In MPEG-2, any particular implementation is defined by a Profile and a Level. A profile defines the tools available to a particular bitstream (like interlaced video, or B-frames), and levels define the parameters of those tools (like maximum resolution and data rate).
For an encoder Profile@Level defines the maximum features you can use to make a compliant bitstream (so it's okay not to use all the tools and still claim to make an Advanced Simple stream - the encoder is compatible, just not maximum quality). For a decoder, Profile@Level defines the minimum level of support - it's always okay to play back illegal streams, but at a minimum all legal streams of the targeted Profile@Level need to be supported.
So our problem is that.mp4 is extremely overloaded, since it doesn't indicate the Profile@Level used for any of its parts. So it's easy to make a file that won't play in the player that automatically takes.mp4 when a file is opened.
Actually, MPEG-1 also supports aspect ratio switching. Not all players do, but most of the modern software ones do.
While all of the above are important for digital broadcasting applications, for storing progressive-scan content, MPEG-1 is pretty much identically good, and players are much more widely distributed (it still costs $2.50 to distribute a MPEG-2 decoder legally).
MPEG-1 playback will certainly will be around until the end of time. MPEG-4 will probably be in the same position in a few years. I'd probably stick to ISMA complaint, using Simple Visual for video and AAC-LC mono or stereo for audio. Those files will play in QuickTime, RealOne (on windows), MPEG4IP, and other players today, and in the future. And MPEG4IP is even open source.
I've done a lot of Bink encoding (Bink is featured in one of the tutorials in my book).
Bink's adavantages are a really mature, widely ported (lots of consoles, Mac, Windows, Linux), fast, and highly-featured API. However, its internal codec isn't really THAT great in terms of compression efficiency (bang for the bit). It's fine for CD-ROM games where having a rich, fast, stable video playback system is more important than keeping space to a minimum. But I certainly wouldn't use it for content to be downloaded from the web.
He should have used PNG, which is lossless RGB 24-bit.
Of course, encoding in JPEG @ 100% quality would have been effectively as good. There really isn't any excuse for having still image artifacts on top of the source artifacts in this kind of article.
I had a Apple//c with an amber monitor (which my parents got since I was supposed to be using it for school work, not games). Had a floppy with some games on it (amazing how much fun could go in 140K!). Went through a period where I was playing Joust and Mario Brothers (no Super back then - OLD school platformer). Joust I had pretty well mastered, but I never got the hang of the extra-icy late levels in Mario Bros. (around level 18?). Very frustrating - had to play quite a while to get to that point, and then I'd lose all my lives in only a couple of minutes. I'd get frustrated, play Joust for a while, then go back to M.B.. Rinse, repeat, for hours, late at night.
So, whenever playing a lot of a game, I'd always start imagining playing it in my head while bored. I'd have the physics down cold, so it actually could be decent practice.
Then, one day on the bus home from school, the two games merged in my mind. I was playing Mario Brothers levels, but with the Joust bird. But the physics just didn't match up, so I was constantly getting killed. And I couldn't stop thinking about it - all day, stuck in my head, the darn bird trying to bang the darn coins, and getting darned killed. All in glowing amber phosphers. Kind of like being in hell for the really, really lame.
Stopped playing both games at that point, and haven't booted up either in, what, 16 years?
So, basically a series-long "Yesterday's Enterprise?"
Or, perhaps, they're busily erasing the history of the past shows, to leave everything wide open? I've felt that the temporal cold war was in some way a result of the events in the 24th century and later.
So, I guess the question is what's happening in Enterprise the cause or the result of the past series. If it's the result, who the heck knows what's going to going to happen, and we can stop worrying about continuity.
Maybe I just don't get out enough to know what people are listening too these days. I had thought all three of those were pretty mainstream, cover-of-Rolling-Stone groups.
Is everyone in college off listening to Goa or something now?
Yep. One great thing about digital distribution is that an album can sell 500 copies and still be profitable! In the long term, this will be good for niches,
For another example, check out CustomFlix, who does on-demand DVD replication and distibution. I've made a tidy bundle selling the DVD-R supplement for my book through them. It hasn't sold anywhere near the 500 copy minimum that a mass-market duplication would have required, but I started netting a profit from them after selling the first SIX copies.
So, I've got 14K songs in my MP3-based iTunes library, replicating my CD collection. I'm a mini-van driving work-from-home 32 year old married father of two who wasn't cool in the day, and certainly am not cool now. I don't have a tattoo, ride a Vespa, or compile kernels from source.
