Summary: Audio content on the internet is in chaos. To reign in the chaos, and to capitalize on internet audio file assets, Google will launch an audio search engine or audio file search tool by 2006, but probably sooner.
Audio Technology on the Edge
Below I have listed some technologies (i.e., BitTorrent, Overnight Loading, Audio Blogs, Podcasting and Speech-to-Text) that allow me to speculate that Google will launch an audio search engine. Please read about these technologies and start thinking about how audio technology has evolved over the last couple of years. We are much, much farther along than most people realize.
I believe that Google is looking at this information and these tools and they are going to do something amazing with audio assets. But I am getting ahead of myself. Read about the technology and then I'll explain how Google's audio search engine will work.
BitTorrent: Audio files are big but bandwidth is limited. However, bandwidth continues to increase all the time. And, technology is slowly emerging to decrease the pain associated with dealing with large files. One way of dealing with massive files, like audio files, is BitTorrent. As I have explained before, BitTorrent is sort of like Napster. It connects your computer to other computers so you can share files. The more people that share or trade the same files, the faster they can be download by everyone. Just like Google PageRank, as popularity increases, the more relevant the content becomes. I know this might be confusing. So, if you want to learn more or if you need help, read my BitTorrent FAQ for Little Monkeys. It is a simple introduction for people who fear technology. (Don't worry, I fear technology sometimes too.)
Overnight Data Loading: Many people understand how RSS works, and subscribing to RSS feeds isn't too hard. There are many new web services available for viewing RSS feeds, and there are also plenty of feed readers. But, I'm less interested in feed services and feed tools. I'm much more interested in overnight loading (a.k.a., RSS payloads) as it applies to audio files and audio blogs. The idea is that you subscribe to audio blogs and preload them behind the scenes on your computer or audio playing device. Since the audio files are preloaded, you simply click and listen. No wait, no fuss. Instant audio gratification.
Audio Blogs: In addition to mainstream web sites, more and more people are generating audio content via blogs. In the past this information was too big to manage for most users due to bandwidth, but it is getting easier as I described above. Also, it is getting easier to create and post audio content. In effect, more audio content is being generated and therefore more value in audio is building up as the pool of content expands. It is getting difficult to track audio blog information for many reasons. Unfortunately, audio blogs aren't very accessible at this point. This technical void is exacerbated by the increasing gravity between all the audio objects being produced by audio bloggers. Remember, as the pool of audio content increases, the value (i.e., potential money to be extracted) of the audio pool directly increases. Keep in mind that this audio content is very speech driven versus music driven. (Bloggers talk, they don't sing.) Speech can be converted to text. It is easy to search through text, folks.
Podcasting: Somewhere I read that Podcasting is sort of like TiVo for your audio player. That seems to be a good but simple way to explain it. The idea is that your audio playing device is used for storing and playing audio files that you have subscribed to via RSS 2.0, with enclosures. In plain English, you subscribe to get audio files, they get loaded on your device, and you play them. No mess, no pain. This is important in that it is a relatively simple way for people to get audio files directly to audio playing devices. When technology is simple, people will use it. RSS is simple, by definition, and podcasting is similarly simple. I advise you to jump on this technology very soon. It is going to get hot. (Will someone please write up a clear "How to Podcast" article?)
Speech-to-Text Conversion: The technology to convert speech to text is already available. It isn't perfect, but it is getting very close. My tablet computer is capable of capturing my speech and converting it to text, although training was required. Imagine if audio blog feeds were converted to text files, and in turn, those text files were parsed and stored in a database. Any company, such as Google, willing to pour resources into speech-to-text conversion, and storage, will be able to dominate the audio space by using simple text-based search technology. Side issues such as storage aren't really issues. It is cheap and getting cheaper. With Google giving out gigabyte email accounts, you know they can handle billions of small, parsed text files.
Smash these technologies together and you should suddenly realize that the future of audio is here today. Napster was certainly the start, and it kicked off the MP3 revolution, but now we are seeing small pieces loosely joined. We're seeing the emergence of the audio internet. Google is certainly watching the audio internet and they will index it and they will make it possible for you to search for and through audio files.
The Four Ways that Google Audio Search Will Work
Assuming the technology described above works, and it does, then Google is in a position to control audio search on the internet. To succeed, they only need to get between the files and the people who want the files. Once they index audio files and once they parse audio files, the game is over. No other search engine will be able to compete with them. They have the money, technology, and (unbelievable) brainpower to stay ahead for a long time.
But what is "audio search" exactly? I've answered this question in four ways below. In effect, I map out four simple ways for how Google audio search would work. If you believe in the technology I talked about above, then you should believe that an audio search engine is entirely feasible. Here's the list; read on.
First, like current MP3 search engines, you will be able to find MP3's (and other audio files) based on file names. That is, if a song is called "John_Rhodes_Sings_About_Google.mp3" you will be able to find that song doing a search such as "Rhodes Google MP3" or "Rhodes MP3" in the search engine. Your keywords will help you track down audio files based on how the files are named. This is already being done and it is extremely simple, especially for a company like Google.
