Well, I guess that NYT (and many others) allowing Google News to login and index their content means that they like them doing that for getting traffic. For whatever reason, NYT wants you to register and they have a right to as well as they have copyright, allowing Google to put in the snippet, but not the whole article without their consent.
And that is the reason for an index, to find the original.
It is good to see they are working this out together, though, without NYT going to court as the first step. This is a far better way than the popular shoot-first-ask-questions-later attitide most media companies have...
Why download it...
on
Random Humor
·
· Score: 5, Funny
I think our two operations were in two very seperate market segments. We had 36 automated streams (using Dalet), programmed by some very talented music directors using Power Gold. This service was not aimed at the purists who shop at Championship Vinyl!
People who just want a consistant sound that doesn't fade into the noise of a crowded office every when played at low volume outnumber purists probably 1000 to 1. More consistant levels equals higher Time Spent Listening. (TSL)
If they want homogenized pre-processed shit, they can download (or merely enable) a plugin for it
Ecxuse me, but I think that I can do a lot better job of processing with a specialy designed, expensive, hardware unit I spent days tweaking than some WMP plugin!
I used to work at a classical music FM station. For the old management, purists, only the mildest gain riding and emergeny limiting was allowed and we played only the more "advanced" works, with a library of 10.000 tracks and carefull anouncements of all the details of the work and performers. This is how classical music needed to be presented, classical listeners were purists and oh so different from the masses. When new management came in, we cut the library to 750 tracks, only the most famous and easy to listen to pieces, and crancked up our Inovonics to rival CHR stations in loudness so we could be listened to in cars, office, and while vacuuming the house. Our TSL doubled and our share trippled in 6 months.
This is the age old purists vs. the masses debate. At the end of the day, the masses eat at McDonalds. If that's what they want, that's what I will provide. I have bills to pay.
Don't know about the state of streaming right now. I guess TCP vs. UDP was much more of an issue over slow connections. With HTTP over broadband, you simply buffer 10 seconds or so on the server side, which you pump out to the client at maximum speed. This way when a client has to ask a packet again, you have 10 seconds to do it. When you send the client that packet, you speed up the stream for a little bit after that to make up for lost time. Hmmm, HTTP/TCP may actualy be _better_ than UDP when the connection is many times faster than the audio bitrate!
Don't get me started on DRM! I remember the discussions with the Music Licensing Mob here in the UK. This was when Napster was still very much active and people were swapping perfectly ripped songs at 192Kbit left, right and center. But these idiots had a problem with us streaming highly processed audio at 64K with cross overs between songs and DJs talking over intros. "People might record it digitaly, so in perfect quality, and share it." And so they wanted highly unreasonable amounts of money. Sigh.
Of course, these days original masters are limited so aggressively that FM can't do any more damage to it.
FM engineers are now complaining about the fact that mastering engineers have discovered multiband limiting and the high end of CDs are very loud. This causes a problem with the emphasis on FM. If the CD's spectrum is flat, and you have to do emphasis, you end up not making the high end louder, but the lower end softer to keep within your allowed modulation! Of course this is not a problem with digital streaming.
I guess? I don't know enough about "the masses" to say. It seems like everyone can play mp3s these days
While true, MP3 is not a streaming format in itself. I do not know the situation now, but in 2000, MP3 streaming was pure HTTP over TCP, while for streaming you need a UDP protocol so the OS's TCP/IP stack doesn't hold things up waiting for a packet. RTSP, et al, use UDP and do their own resend requests and putting the packets back in the right order. That way if one fails to show on time, you can just skip it without breaking the whole stream. That and the fact we needed an embedable player, so it could have a nice branded Flash "skin" and ads...
Truth is, I've long since given up on understanding "the masses" vs. music
Music aside, this is more a technical issue. These are simple consumers, and you need something that works without them having to put in any effort. In that respect, Windows Media is a good choice. Nothing to install for the user and reasonable quality.
Plain old AGC + multiband compression + limiter that has made FM radio sound the way it does for the last 30 years or so. You have to realise, when I say "good", it is relative to other streaming services, NOT a CD or even broadcast radio.