But, to my surprise, I'm revealed to be an elitist indy music prig! To test the new service, I sorted my iTunes library by play count so I could compare the quality of my current rips with the new service's previews.
And darn it off if pretty much all the music I listen to isn't from one of the majors! In order of "play count":
Sleater-Kinney - nope The Hives - yep (one album) The White Stripes - nope Husker Du - nope Man or Astroman - no Len - just three tracks Rancid - no Veruca Salt - yep (two albums) The Clash - pretty much everything Riverdales - no Screeching Weasel - no Beastie Boys - no (an on a major, I thought) Cast of Buffy the Vampire Slayer - nope! Fugazi - no Renegade Soundwave - no Gang of Four - no Luscious Jackson - yep Pop Will Eat Itself - just one album Tricky - lots of stuff X - just one album
So, of my top twenty-by-listening bands (which is surprisingly different from what i would have guessed they would be), only 4 have a substantial body of their work available among the 200,000 tracks available through iTunes.
Blockbuster+DVD works well because you can get 18 GB in 20 minutes. It's a high latency connection, but tons of bandwidth!
While broadband connections and modern codecs are getting good enough to do a reasonable quality standard definition broadcast, we're not too far from blue-laser HD DVD.
However, the number of bits for "good enough" audio is a LOT lower than for video. For any reasonable extrapolation of high-end stereo system, (like a full home 7.1 system that can usefully deal with dynamic ranges and frequcies well beyond what speakers can do), we can already get to the "so good golden ears can't hear the difference" with a typical broadband connection. For people with typical systems, even 48 Kbps with some modern codecs like HE AAC has been shown to be as good as CD.
Now, for video, things get trickier. I doubt 8 Mbps will be enough for percepturally lossless 1920x1080 in a real-time stream, even with WMV9 or AVC.
And in fact the QuickTime AAC encoder got a LOT better in QuickTime 6.1. It added a broader range of sample rates that could be used at a given data rates, and both faster (but lower quality) and slower (but higher quality) modes.
A study like this is comparing codecs, not bitstreams. While a good MP3 encoder will certainly beat a poor AAC encoder, the more powerful tools availble in the AAC bitstream will enable it to provide better compression efficiency with encoders of equal quality.
Re:How to convert from MP3 to AAC with iTunes 4?
on
AAC vs. OGG vs. MP3
·
· Score: 1
Ah! Just the trick.
Next question - where does the metadata live one they're ripped? In the.m4a, or the library, or what?
How to convert from MP3 to AAC with iTunes 4?
on
AAC vs. OGG vs. MP3
·
· Score: 2, Interesting
In my workflow, I want to keep a big bunch of high data rate files on the home server (about 140 GB of 320 Kbps MP3 files), and then recompress to more portable formats to carry around on the PowerBook or whatever. This used to work fine. I'd use the Import feature of iTunes, and would convert from the 320 Kbps master file to ~150 Kbps VBR MP3 files for the road. While the lower data rates wouldn't work on my home Paradigm speakers, they were fine for listening to on airplanes.
However, this doesn't seem to work in iTunes 4. I see the Import option, but all the MP3 files in my current library are grayed out. Is this operator error, or does this not work anymore? If not, what is the Import function for?
Obviously I'd like to switch to 128 Kbps AAC-LC for my mobile music. But heck, I'd live with being able to make my old MP3 files!
While I doubt iLife makes a lot of money, the professional apps like Final Cut Pro absolutely do. $1000 per boxed copy - with a higher gross margin than most of Apple's computers I'd warrant.
Although with Apple's pricing strategy for Shake ($5K for Mac w/ free network rendering, $10K for Linux, plus ~3K per rendering node), they're definitely trying to sell more Xserves there right now.
I had a T68i via ATT for a week, and returned it for a Motorola T720. the T68i was a great phone in pretty much every way except that its signal strength simply wasn't adequate in lots of places I need to be (Portland airport, San Jose). The T720 works great for me in places where the T68i would cut out constantly. I do miss the Bluetooth, though.
I looked at the new Nokias, but I just couldn't hack the weird number pad. Yeesh.
It likely would. Although I haven't seen much in the way of RTSP/multicast OGG content out there. I'm not sure if there is any technical reason for this. Experts seem to think that OGG isn't very well designed as a RTSP format, but I don't understand those particular issues in enough depth to state what the problem there is.
