly much less corrupt than any one of the 19th century. Compared to other modern countries?
And which kind of corruption are you thinking about? License Raj? Widespread bribary?
Things certainly should be better, but our current government is better than those that at least 95% of humans have lived under.
Speaking as a respectibly left wing Democrat, I'd vote for Jobs over Nader in a heartbeat. Nader has simply shown zero ability for that kind of a job. When he says there is no difference between Democrats and Republicans, he's either lying, or an idiot. Who is president MATTERS, no matter what he says.
Leadership requires the right mix of idealism and pragmatism, and Nader badly fails that test. If he actually WON the presidency, he'd be disasterous at it. And since even he knows that he isn't going to win, running mainly makes him just the Perot-of-the-left, working as a spoiler to get Bush reelected.
Josh Goldberg, as I live and breath.
I've still got your copy of "Modern Primatives!" That was, what, ten years ago?
Anyway, I won't argue that Jobs was a jerk, but given what he's been able to accomplish at Apple, he's clearly learned a lot since those days.