Of course, at 64K, it will never sound as good as uncompressed, but that never stopped AM radio from becoming big. It's all about content. (OK, AM radio had no competition)
This was late 2000 and MP3 and Ogg were, and still are, useless for reaching the masses, for that you need Windows Media. Even Real was a mistake I think, not everybody had it installed and everyone seemed to connect at 64K anyway, making their bandwidth negotiation a moot point. Although when we tried to make Windows Media work, in true MS style it was full of bugs, Real was a much more mature product. And MS sacrificed artifacts for high end. While this makes it sound better to many people at first, after a while the extra artifacts are much more tiring. Real decided to minimize artifacts by simply not wasting any bytes on the top end.
I used to be CTO of a now defunct online radio service in the UK (puremix.com), way back in 2000/2001 and worked at various radio stations before that.
We streamed 64K Real Audio and it sounded great. The secret to making it sound good is audio proecessing, just like an analog radio station does. I am not advocating New York style maximizing of loudness at all cost, but any signal needs some work.
That work is missing on most not only amateur, but also professional streams or it is done by very bad software solutions. Online music services are often created by people who love and know their music and are geeks. Few of them are actualy audio wizards. (Even at radio stations, engineers are often under valued because the "creative" people don't understand what's involved) The result is that even peak signals are below maximum modulation and missing (multiband) compression and limiting makes sure there is no consintancy in quality and loudness between songs, which brings out encoding articfacts much more. And that is a real shame.
Even if you stick to ANSI SQL (which version?) most databases have too many differences and you can greatly optimize things and make life lot easier for yourself by using the tools available in the database. And then I haven't even touched stored procs yet!
Middleware suck. The whole pg_connect(), syb_connect() in PHP is silly, though. The least you can do is create your own getConnection() and such. ADODB is OK for useability, but it is painfully slow to include on each page.
The only thing I care about is if I am getting what I pay for, which I seem to get. I have the standard service here in the UK (512/256, although actual line speed is 10% higher) and I always get full up/download speeds. Yes I could buy a 1Mbit cable service (if there was cable in my area) and that would be faster, but that is comparing two completely different things, a 1Mb vs. a 512K service, the delivery method is a moot point.
Just checked a fair few DVD. In most normal scenes, it seem between 3-4, sometimes 5. In complex scenes it does go up to 5-7. Still average seems to be quite low. This is more than enough. Hardware MPEG decoders in a proper DVD player are much better than software decoders and judging bitrate/artifacts on a PC is not a very good idea, they will lose most of the time and 8Mbit seems so much better, even though you wouldn't seem the difference on a normal TV.
One major problem for the US is (where most readers seem to come from and this site is biased towards) that your NTSC system has 100 lines less than PAL and us Europeans (and Aussies) "Paying for Added Luxury" have a much better deal.
IP broadcasting can be done, especialy when IPv6 arrives but even then the carriers will want more money than broadcasters have. On the internet, both sending and recieving party pay.
Peercast will die the fate of popular P2P networks. Hailed as "the more users the better it works", the reality is: "too many users and it dies becasue non of these users have enough bandwidth to be hub". I have seen 10Mbit connections die because a PC was a Kazaa host. Kazaa saw it had a lot of bandwidth and made it a master. Ofcourse IT infrastructure shut this PC down.
Either that or the horizon will be too close and many stations are out of range, just like cool FM station in the other town.
Their servers would die because of the overload as these indie bands can't pay for servers and bandwidth themselves and there is still no viable business model for online radio, wether it is indie or Britney...
That's only the ActiveTcl Windows binaries, the source is a bit easier to get to, just like any other scripting language. But I do agree with you and wish someone would make an easier available windows binary. I still use 8.2 on windows for that reason!
...With no components to re-use when you want to upgrade in a few years time. I guess that feature is worth the extra expense. I'll save my money for a Shuttle mini PC...
The reason these people found MP3 better than the original can be for various reasons. First I doubt that the MP3 and the CD were played through the same D/A converter. Different converters sound different.