No multicasting No native RTSP support No good loss recovery mechanism
It's amazing that MP3 works as well as it does for pseudo-streaming, but a true streaming format it ain't. Personally, I'd like to see them adopt a MPEG-4 AAC-LC stream, which QuickTime, RealOne for Windows, and other ISMA MPEG-4 compliant players could tune into. Better quality at lower data rates than MP3.
Pretty soon we'll have AAC High Effeciency, which can do ~FM quality at 32 Kbps for 44.1 stereo. Astonishingly better than other propritary codecs in head-to-head at these low bitrates.
I'm actually in the process of designing a digital cinema system along these lines, so here's a few comments
Everyone is focusing on the codec being the quality limitation, but that's not true. In fact, the projector is the biggest deal. There are plenty of modern codecs that can give you visually lossless quality if you throw enough bits at them. The issue with codecs is getting compression efficiency up so that transmission and storage is cheaper, and keeping decode complexity down so you don't need to have expensive hardware in the projector. The WM9 system is pretty much a high end (but not the highest end) Dell workstation, strapped to a cart with XLR audio out, a control pad, and a big data projector on the top. All off the shelf parts, which makes implementation cheap, and upgrading the computer very cheap. But those are nice things to have, but not strictly required for digital projection.
But we could do the same thing with MPEG-4, or other formats. WM9 has a more mature DRM solution and some other advantages, but it is absolutely possible to use another format.
The big limit is in having a projector that is bright enough to fill the room, with a dark black, and high resolution. Moore's law gives us improvements in compression faster than we get improvements in projection, so the big photon cannon will be the true limit on quality for a while.
Your description is more like an API. The conformance point of MPEG-4 is a bitstream. Any encoder that makes a legal bitstream is legal, and any player than can take any legal bitstream is legal. Within that, players and encoders are limited by Profiles and Levels, which define the tools, and the maximum paramters of the tools, that can be used.
WMV9 is derived form a draft standard of MPEG-4, but isn't MPEG-4 in any meaningful (=interoperable) way. It isn't "just" a propritary codec, but is a fabulously good one.
ALL delivery video codecs use I frames plus P frames. Some even use B-frames. But that's like saying XML is just a derivative of ASCII.
Slashdot Mirror
In MPEG-2, any particular implementation is defined by a Profile and a Level. A profile defines the tools available to a particular bitstream (like interlaced video, or B-frames), and levels define the parameters of those tools (like maximum resolution and data rate).
.mp4 is extremely overloaded, since it doesn't indicate the Profile@Level used for any of its parts. So it's easy to make a file that won't play in the player that automatically takes .mp4 when a file is opened.
For an encoder Profile@Level defines the maximum features you can use to make a compliant bitstream (so it's okay not to use all the tools and still claim to make an Advanced Simple stream - the encoder is compatible, just not maximum quality). For a decoder, Profile@Level defines the minimum level of support - it's always okay to play back illegal streams, but at a minimum all legal streams of the targeted Profile@Level need to be supported.
So our problem is that
Actually, MPEG-1 also supports aspect ratio switching. Not all players do, but most of the modern software ones do.
While all of the above are important for digital broadcasting applications, for storing progressive-scan content, MPEG-1 is pretty much identically good, and players are much more widely distributed (it still costs $2.50 to distribute a MPEG-2 decoder legally).
MPEG-1 playback will certainly will be around until the end of time. MPEG-4 will probably be in the same position in a few years. I'd probably stick to ISMA complaint, using Simple Visual for video and AAC-LC mono or stereo for audio. Those files will play in QuickTime, RealOne (on windows), MPEG4IP, and other players today, and in the future. And MPEG4IP is even open source.
I've done a lot of Bink encoding (Bink is featured in one of the tutorials in my book).
Bink's adavantages are a really mature, widely ported (lots of consoles, Mac, Windows, Linux), fast, and highly-featured API. However, its internal codec isn't really THAT great in terms of compression efficiency (bang for the bit). It's fine for CD-ROM games where having a rich, fast, stable video playback system is more important than keeping space to a minimum. But I certainly wouldn't use it for content to be downloaded from the web.
He should have used PNG, which is lossless RGB 24-bit.
Of course, encoding in JPEG @ 100% quality would have been effectively as good. There really isn't any excuse for having still image artifacts on top of the source artifacts in this kind of article.
Back in the chipwich era...
//c with an amber monitor (which my parents got since I was supposed to be using it for school work, not games). Had a floppy with some games on it (amazing how much fun could go in 140K!). Went through a period where I was playing Joust and Mario Brothers (no Super back then - OLD school platformer). Joust I had pretty well mastered, but I never got the hang of the extra-icy late levels in Mario Bros. (around level 18?). Very frustrating - had to play quite a while to get to that point, and then I'd lose all my lives in only a couple of minutes. I'd get frustrated, play Joust for a while, then go back to M.B.. Rinse, repeat, for hours, late at night.