Second, the search engine will be able to hunt down semantic web information. In plain terms, the search functionality will allow you to search through the metatags (or metadata) of songs, such as artist, album, length, year, and more. Perhaps this will be based on ID3 tag search. Regardless, audio files are more than audio. There is simple text information available under the surface. Google will index this metadata and your keywords again will help you find audio files through simple text mapping.
Third, the search engine will allow you to find songs based on the words used in the song. It'll work like this. MP3's (audio files) will be converted to text and the text will be stored. Google already caches web pages, and they will cache both the audio file and the text conversion of the audio file. As part of this functionality, they might even offer transcription services whereby audio files are parsed and the text is made available, perhaps for a fee. Once again, using keywords you will be able to find audio files. However, the big difference is that your keyword text search will actually allow you to find audio files based on the words in the audio file. That's right, just like Amazon's "Search Inside the Book", you will be able to search through audio content using text. Amazing. Of course, audio blog searches will be first because they are easier to parse, whereas the music in songs will be problematic for some time. In this sense, all music is noise.
Finally, the audio search will allow users to find files based on associations between songs. Audio files aren't currently well indexed and related to each other by any central source. Unlike HTML files hyperlinked to other HTML files, audio files aren't well linked at all. However, Google will associate file names, metadata, and speech-to-text content, so that you can literally link from one song to another song. In simple terms, the text (or text extracts) associated with audio files are easy to link to if your database is big enough. Google can convert and store audio files to text, index that data, associate it, and make it available to you. Interestingly, Google will link audio files to audio files, using simple text. Searching by association will be natural to you. We all talk about songs in reference to other songs, and audio (e.g., conversations) to audio, based on many category types. We make associations all the time. Search functionality that allows users to find similar songs and audio files (e.g., artist, year, time, blog name, speaker, popularity, types of instruments, similar sounds, similar topics) is coming soon. It will be elegant but simple; the Google way. As Dave Winer would say, this is a mindbomb. Definitely.
Aside: Google Audio Search for Dollars
Making money from internet audio search is the topic of another article. However, the usual suspects for generating revenue are alive and well, such as advertising. With a little thinking, I'm sure you can list plenty of other ways for Google to make money by controlling audio search on the internet. It is an untapped goldmine. We're not talking about millions of dollars, we're probably talking about billions. (By the way, there are many ways to make money on the "plain old" internet.)
Summary Time
So that's the story. The audio web has come together and many people didn't even know it. On top of that, I predict that Google (or some competitor) will use its strength and intelligence to make the audio web available to everyone. Google Audio Search will become a reality because audio information is becoming a true internet asset, and when assets are unclassified there is a search void. Google will fill the audio search void, and make a lot of money. You heard it first right here, right now.
Audio Technology on the Edge
Below I have listed some technologies (i.e., BitTorrent, Overnight Loading, Audio Blogs, Podcasting and Speech-to-Text) that allow me to speculate that Google will launch an audio search engine. Please read about these technologies and start thinking about how audio technology has evolved over the last couple of years. We are much, much farther along than most people realize.
I believe that Google is looking at this information and these tools and they are going to do something amazing with audio assets. But I am getting ahead of myself. Read about the technology and then I'll explain how Google's audio search engine will work.
BitTorrent: Audio files are big but bandwidth is limited. However, bandwidth continues to increase all the time. And, technology is slowly emerging to decrease the pain associated with dealing with large files. One way of dealing with massive files, like audio files, is BitTorrent. As I have explained before, BitTorrent is sort of like Napster. It connects your computer to other computers so you can share files. The more people that share or trade the same files, the faster they can be download by everyone. Just like Google PageRank, as popularity increases, the more relevant the content becomes. I know this might be confusing. So, if you want to learn more or if you need help, read my BitTorrent FAQ for Little Monkeys. It is a simple introduction for people who fear technology. (Don't worry, I fear technology sometimes too.)
Overnight Data Loading: Many people understand how RSS works, and subscribing to RSS feeds isn't too hard. There are many new web services available for viewing RSS feeds, and there are also plenty of feed readers. But, I'm less interested in feed services and feed tools. I'm much more interested in overnight loading (a.k.a., RSS payloads) as it applies to audio files and audio blogs. The idea is that you subscribe to audio blogs and preload them behind the scenes on your computer or audio playing device. Since the audio files are preloaded, you simply click and listen. No wait, no fuss. Instant audio gratification.
Audio Blogs: In addition to mainstream web sites, more and more people are generating audio content via blogs. In the past this information was too big to manage for most users due to bandwidth, but it is getting easier as I described above. Also, it is getting easier to create and post audio content. In effect, more audio content is being generated and therefore more value in audio is building up as the pool of content expands. It is getting difficult to track audio blog information for many reasons. Unfortunately, audio blogs aren't very accessible at this point. This technical void is exacerbated by the increasing gravity between all the audio objects being produced by audio bloggers. Remember, as the pool of audio content increases, the value (i.e., potential money to be extracted) of the audio pool directly increases. Keep in mind that this audio content is very speech driven versus music driven. (Bloggers talk, they don't sing.) Speech can be converted to text. It is easy to search through text, folks.