In the different converters scenario, sound levels might have been different and the MP3 might have been a fraction louder, which people like.
But even if they were using the same converters, MP3 introduces distortion. Distortion sounds louder to the ear than the original. And again, to most people that little extra loudness sounds better.
And no, artists don't put their stuff through encoding cycles. There are other ways to make a 'finished' recording sounds better or at least more appealing in the short term. That is why cream of the crop mastering engineers like Bob Ludwig are celebrities in the industry.
You don't need full ACID to be reliable, Just look at Sybase or MS SQL server, they both aren't truly due to the lack of multi-versioning that Oracle and Postgress have. That doesn't mean they are not reliable!
That said: I wouldn't trust MySql for anything. What makes it's acceptance in the open source community even harder to understand for me is the lack of a true GPL. Postgress is very reliable and is GPL. I also wouldn't know how to survive without referential integrity constraints, outer joins, subselects and nested queries!
But for a project like this, I would certainly go for Sybase or Oracle, I love free/OS software, which is often better than commercial products, but when it comes to database, I am not convinced yet!
Beter close the curtains next time I take a whiz or get some Haagen Dasz from the fridge during the ad breaks, for all I know a neighbour might be watching and snitch on me!
"Let's create a website on which people can create their homepages and when people come to visit those homepages, we put OUR ads on them and make money that way"
"The web site you are trying to access has exceeded its allocated data transfer."
Anyone else missing something here? Or is it just me?
"5 Hours of CD quality music", on a 500Mb disk? 500Mb will only hold 50 minutes of CD quality audio. I don't mind compression, I love my MD player, but cramming 5 hours in 500 megs is just ~225Kbit. Good enough to be used as a portable device, but buying pre-recorded ones? I'll just stick to the CD to enjoy at home and record that onto a compressed format for use on the tube.
Slashdot Mirror
Well, I guess that NYT (and many others) allowing Google News to login and index their content means that they like them doing that for getting traffic. For whatever reason, NYT wants you to register and they have a right to as well as they have copyright, allowing Google to put in the snippet, but not the whole article without their consent.
And that is the reason for an index, to find the original.
It is good to see they are working this out together, though, without NYT going to court as the first step. This is a far better way than the popular shoot-first-ask-questions-later attitide most media companies have...
Don't forget to also get the 1997 video "It Could Have Been So Easy"!
I think our two operations were in two very seperate market segments. We had 36 automated streams (using Dalet), programmed by some very talented music directors using Power Gold. This service was not aimed at the purists who shop at Championship Vinyl!
People who just want a consistant sound that doesn't fade into the noise of a crowded office every when played at low volume outnumber purists probably 1000 to 1. More consistant levels equals higher Time Spent Listening. (TSL)
If they want homogenized pre-processed shit, they can download (or merely enable) a plugin for it
Ecxuse me, but I think that I can do a lot better job of processing with a specialy designed, expensive, hardware unit I spent days tweaking than some WMP plugin!
I used to work at a classical music FM station. For the old management, purists, only the mildest gain riding and emergeny limiting was allowed and we played only the more "advanced" works, with a library of 10.000 tracks and carefull anouncements of all the details of the work and performers. This is how classical music needed to be presented, classical listeners were purists and oh so different from the masses. When new management came in, we cut the library to 750 tracks, only the most famous and easy to listen to pieces, and crancked up our Inovonics to rival CHR stations in loudness so we could be listened to in cars, office, and while vacuuming the house. Our TSL doubled and our share trippled in 6 months.
This is the age old purists vs. the masses debate. At the end of the day, the masses eat at McDonalds. If that's what they want, that's what I will provide. I have bills to pay.
Don't know about the state of streaming right now. I guess TCP vs. UDP was much more of an issue over slow connections. With HTTP over broadband, you simply buffer 10 seconds or so on the server side, which you pump out to the client at maximum speed. This way when a client has to ask a packet again, you have 10 seconds to do it. When you send the client that packet, you speed up the stream for a little bit after that to make up for lost time. Hmmm, HTTP/TCP may actualy be _better_ than UDP when the connection is many times faster than the audio bitrate!