I had a Apple
So, whenever playing a lot of a game, I'd always start imagining playing it in my head while bored. I'd have the physics down cold, so it actually could be decent practice.
Then, one day on the bus home from school, the two games merged in my mind. I was playing Mario Brothers levels, but with the Joust bird. But the physics just didn't match up, so I was constantly getting killed. And I couldn't stop thinking about it - all day, stuck in my head, the darn bird trying to bang the darn coins, and getting darned killed. All in glowing amber phosphers. Kind of like being in hell for the really, really lame.
Stopped playing both games at that point, and haven't booted up either in, what, 16 years?
So, basically a series-long "Yesterday's Enterprise?"
Or, perhaps, they're busily erasing the history of the past shows, to leave everything wide open? I've felt that the temporal cold war was in some way a result of the events in the 24th century and later.
So, I guess the question is what's happening in Enterprise the cause or the result of the past series. If it's the result, who the heck knows what's going to going to happen, and we can stop worrying about continuity.
Maybe I just don't get out enough to know what people are listening too these days. I had thought all three of those were pretty mainstream, cover-of-Rolling-Stone groups.
Is everyone in college off listening to Goa or something now?
Yep. One great thing about digital distribution is that an album can sell 500 copies and still be profitable! In the long term, this will be good for niches,
For another example, check out CustomFlix, who does on-demand DVD replication and distibution. I've made a tidy bundle selling the DVD-R supplement for my book through them. It hasn't sold anywhere near the 500 copy minimum that a mass-market duplication would have required, but I started netting a profit from them after selling the first SIX copies.
http://www.customflix.com
So, I've got 14K songs in my MP3-based iTunes library, replicating my CD collection. I'm a mini-van driving work-from-home 32 year old married father of two who wasn't cool in the day, and certainly am not cool now. I don't have a tattoo, ride a Vespa, or compile kernels from source.
But, to my surprise, I'm revealed to be an elitist indy music prig! To test the new service, I sorted my iTunes library by play count so I could compare the quality of my current rips with the new service's previews.
And darn it off if pretty much all the music I listen to isn't from one of the majors! In order of "play count":
Sleater-Kinney - nope
The Hives - yep (one album)
The White Stripes - nope
Husker Du - nope
Man or Astroman - no
Len - just three tracks
Rancid - no
Veruca Salt - yep (two albums)
The Clash - pretty much everything
Riverdales - no
Screeching Weasel - no
Beastie Boys - no (an on a major, I thought)
Cast of Buffy the Vampire Slayer - nope!
Fugazi - no
Renegade Soundwave - no
Gang of Four - no
Luscious Jackson - yep
Pop Will Eat Itself - just one album
Tricky - lots of stuff
X - just one album
So, of my top twenty-by-listening bands (which is surprisingly different from what i would have guessed they would be), only 4 have a substantial body of their work available among the 200,000 tracks available through iTunes.
Blockbuster+DVD works well because you can get 18 GB in 20 minutes. It's a high latency connection, but tons of bandwidth!
While broadband connections and modern codecs are getting good enough to do a reasonable quality standard definition broadcast, we're not too far from blue-laser HD DVD.
Well, to a certain degree.
However, the number of bits for "good enough" audio is a LOT lower than for video. For any reasonable extrapolation of high-end stereo system, (like a full home 7.1 system that can usefully deal with dynamic ranges and frequcies well beyond what speakers can do), we can already get to the "so good golden ears can't hear the difference" with a typical broadband connection. For people with typical systems, even 48 Kbps with some modern codecs like HE AAC has been shown to be as good as CD.
Now, for video, things get trickier. I doubt 8 Mbps will be enough for percepturally lossless 1920x1080 in a real-time stream, even with WMV9 or AVC.
And in fact the QuickTime AAC encoder got a LOT better in QuickTime 6.1. It added a broader range of sample rates that could be used at a given data rates, and both faster (but lower quality) and slower (but higher quality) modes.
A study like this is comparing codecs, not bitstreams. While a good MP3 encoder will certainly beat a poor AAC encoder, the more powerful tools availble in the AAC bitstream will enable it to provide better compression efficiency with encoders of equal quality.
Ah! Just the trick.
.m4a, or the library, or what?
