|
Streaming Media Report Original URL: http://www.nettime.org/nettime.w3archive/199911/msg00138.html This is a report written for the Xchange (http://www.xchange.re-lab.net) community about a commercial streaming media conference I attended in October of this year. For those unfamiliar with Xchange, its a network of practioners experimenting with creative applications of audio and video streaming. The first section of this report is a pre-amble specifically for Xchange. The second section is about the the conference I attended, my perception of streaming media on the net at this time, and some suggestions for streaming media artists and the Xchange community.
SECTION ONE Hi All, On October 19-20 this year I attended Streaming Media Europe a conference about the commercial streaming media industry. Approximately 500 executives, developers, and speakers came together to exchange ideas and discuss the direction of the industry in Europe. It was a very commercially orientated conference and I was representing the Amsterdam ISP XS4ALL, however I thought it may also be of interest to Xchange. So, I have also written a report for you all and included some of my own thoughts about issues that the conference raised for Xchange. Before I start, many thanks to Honor Harger for her thorough feedback about the report and her generous contribution of her own ideas. Many thanks also to Geert Lovink, David Garcia, and Josephine Bosma, for their feedback. During the conference it quickly became apparent to me that the Xchange community is very advanced in its use and thinking about streaming media. In fact many of the ideas I heard coming from business people working with streaming media did not (generally) reflect that same quality of ideas and breadth of understanding about the medium that I know to exist in Xchange. I found this very surprising and exhilarating as I personally was beginning to think that we (Xchange) were being left behind by industry. There are however some very sophisticated commercial sites, I have outlined these throughout the report and commented on them and my perception on their relevance to us. As stated in an earlier email that Zina Kaye and I sent to Xchange, I do believe that Xchange has to act quickly as a community to crystallise and finalise some of the ideas we have had for a long time. Janis' email about the possibilities for updating the Xchange site was very enlightening and gave good reason to think that we can make our presence on the net more sophisticated and interesting. I think this is absolutely necessary because if we don't act together to unify our identity I believe we will all be lost as individual sites in a mire of horrible commercial content and 'content portals'. It is also very satisfying, and gives me a lot of faith in Xchange, that Rasa and Raitis have said that they are willing for Xchange members to be involved in the process of updating the Xchange site. They perceive Re-lab as being 'another board member' rather than a controlling voice. This is a great tribute to their vision and sense of community that they have helped develop since they instigated the mailing list in 1997. I am also aware that each idea we have takes time to implement, but I am not sure we have so much time left to establish ourselves within this rapidly expanding field. My suggestion would be that the responsibility for doing everything should not fall just on Janis' shoulders (for example) but we should consider some kind of co-ordinated effort to contribute to a further development of Xchange. Specifically I would recommend setting a fixed date, perhaps on a weekend, where we could meet online and assign some tasks to work on during that weekend. The more I think about it the more a co-ordinated open policy for the site development looks attractive....