Don't get me started on DRM! I remember the discussions with the Music Licensing Mob here in the UK. This was when Napster was still very much active and people were swapping perfectly ripped songs at 192Kbit left, right and center. But these idiots had a problem with us streaming highly processed audio at 64K with cross overs between songs and DJs talking over intros. "People might record it digitaly, so in perfect quality, and share it." And so they wanted highly unreasonable amounts of money. Sigh.
FM engineers are now complaining about the fact that mastering engineers have discovered multiband limiting and the high end of CDs are very loud. This causes a problem with the emphasis on FM. If the CD's spectrum is flat, and you have to do emphasis, you end up not making the high end louder, but the lower end softer to keep within your allowed modulation! Of course this is not a problem with digital streaming.
I guess? I don't know enough about "the masses" to say. It seems like everyone can play mp3s these days
While true, MP3 is not a streaming format in itself. I do not know the situation now, but in 2000, MP3 streaming was pure HTTP over TCP, while for streaming you need a UDP protocol so the OS's TCP/IP stack doesn't hold things up waiting for a packet. RTSP, et al, use UDP and do their own resend requests and putting the packets back in the right order. That way if one fails to show on time, you can just skip it without breaking the whole stream. That and the fact we needed an embedable player, so it could have a nice branded Flash "skin" and ads...
Truth is, I've long since given up on understanding "the masses" vs. music
Music aside, this is more a technical issue. These are simple consumers, and you need something that works without them having to put in any effort. In that respect, Windows Media is a good choice. Nothing to install for the user and reasonable quality.
Plain old AGC + multiband compression + limiter that has made FM radio sound the way it does for the last 30 years or so. You have to realise, when I say "good", it is relative to other streaming services, NOT a CD or even broadcast radio.
Of course, at 64K, it will never sound as good as uncompressed, but that never stopped AM radio from becoming big. It's all about content. (OK, AM radio had no competition)
This was late 2000 and MP3 and Ogg were, and still are, useless for reaching the masses, for that you need Windows Media. Even Real was a mistake I think, not everybody had it installed and everyone seemed to connect at 64K anyway, making their bandwidth negotiation a moot point. Although when we tried to make Windows Media work, in true MS style it was full of bugs, Real was a much more mature product. And MS sacrificed artifacts for high end. While this makes it sound better to many people at first, after a while the extra artifacts are much more tiring. Real decided to minimize artifacts by simply not wasting any bytes on the top end.
I used to be CTO of a now defunct online radio service in the UK (puremix.com), way back in 2000/2001 and worked at various radio stations before that.
We streamed 64K Real Audio and it sounded great. The secret to making it sound good is audio proecessing, just like an analog radio station does. I am not advocating New York style maximizing of loudness at all cost, but any signal needs some work.
That work is missing on most not only amateur, but also professional streams or it is done by very bad software solutions. Online music services are often created by people who love and know their music and are geeks. Few of them are actualy audio wizards. (Even at radio stations, engineers are often under valued because the "creative" people don't understand what's involved) The result is that even peak signals are below maximum modulation and missing (multiband) compression and limiting makes sure there is no consintancy in quality and loudness between songs, which brings out encoding articfacts much more. And that is a real shame.
Use mod_rewrite to have certain URL patterns handled by a PHP page. Works wonders!
Even if you stick to ANSI SQL (which version?) most databases have too many differences and you can greatly optimize things and make life lot easier for yourself by using the tools available in the database. And then I haven't even touched stored procs yet!
Middleware suck. The whole pg_connect(), syb_connect() in PHP is silly, though. The least you can do is create your own getConnection() and such. ADODB is OK for useability, but it is painfully slow to include on each page.
The best way is to set up your own company, being employed by a foreign company is a PITA, both for you as well as their HR department.
The only thing I care about is if I am getting what I pay for, which I seem to get. I have the standard service here in the UK (512/256, although actual line speed is 10% higher) and I always get full up/download speeds. Yes I could buy a 1Mbit cable service (if there was cable in my area) and that would be faster, but that is comparing two completely different things, a 1Mb vs. a 512K service, the delivery method is a moot point.
