We fought you off for as long as possible, but in the end we had to give in 🙂 What decided it was that OSnews is not a mainstream source of news like Engadget or even sites where they actually employ people, we are not an entity like AOL that fails to grasp the reality of the web today and the issues surrounding the technology landscape. We have the technical background, and the community to boot to know that when we criticise others, we had better be doing the right thing ourselves. Get the feed here.
With my work utilising HTML5, I understand the importance of free access to media, and the huge furor of news that was generated over the iPad’s announcement regarding the lack of Flash, and ergo what that meant for HTML5, and ergo what that meant for H.264 licencing issues.
The case for OGG is not so much about hardware availability, as it is for hackability. Openness means that no one person is dependent on one provider for any solution, and I know—however unlikely it is—that someone in a developing nation (or even in the impoverished parts of developing nations) can take OGG software and our content and put together a solution that works for them, perhaps even invent the Next Big Thing™ in media (even the adoption of podcasting is a new thing). The truth is that you never know where the next innovation is going to spring from when you have an open playing field.
The community at OSnews is skilled enough to not require their media handed to them through layers of DRM and Approved Channels® and can hack together any solution that suits them. Providing OGG audio (and hopefully video in the future too) is just a small step—by the volunteer team—to open OSnews up to more of that creative hacking. We have a number of things in consideration including the moving to CC-BY licenced content, a JSON API and clear and well-documented HTML/JS/CSS that will allow for inventive user-styles and greasemonkey scripts.
The Flash audio player has been replaced by an HTML5 audio element. Please report any problems you experience.
Doing the actual transcoding to OGG and uploading has taken over 20 hours to do, all I can say is that I hope it all goes to good use, I’m knackered.
Providing OGG is really great for a site where alternative OSes are covered.
Thanks for all the work to make it happen
PS. So does this use some kind of Audio For Everybody? (as in either provide MP3 or OGG depending on your browser)
It’s just an Audio element, no fallback. I haven’t developed an AfE, though I got an e-mail a long time back from a reader of my site who created an AfE chunk, but I haven’t tested it broadly yet to see if it fits the bill.
Time permitting, I will look into making AfE, but it does take time (some 50+ browser configs to test).
Correction: I should add that it’s an Audio element with the MP3 and OGG sources, so assuming that all HTML5 browsers can play those, it should work. But it’s not been tested at all. If there isn’t universal support for OGG and MP3, there’s no way I could provide a third format (like WAV).
The in-page player is a minor thing anyway, as it stops playing when you navigate away from the page, so it’s hardly useful.
Edited 2010-02-16 14:12 UTC
I remain convinced that an “ogg” version adds 100% more time to the editors job as they podcast, and satisfies the hungry need of about 6 people.
Sometimes, it’s not about catering to everyone’s wish, but rather, about providing the most realistic and widest reaching solution.
I suggested that perhaps we also encode in FLAC, the leading lossless format, AAC, which features far better compression, and for older hobbiest PCs, Real Audio.
No sense in not catering to the whim of EVERY contingent of single-digit group demands.
Edit: Nonetheless, we’re doing it, and we’ll either be proven to have caught on early or to be guilty of wishful thinking.
Edited 2010-02-16 14:12 UTC
I don’t think that that should always apply to the OSnews community, it’s just not mainstream—and we shouldn’t tar them with the same brush that AOL would. We could drop IE support tomorrow and it wouldn’t affect that many people.
The solution to the chicken and egg problem is to keep providing chickens until you get an egg. It’s like podcasting—you don’t have a big audience with the first episode, sometimes you have to slog it for years making episodes that only a handful of people listen to until you hit the big time. The same is of OGG. If we just stick with it long enough, the audience will come. I hate the fact that it increases my workload, doubles the upload time and the tools to work with OGG are a joke, but that’s the reality and it won’t change until we’ve done the slogging _first_.
I agree with you Adam, but it’s necessary to find out what’s possible with the community. You have to throw it out there and see if it sticks.
(Since you mention working with OGG…) You wouldn’t happen to know a good tool for converting OGG and other such audio formats into something standard, would you?
‘standard’, heh.
Anyway, mplayer/mencoder is basically a pretty good everything-to-everything converter for audio and video.
If you’re on Windows, CDex should do the trick with the “Re-encode compressed audio files” option.
Bull, because if you want to go HTML5 then OGG is the widest reaching solution.
But sure ignoring standards is more convenient.
BTW: If you have shell access to the server you can convert the MP3 to OGG on the machine. That would cut your upload time in half.(And I guess the Linux tools are a lot more mature than the OSX ones)
In the case of GUI audio tools, they definitely are better with Vorbis support on Linux (not really surprising). In the case of the cli all platforms are pretty much on par since the standard vorbis tools apply in most cases.
Supported in IE? Nope.
Supported in Safari? Nope.
Supported in Opera? Nope.
Supported in Gecko? Yes.
Supported in Chrome? Yes.