SECTION TWO Keynote Speakers Archived streamed video keynote speeches available at http://www.streamingmedia.net/streamingmedia/europe/program.html James Bethell, CEO, Ministry of Sound (http://www.ministryofsound.com/) James Bethell was the token European Keynote speaker in this largely American dominated line-up. His speech was very entertaining and informative. Essentially James spoke about the successes and failures of the Ministry of Sounds experiments with streaming media. Ministry of Sound (MoS) started as a very exclusive dance club that progressed into marketing and selling a lifestyle, namely the well-known club scene in London. Ministry of Sound's first experiments with audio on the net came selling music of associated DJs (linked directly to CDNOW an online CD reseller similar to Amazon.com - http://www.cdnow.com). They also have DJ sessions available for listening to. They experimented with using Streaming Media to distribute audio (remixes and DJ tracks), and to stimulate traffic to the site around which they could sell advertising. They claim that some of their MP3 tracks were downloaded by so many users that they would have forced the track onto the charts if the charting system recognized this as a legitimate gauge of public 'sales'. One of their most successful uses of streaming media was to put a live camera in the event so those that could not get into the club could still participate by watching the live stream. MoS also tried to run a mini-ISP for which they attracted 15,000 users but the free-service providers brought that investment to a very definite close. They see the internet as a necessary 'frontier' that they must absolutely dominate with the MoS brand. The time period they give themselves for this is about 1 year, if they aren't the most favored 'Club-lifestyle'' site within this time-period they will consider that they have lost the game. They see online branding and direct ecommerce sales (of merchandise) as a way to export their brand from one centralized, locally-controlled point, hence reducing the overheads (and logistical problems) of setting up international marketing and distribution offices. MoS see the net as the only method available for spreading their brand globally. It was interesting to see MoS represented here as they didn't altogether fit the line-up of 'heavy-weight' commercial developers. From James' presentation however it was evident that we could begin thinking of Xchange as a unified umbrella or (in commercial terms) 'brand'. MoS have achieved a very high profile on the net through developing a very sophisticated site that enhances (or tries to!), at every opportunity, their reputation as a leading expert on mainstream dance culture. Before continuing, to avoid misunderstanding, I must emphasise that I am not suggesting Xchange or any of its members become commercial. I am suggesting that there are some interesting ideas that the commercial sector uses that we could use to increase our visibility on the net. The first idea is to consider Xchange as having an identity that we can promote through as many available channels as possible. We must develop the reputation of Xchange as the place on the net to go to find interesting creative content. This involves a more sophisticated development of the site as Janis and Rasa have discussed and a more active strategy for making people aware of us. Because we do not have the money to spend as individuals or as a group to help us with this I think we need a much more 'street level' strategy. My first suggestion is for all of us to take the responsibility to ensure that the name 'Xchange' is taken with us where ever we go. We are infact all autonomous agents of Xchange (which is an important aspect of the network), and we should each act to spread the word as far as possible in whatever means we have available to us. For example:
Nicholas Butterworth, President
& CEO of MTVi (mtv.com and sonicnet.com) FlashRadio is certainly an interesting model and I am sure that something like this could be developed within Xchange for Xchange. A FlashTuner working with embedded realaudio (see http://www.radioqualia.va.com.au for an example - this is an experiment only, it requires Internet Explorer with Flash and RealAudio installed) for tuning through Xchange channels could be very useful. This is an area that I believe we could all catch up. I remember when I first saw 'Interface's Pirate Player' (http://interface.pirate-radio.co.uk/ ) and thinking how fantastic it was, but the commercial industry is now starting to create customised players and I believe we could go far to start experimenting more with this medium... The FreeB92 Player is a good example 1. More Sonicnet.... The belief of SonicNet is that "Every user can have their own station". In this context Nicholas Butterworth is talking about users as consumers (i.e. listeners) of programs and not as producers. They very much believe that the future of streaming media is in providing interactive channels for users. Hence they have developed some interesting ways to be able to search and access content, many of which are ideas we have all had and discussed but we have rarely implemented. I believe many people within Xchange have ideas that can be implemented on this level that outstretch the possibilities of what the commercial industry is doing. We must ask ourselves why, for example, is content by individual members of Xchange not easily searchable and locatable through one gateway (i.e. Xchange?). A creation of a searchable database for content created by all Xchange members would be great. Radio Internationale Stadt (http://orang.orang.de) is of course an excellent model for this but a significant number of Xchange members do not use this facility. Some centralisation of content information would be a great asset to assist potential listeners/users in finding our content. SonicNet also strongly believe that the great untapped mass of net consumers is ready for "mass consumption of streaming audio" but streaming video is still on the horizon. In their terms, the streaming audio industry is now poised to provide "a soundtrack to the web" (sic). MTV.com is of course the online MTV brand. Soon MTV is launching two new products, "WebRiot" and "Rock 'n Roll Collector", both of which are television programs broadcast on MTV with interactive audience participation components on the web. Nicholas Butterworth offered