Just checked a fair few DVD. In most normal scenes, it seem between 3-4, sometimes 5. In complex scenes it does go up to 5-7. Still average seems to be quite low. This is more than enough. Hardware MPEG decoders in a proper DVD player are much better than software decoders and judging bitrate/artifacts on a PC is not a very good idea, they will lose most of the time and 8Mbit seems so much better, even though you wouldn't seem the difference on a normal TV.
One major problem for the US is (where most readers seem to come from and this site is biased towards) that your NTSC system has 100 lines less than PAL and us Europeans (and Aussies) "Paying for Added Luxury" have a much better deal.
My DVD player can show the current bitrate and 3-4 seems more like it. No wonder this miracle compression algorithm works miracles at 5-8!
IP broadcasting can be done, especialy when IPv6 arrives but even then the carriers will want more money than broadcasters have. On the internet, both sending and recieving party pay.
Peercast will die the fate of popular P2P networks. Hailed as "the more users the better it works", the reality is: "too many users and it dies becasue non of these users have enough bandwidth to be hub". I have seen 10Mbit connections die because a PC was a Kazaa host. Kazaa saw it had a lot of bandwidth and made it a master. Ofcourse IT infrastructure shut this PC down.
Either that or the horizon will be too close and many stations are out of range, just like cool FM station in the other town.
Their servers would die because of the overload as these indie bands can't pay for servers and bandwidth themselves and there is still no viable business model for online radio, wether it is indie or Britney...
WiFi or not, I'd rather still have a Turtle Beach Audiotron connected to my home stereo!
That's only the ActiveTcl Windows binaries, the source is a bit easier to get to, just like any other scripting language. But I do agree with you and wish someone would make an easier available windows binary. I still use 8.2 on windows for that reason!
...With no components to re-use when you want to upgrade in a few years time. I guess that feature is worth the extra expense. I'll save my money for a Shuttle mini PC...
The reason these people found MP3 better than the original can be for various reasons. First I doubt that the MP3 and the CD were played through the same D/A converter. Different converters sound different.
In the different converters scenario, sound levels might have been different and the MP3 might have been a fraction louder, which people like.
But even if they were using the same converters, MP3 introduces distortion. Distortion sounds louder to the ear than the original. And again, to most people that little extra loudness sounds better.
And no, artists don't put their stuff through encoding cycles. There are other ways to make a 'finished' recording sounds better or at least more appealing in the short term. That is why cream of the crop mastering engineers like Bob Ludwig are celebrities in the industry.
You don't need full ACID to be reliable, Just look at Sybase or MS SQL server, they both aren't truly due to the lack of multi-versioning that Oracle and Postgress have. That doesn't mean they are not reliable!
That said: I wouldn't trust MySql for anything. What makes it's acceptance in the open source community even harder to understand for me is the lack of a true GPL. Postgress is very reliable and is GPL. I also wouldn't know how to survive without referential integrity constraints, outer joins, subselects and nested queries!
But for a project like this, I would certainly go for Sybase or Oracle, I love free/OS software, which is often better than commercial products, but when it comes to database, I am not convinced yet!
Beter close the curtains next time I take a whiz or get some Haagen Dasz from the fridge during the ad breaks, for all I know a neighbour might be watching and snitch on me!
"Let's create a website on which people can create their homepages and when people come to visit those homepages, we put OUR ads on them and make money that way"
"The web site you are trying to access has exceeded its allocated data transfer."
Anyone else missing something here? Or is it just me?
"5 Hours of CD quality music", on a 500Mb disk? 500Mb will only hold 50 minutes of CD quality audio. I don't mind compression, I love my MD player, but cramming 5 hours in 500 megs is just ~225Kbit. Good enough to be used as a portable device, but buying pre-recorded ones? I'll just stick to the CD to enjoy at home and record that onto a compressed format for use on the tube.
Now where are those 24/96 audio DVDs!?