Next question - where does the metadata live one they're ripped? In the
In my workflow, I want to keep a big bunch of high data rate files on the home server (about 140 GB of 320 Kbps MP3 files), and then recompress to more portable formats to carry around on the PowerBook or whatever. This used to work fine. I'd use the Import feature of iTunes, and would convert from the 320 Kbps master file to ~150 Kbps VBR MP3 files for the road. While the lower data rates wouldn't work on my home Paradigm speakers, they were fine for listening to on airplanes.
However, this doesn't seem to work in iTunes 4. I see the Import option, but all the MP3 files in my current library are grayed out. Is this operator error, or does this not work anymore? If not, what is the Import function for?
Obviously I'd like to switch to 128 Kbps AAC-LC for my mobile music. But heck, I'd live with being able to make my old MP3 files!
-Ben
While I doubt iLife makes a lot of money, the professional apps like Final Cut Pro absolutely do. $1000 per boxed copy - with a higher gross margin than most of Apple's computers I'd warrant.
Although with Apple's pricing strategy for Shake ($5K for Mac w/ free network rendering, $10K for Linux, plus ~3K per rendering node), they're definitely trying to sell more Xserves there right now.
I had a T68i via ATT for a week, and returned it for a Motorola T720. the T68i was a great phone in pretty much every way except that its signal strength simply wasn't adequate in lots of places I need to be (Portland airport, San Jose). The T720 works great for me in places where the T68i would cut out constantly. I do miss the Bluetooth, though.
I looked at the new Nokias, but I just couldn't hack the weird number pad. Yeesh.
CELP? No, that's a different codec.
CELP might be largely bypassed in favor of AMR - Adaptive Multi Rate, which offers real-time scalability for speech. Good stuff.
It likely would. Although I haven't seen much in the way of RTSP/multicast OGG content out there. I'm not sure if there is any technical reason for this. Experts seem to think that OGG isn't very well designed as a RTSP format, but I don't understand those particular issues in enough depth to state what the problem there is.
A few reasosn not to use MP3:
No multicasting
No native RTSP support
No good loss recovery mechanism
It's amazing that MP3 works as well as it does for pseudo-streaming, but a true streaming format it ain't. Personally, I'd like to see them adopt a MPEG-4 AAC-LC stream, which QuickTime, RealOne for Windows, and other ISMA MPEG-4 compliant players could tune into. Better quality at lower data rates than MP3.
Pretty soon we'll have AAC High Effeciency, which can do ~FM quality at 32 Kbps for 44.1 stereo. Astonishingly better than other propritary codecs in head-to-head at these low bitrates.
Er?
At Quality=X, WMV9=Z, MEPG-2=2Z
Thus, 100% larger.
I'm actually in the process of designing a digital cinema system along these lines, so here's a few comments
Everyone is focusing on the codec being the quality limitation, but that's not true. In fact, the projector is the biggest deal. There are plenty of modern codecs that can give you visually lossless quality if you throw enough bits at them. The issue with codecs is getting compression efficiency up so that transmission and storage is cheaper, and keeping decode complexity down so you don't need to have expensive hardware in the projector. The WM9 system is pretty much a high end (but not the highest end) Dell workstation, strapped to a cart with XLR audio out, a control pad, and a big data projector on the top. All off the shelf parts, which makes implementation cheap, and upgrading the computer very cheap. But those are nice things to have, but not strictly required for digital projection.
But we could do the same thing with MPEG-4, or other formats. WM9 has a more mature DRM solution and some other advantages, but it is absolutely possible to use another format.
The big limit is in having a projector that is bright enough to fill the room, with a dark black, and high resolution. Moore's law gives us improvements in compression faster than we get improvements in projection, so the big photon cannon will be the true limit on quality for a while.
Your description is more like an API. The conformance point of MPEG-4 is a bitstream. Any encoder that makes a legal bitstream is legal, and any player than can take any legal bitstream is legal. Within that, players and encoders are limited by Profiles and Levels, which define the tools, and the maximum paramters of the tools, that can be used.
WMV9 is derived form a draft standard of MPEG-4, but isn't MPEG-4 in any meaningful (=interoperable) way. It isn't "just" a propritary codec, but is a fabulously good one.
ALL delivery video codecs use I frames plus P frames. Some even use B-frames. But that's like saying XML is just a derivative of ASCII.
Actually, WMV9 isn't a MPEG-4 codec. Earlier versions were based on draft MPEG-4 standards, but they forked quite a while ago.
Also, the difference is a lot bigger than 30%. It's more like 100% more for MPEG-2, with the gap increasing as data rates get lower.