some interesting insights into how they visualize their online
branding and how they cross-fertilize television consumption
with online activities. There is at the moment, he believes,
a three tiered system of users/watchers which co-relates to three
separate strategies. The interesting aspect of this model is that it reverses the commonly held belief that high bandwidth necessitates high interactivity. High bandwidth is primarily the domain, as MTV sees it, of those that want good quality streaming and are not so concerned with the interactive aspects of the medium. For each of these tiers MTVi believes they have to develop a distinct channel for each brand. The revenue models for MTVi
revolve very much around creating an audience and then selling
advertising to "meet the needs of the audience". Nicholas
Butterworth does not see micropayments (e.g. Pay per play) as
being viable because the user expectations are raised when they
pay for content, and the delivery mechanism (streaming media
and bandwidth provision) is not yet to the standard that any
streaming content provider can guarantee perfect delivery. Nicholas
Butterworth introduced 2 ideas that were re-iterated many times
throughout the conference: Tightly targeted, user-specific, advertising is the most preferred model for generating revenue by these businesses. Nicholas Butterworth also believes that the idea of "audio logos" is coming of age. An Xchange jingle, sting or ID would be a great plan. Like the net.radio jingles that Convex Tv (http://www.art-bag.net/convextv/) did, for instance. We could turn the audio logo idea inside out and turn it into a creative networking strategy - tiny embedded midi file stings on Xchange network member sites, for instance....
Anthony Bay, General Manager,
Streaming Media Division, Microsoft Corp They also have a system for encoding that allows the encoder to determine the 'access privileges' of the content.. For example, you can specify whether you want the user to view the content directly, view the content after paying, or view the content after registering your name (thereby giving away their personal data for which you can do with as you like). MS also promise to integrate the player into the OS.
LarryLux, CEO, Pseudo Programs,
Inc. http://www.pseudo.com/)
FORUMS: Get Up and Streaming This discussion was a lightweight look at video production for the internet for live events and pre-recorded files. The discussion centered on some basics including the type of cameras recommended in commercial applications, lighting, and how to deliver clean audio. No really interesting tips were given apart from the basics: Audio: keep the audio and encoding apparatus power supply separate from the lighting power supply to reduce hum. Video : use keyframes often (keyframes are a 'snapshot' between motion frames that appears much clearer). Get your lighting right. Use the best camera available to you. I think we should put together a cgi script 'database' that Xchange members can add tips and scripts to for general use. 2 Most of the panel agreed that there was very little you could do to a video signal to improve it before the encoding began. The main reason given for this is the fact that each organization was using a variety of codecs and hence each platform (real/windows/mpeg/QuickTime) dealt with video differentially so it is difficult to optimise one source for 2-4 different formats simultaneously. The really high-technology commercial sites did actually do some signal processing but they would then do this for each specific codec. However it was stressed that this was marginally effective for any output other than high bandwidth production. For low or medium bandwidth production (200k or less) the advice was to use a flat signal and make sure that your lighting was right and you had the best camera you had available.
Encoding Your Streaming Content
This centered largely on a discussion
about which codec was best for streaming video. · Real has 70 million
players distributed. It was generally acknowledged by content producers that WindowsMedia is now becoming popular and is taking a large share of the market away from Real. 30 frames per second was generally the agreed target for full motion streaming. For low bandwidth encoding the advice was to stay with a crisp image rather than trying to capture movement. This evolved into a discussion about different strategies for video content production. This was a very interesting discussion and focused on the strategies needed to produce interesting content for low bandwidth delivery. The most insightful advice was to forget about the 'main event' if it featured motion (sports events, theatre etc) and focus on 'behind the scenes' activities which have less motion and can be portrayed more effectively over low bandwidth video streaming. 80k audio was generally accepted to be so close to CD audio that a large percentage of the public would not be able to discern the difference between 80k audio and CD. Nice quote: "A PC is your typical attention deficit disorder box" ...from one of the panel members.