Podcasting: Somewhere I read that Podcasting is sort of like TiVo for your audio player. That seems to be a good but simple way to explain it. The idea is that your audio playing device is used for storing and playing audio files that you have subscribed to via RSS 2.0, with enclosures. In plain English, you subscribe to get audio files, they get loaded on your device, and you play them. No mess, no pain. This is important in that it is a relatively simple way for people to get audio files directly to audio playing devices. When technology is simple, people will use it. RSS is simple, by definition, and podcasting is similarly simple. I advise you to jump on this technology very soon. It is going to get hot. (Will someone please write up a clear "How to Podcast" article?)
Speech-to-Text Conversion: The technology to convert speech to text is already available. It isn't perfect, but it is getting very close. My tablet computer is capable of capturing my speech and converting it to text, although training was required. Imagine if audio blog feeds were converted to text files, and in turn, those text files were parsed and stored in a database. Any company, such as Google, willing to pour resources into speech-to-text conversion, and storage, will be able to dominate the audio space by using simple text-based search technology. Side issues such as storage aren't really issues. It is cheap and getting cheaper. With Google giving out gigabyte email accounts, you know they can handle billions of small, parsed text files.
Smash these technologies together and you should suddenly realize that the future of audio is here today. Napster was certainly the start, and it kicked off the MP3 revolution, but now we are seeing small pieces loosely joined. We're seeing the emergence of the audio internet. Google is certainly watching the audio internet and they will index it and they will make it possible for you to search for and through audio files.
The Four Ways that Google Audio Search Will Work
Assuming the technology described above works, and it does, then Google is in a position to control audio search on the internet. To succeed, they only need to get between the files and the people who want the files. Once they index audio files and once they parse audio files, the game is over. No other search engine will be able to compete with them. They have the money, technology, and (unbelievable) brainpower to stay ahead for a long time.
But what is "audio search" exactly? I've answered this question in four ways below. In effect, I map out four simple ways for how Google audio search would work. If you believe in the technology I talked about above, then you should believe that an audio search engine is entirely feasible. Here's the list; read on.
First, like current MP3 search engines, you will be able to find MP3's (and other audio files) based on file names. That is, if a song is called "John_Rhodes_Sings_About_Google.mp3" you will be able to find that song doing a search such as "Rhodes Google MP3" or "Rhodes MP3" in the search engine. Your keywords will help you track down audio files based on how the files are named. This is already being done and it is extremely simple, especially for a company like Google.
Second, the search engine will be able to hunt down semantic web information. In plain terms, the search functionality will allow you to search through the metatags (or metadata) of songs, such as artist, album, length, year, and more. Perhaps this will be based on ID3 tag search. Regardless, audio files are more than audio. There is simple text information available under the surface. Google will index this metadata and your keywords again will help you find audio files through simple text mapping.
Third, the search engine will allow you to find songs based on the words used in the song. It'll work like this. MP3's (audio files) will be converted to text and the text will be stored. Google already caches web pages, and they will cache both the audio file and the text conversion of the audio file. As part of this functionality, they might even offer transcription services whereby audio files are parsed and the text is made available, perhaps for a fee. Once again, using keywords you will be able to find audio files. However, the big difference is that your keyword text search will actually allow you to find audio files based on the words in the audio file. That's right, just like Amazon's "Search Inside the Book", you will be able to search through audio content using text. Amazing. Of course, audio blog searches will be first because they are easier to parse, whereas the music in songs will be problematic for some time. In this sense, all music is noise.
Finally, the audio search will allow users to find files based on associations between songs. Audio files aren't currently well indexed and related to each other by any central source. Unlike HTML files hyperlinked to other HTML files, audio files aren't well linked at all. However, Google will associate file names, metadata, and speech-to-text content, so that you can literally link from one song to another song. In simple terms, the text (or text extracts) associated with audio files are easy to link to if your database is big enough. Google can convert and store audio files to text, index that data, associate it, and make it available to you. Interestingly, Google will link audio files to audio files, using simple text. Searching by association will be natural to you. We all talk about songs in reference to other songs, and audio (e.g., conversations) to audio, based on many category types. We make associations all the time. Search functionality that allows users to find similar songs and audio files (e.g., artist, year, time, blog name, speaker, popularity, types of instruments, similar sounds, similar topics) is coming soon. It will be elegant but simple; the Google way. As Dave Winer would say, this is a mindbomb. Definitely.
Aside: Google Audio Search for Dollars
Making money from internet audio search is the topic of another article. However, the usual suspects for generating revenue are alive and well, such as advertising. With a little thinking, I'm sure you can list plenty of other ways for Google to make money by controlling audio search on the internet. It is an untapped goldmine. We're not talking about millions of dollars, we're probably talking about billions. (By the way, there are many ways to make money on the "plain old" internet.)
Summary Time
So that's the story. The audio web has come together and many people didn't even know it. On top of that, I predict that Google (or some competitor) will use its strength and intelligence to make the audio web available to everyone. Google Audio Search will become a reality because audio information is becoming a true internet asset, and when assets are unclassified there is a search void. Google will fill the audio search void, and make a lot of money. You heard it first right here, right now.
No comments:
Post a Comment