THAT’S the widest reach for audio in HTML5? Actually, the widest support is still Flash.
Quote my whole sentence, not just parts of it.
Major HTML5 browsers:
Firefox
Chrome
Safari
Only Safari does not support OGG and it is the one with the smallest market share.
C’mon dude, don’t play that game. You did NOT say “major html browsers.” You said HTML5.
??? So?
I said “Bull, because if you want to go HTML5 then OGG is the widest reaching solution.” and going HTML5 makes a HTML5 browser a requirement, doesn’t it? HTML5 needs HTML5 browser support.
Then I listed the major HTML5 browsers to show that amoung those OGG support is something like 80% there. (Very rough guestimatin. Chrome + Mozilla = 30%, Safari = 5%)
IE8 implements some HTML5 (cross-document messaging and “non-SQL storage”). It also lays the groundwork for additional HTML5 support. But in the classic sense, it doesn’t implement new elements, such as video or canvas.
But to say that IE can’t handle HTML5 and therefore was exempt from your statement is stretching it, because IE is still the dominant browser and can still graceful fail to some subset of HTML5. You weren’t suggesting we lock out 75% of the market, were you?
In this context I am obviously talking about the media tags in HTML5 and only Safari, Chrome, Opera and Firefox support those.
And IE has 65% market share worldwide, I don’t think it has 75% on OSnews, probably more like 30% or even less. Do you have any real numbers (facts) about browser usage on OSnews?
And you could help IE users and suggest to them to install Chrome (frame) or Firefox.
Isn’t bringing Flash into the equation a bit unfair? Isn’t there also a plugin for IE that allows Ogg/Theora usage via Silverlight for example? Or even Chrome Frame? AFAIK, there are no browsers out there that have native Flash support either.
True, but Flash is pre-installed on many OEM machines, but OGG, to my knowledge, is not pre-installed on any major manufacturer’s distribution.
Ogg Vorbis and Ogg Theora are supported out-of-the-box for most desktop Linux distributions (including the software shipping with the One Laptop Per Child XO) and there are several vendors (e.g. Dell) selling machines with Linux pre-installed. Plus, there is the embedded Linux mobile market. It may not be as sizable as the Windows market but it is still significant.
Ogg Vorbis and Ogg Theora are supported by VLC, which is the most cross-platform media player there is.
http://en.wikipedia.org/wiki/VLC_media_player
There is no need for the sizeable Windows market to miss out when it comes to Ogg Vorbis and Ogg Theora.
Edited 2010-02-17 00:22 UTC
“Opera is committed to both Ogg Theora and Ogg Vorbis for video and audio respectively.”[1]
–David Storey, Opera Employee
[1] http://my.opera.com/chooseopera/blog/2009/10/16/opera-talks-david-s…
Flash is supported on no browser natively.
The widest support for audio is ogg.
http://en.wikipedia.org/wiki/Vorbis#Application_software
For full platform support, copy then paste the podcast URL into VLC.
http://en.wikipedia.org/wiki/VLC_media_player#Features
If that doesn’t work, then add a VLC plugin for your browser:
http://en.wikipedia.org/wiki/VLC_media_player#Browser_plugins
You should be able to get it to wrok on any browser on any platform.
i fail to get this part, i thought the podcast is recorded first in whatever format it is recorded in and then transcoded to the final format or its transcoded on the fly at record time. How does it add more time? ..isnt this one of the those things? you set them up once and keep on using the set up? isnt this what computers are best at?
sometimes its best to be where everybody else because its easiest and sometimes its best to be inconvenienced to try to influence the future.
There is no sense in this but there is sense in trying to influence were the future is going. Most things starts at single digit shares and they grow from there. Its not easy to know which one will strike big and when but it doesnt mean we shouldnt try to support anything until it hit big. There is value is having a goal and work at achieving it. My hat is off for OSNEWS here
Nobody can predict the future, ideas need to be tested in the field and you guys are a part of that test ..regardless of how things will turn out, you guys will be winners to most of us who know what is at stake here
If you surround yourself with one kind of person, you start to think you’re viewing reality. OGG may seem important in this little silo, but for the rest of the world, no one gives a shit, and H264 and MP3 are fine, for better or worse. You don’t have to be a visionary or an envelope-pusher on every single cause.
But the fact is, OGG seems to be like RSS… it will take a LONG TIME before it catches on outside of a select crowd. Until Windows and Mac machines (or IE and Safari) ship with an OGG decoder, this is all just piss in the wind.
For me, this is the perfect example of why shouting loudly is a crowded room sometimes pays off – even if the result is a pile of MEH.
To say this is pointless would be to understate how bigger mugs you guys seem to be.
Edited 2010-02-16 15:37 UTC
What does that even mean? Are you mocking us for delivering the solution you requested?
I never asked for OGG; I actually said the same as others – MP3 is far more widely supported, AAC is a superiour format.