Show Me the Money Phil Wilson, CMN.com MJ Rauch, The Auto Channel Marc Lewis, Web Marketing Ltd David Chester, HotelView Corp. "Ownership of customers
(Marketshare) is the ingredient for success" This session was surpassingly lightweight but some interesting facts were highlighted. Freereportex<http://www.wordtracker.com/freereportex.htm>
reported that the most commonly searched keyword on the net is
MP3, replacing 'sex' for the first time as the most frequently
searched word: Another model for ensuring that site traffic is increased is to form alliances with 'portal' sites. The phenomenal increase in streaming media content on the web has inevitably lead to some of the most successful recent commercial enterprises are now streaming media 'portals'. An interesting discussion topic
was the rise of video banner advertising on websites. One claim
by Marc Lewis is that PC consumption time is now surpassing television
consumption in the United States. Hence web advertising is becoming
vastly more sophisticated. Video banners are the latest development
in online advertising. The advantages of video banner advertising
over television advertising are:
Streaming Media and Portals
This panel focussed on the need to create streaming media portal business models. Nothing surprising here, the portal sites main source of income was through gathering click traffic and then displaying advertising in the form of banner ads. Most agreed that 'instream advertising' (playing advertisements before, during or after content similar to television advertising strategy) was intrusive and generally not tolerated by most users. However it was admitted that there is, in many cases, an unspoken 'user contract', where users do recognise that they are being delivered something for 'free' and hence may tolerate some advertising with its delivery. There was some contention about how far this 'user contract' extended but generally most of the panel were careful when considering advertising strategies that might be considered intrusive. It was apparent from this discussion however that the streaming media portals are on the rise as their is a huge amount of streaming content being created for the internet. One problem faced by portals is that none of the main 4 streaming protocols (real, windows, MPEG, QuickTime) have searchable indexes embedded in the file. Hence content must be manually catagorised and indexed. It was also interesting to note that Excite do not have a sophisticated streaming media search system and that there seemed to be no strategy for them to implement one. Dela Quist from Excite was loath to discuss the reasons why, and gave the appearance that Excite is reluctant to enter into this business. He did however mention briefly that Excite intend on developing a children's version of Excite and an SMS search system (searchable using a mobile phone). Portals for streaming media are very definitely on the rise.
Developing and Managing a Streaming
Network Many ISPs are now considering Streaming Media as a necessary part of their long-term business strategies. It is acknowledged that in many countries average available bandwidth is getting higher. Hence the demand for high bandwidth content is rising. At the moment streamed audio is considered ready for mass consumption, in fact it is already one of the most popular mediums on the internet. However, while streaming video is not yet widely consumed the panel for this session agreed that codec improvements and the growth of high bandwidth connections will mean streaming video will eventually follow the same path as streaming audio. Globix (http://www.globix.com) is an American ISP that is beginning to introduce streaming media services to its customers. They consider that, as demand for rich media increases, the ISPs that can deliver uninterrupted delivery of rich media content will have a major advantage over those limited to low bandwidth connectivity or hindered by net congestion. The first part of the strategy for Globix is to educate their users and their company about the possibilities of streaming media, therefore sowing the seeds for demand of these services. This stimulates the need for high bandwidth and encourages users to upgrade their connection. Another discussion topic centered on the need to develop sophisticated caching servers for the delivery of streaming media content locally.
Content Management Strategies
This session really focused on one topic : indexing and organising content. There is an entire business sector that earns income from categorising and indexing content for media developers. CNN, for example, require extremely fast encoding, cateegorising and indexing of their television news segments for placement on CNN.com. Hence they employ the services of an indexing business. These businesses have sophisticated voice and face recognition software for creating searchable databases of content information, they also manually classify content. It would be great to develop a cgi script for Xchange that could be used to create ram files but that contained other information which can be searched. 3 Interestingly CNN stated that they prefer to work in both Real and Windows but Windows was definitely starting to bite into Reals marketshare. Additionally CNN hits are 50% 80k streams and 50% 20k streams, indicating that average bandwidth availability is rising. One problem CNN also faces is the problems with buffering of streamed content and they recommended that real (etc) should consider the "caching of personalised add content to play during buffering".
Streaming copyright and the
Law This was an extensive discussion
on the need to develop coherent and appropriate copyright protection
legislation for the internet. The consensus was that there are
no appropriate models for copyright protection on the net at
present. The problems that copyright protection organisations
face include: The final advice on this issue was that because copyright legislation for the internet is a very vague issue it is important for commercial enterprises to get a good legal insurance policy incase they inadvertently become a precedent case in court.
SUMMARY As stated above I think Xchange is at a very important crossroads, we have the opportunity to affirm our place as one of *the* places to go for artistic and creative approaches to streaming culture. Generally I think that Xchange could develop more into a 'portal' for creative streaming media. I also believe that we need to develop more mechanisms for sharing expertise about streaming media, to develop our own skills and to help others. To assist these developments I would like to suggest the following additions to Xchange: 1) A tips and Scripts database/ezine,
this could possibly be in collaboration with Rachel Bakers' TMselector
<http://www.tmselector.net>, which is already the embryo
for this idea. Plus we need to promote more
through: In addition I believe that we all have to consider diversifying our practice more and more into streaming video, or invite people to join Xchange that are already doing this.
I hope that you find this report interesting. I would love it if this could provide a basis for discussion either on the Xchange list or personally with me. Please feel free to critique and comment! Adam Hyde adam@va.com.au, adam@xs4all.nl
FOOTNOTES 1. The FreeB92 Player <http://www.freeb92.net/freeb92player/freeb92player.zip> is a piece of streaming media software, developed as a response to the events in Yugoslavia which have made it difficult for B92 to operate in a normal way. It is symbolic as much as it is functional. Symbolic because it represents a manifestation of B92 when the conflict in the region has ceased B92's continued operation. It is also another example of B92's innovative use of media as a response to a repressive circumstance Functionally it does the following:
All these presets can be updated.
There is also a SMIL FreeB92 Player Solution for the Macintosh (SMIL = Synchcronised Multimedia Integrated Language, a tagging language similar to HTML. It is like a media rich version of HTML in that it deals efficiently with Shockwave, Flash, RealAudio/video etc. It is being used by many companies including RealNetworks. In the context of RealNetworks it is being used to make the G2 player a "Rich Media Browser". This gives the G2 player an amazing amount of versatility in integrating media protocols within the G2 player window). RealNetworks was assisting in developing an SMIL version of the Freeb92player, so that there is a cross-platform version of the Freeb92 player. However this development was stalled near completion due to the war ending and B92 gaining access to the airwaves again. It would be great if Xchange, or someone within the network, could put some time into developing a similar application for playing Xchange material.
2. A cgi script 'database' could include diverse topics, from "how to find the right Realaudio encoder" (!) ... to providing adaptable scripts such for randomizing .ram-file contents etc...this would also help us to share information about new protocols. MP3, for example, is a protocol that we are all learning from the beginning. If there was a shared knowledge base we could understand and adapt to new developments a lot faster.
3. A cgi script that could be used to create ram files would be very useful. Ram files can contain other information which can be searched. Ram files can contain "comment" fields that do not effect the functionality of the ram file but can be used to include keywords for searching or content descriptions. If anyone has the time to do this and could make it available for all members of Xchange then we have the basis of a searchable system for Xchange content. The next logical step would be to have one centralised location for all Xchange ramfiles and create a simple cgi search mechanism. The cgi would search all the ramfile comment fields and return pnms directing the user to that content. |