Yes. A “mug” is a “fool” and being a mug implies someone/thing is foolish. Kroc would have been able to translate.
I know what a mug is, but sorry, that sentence doesn’t make sense… at least in English.
Let’s break it down then:
To say this [i.e. OGG feed] is pointless would be to understate [i.e. stating that OGG is pointless understates, belittles, does not give enough emphasis] how bigger mugs you guys seem to be [i.e. it is being impled that you are giant mugs for creating an OGG feed for such a small audience.]
Or to rephrase it in a more American English compatible way:
Saying “this is pointless”, seriously understates how badly you have been duped.
Well, now, in my hypothetical free time, I’m going to explore converting that file to text automatically via speech detection and putting it up in a wiki for people to correct the inevitable mistakes. It will make it more accessible to shmoe’s like me that can’t sit still and listen to audio for very long. Plus, I’ll get to explore open source speech recognition.
If it works well enough, you’ll probably have more than 6 extra listeners/readers.
Good luck getting through my accent and mumbling. It was very difficult transcribing episode 31.5 and I was listening to myself! I could not imagine any software being able to transcribe our podcast given the accents, the speed that we talk and the exceptionally poor quality provided by my mic, Skype, and the OGG/MP3 bitrate.
If you can get software to transcribe our podcast accurately, you could make yourself a very rich person!
Enter Google
Seriously, with their Google voice transcription service, and free Google 411 service, they’ve really been refining their speech recognition software in the recent years…
How well it deals with accents, I don’t know, my wife already complains that it never seems to work right when she calls the 411 service
…and what about people who speak Foreign. Like me.
I’m aware of the inherit problems with the approach. I can run all sorts of audio filters to try and clean up the sound as best as possible. But even then, yeah, I’d predict failure too. It will be fun to try. Any level of success would be considered a bonus.
Plus, considering I’ll just be using open source software, I’m don’t think I should even deserve most of the credit if it happens to work, much less any of the revenue that could be realized by accurate voice transcription.
Adam,
It only adds 100% more time (by the way, you should be more specific since 60 seconds is 100% more time of 30 seconds) if you choose to still provide the audio in a proprietary format in addition to an open format. Also, I realize that you are being sarcastic about FLAC as well as the other codecs but if you were to (only) provide the audio in FLAC (which is also unencumbered like Ogg Vorbis) then your listeners could transcode it into whatever format their hearts desire.
By the way, it would be interesting to log and publish the number of downloads for both the MP3 and Ogg Vorbis files so that we could see for ourselves if Ogg Vorbis is only satisfying the “hungry need of about 6 people”.
In any case, I still appreciate your effort as part of the OSNews team to make the transition to an open format possible. Thank you.
It’s one thing to provide an ogg feed for a small group of people clamoring for it, it’s another to raise the barrier of entry by removing the de facto standard altogether. Removing MP3 means taking our toe out of iTunes, removing the ability for our group of corporate users without access to add new codecs, etc. Removing MP3 would be moronic, especially given that we are able to legally produce and distribute them.
i dont think anybody is for having only ogg feed. People are happy that osnews is playing a part in promoting an open standard. It is one thing to force ogg on people’s throats and it is another putting front and center as a recommended format and give alternatives for those prefer it.
Putting one format as a recommended and default format and another as an option to it can go a long way in supporting ogg
Cool, then we’re on the same page. We do offer OGG, but the default format is MP3, because well… there’s no central pingable service for oggcast updates, is there? I’m not aware of any.
As one of the (probably not actually single digits) number of users who will prefer the OGG, I’d like to chime in and agree that OGG + MP3 would be the best option, but I would vote for setting OGG as the default.
Then we would have one more thing of interest to the OSNews community: an exact figure of how many users “need MP3”
I just want to say thanks for stepping up and doing this. For me it wasn’t about the ability to play it (everything can play mp3 after all) but following the standards the editors themselves have advocated. I’m glad to see this and I’ve got the ogg feed in Rhythmbox now.
A question though: How does doing this double the workload? It’s not as though you have to record the podcast twice. Is it the double encoding or the actual uploading process that’s the hassle here? The double encoding, at least, could be scripted relatively easily.
And if you’re using OS X, as I believe you are Kroc, I agree the ogg tools at least for the GUI are pathetic. I have to drop to the cli and use the standard vorbis tools to do anything productive with ogg on OS X, or else use an audio editor like Amadeus Pro that has proper ogg support implemented via the vorbis libraries. Ogg encoding in GB or Quicktime, might as well just forget it. If I need to do multiple encodes of a gb project I save it out to AIFF and run the encodes in parallel in a second step. At least that can be scripted via shell scripts plus Automater.
Isn’t there a version of Audacity for OS X? It should be able to export/convert to OGG.
There is but it doesn’t work very well when compared to its other ports. You get the spinning beachball far too much and it’s slow compared with Mac-only editors like Amadeus or Sound Studio.
That’s a shame… I hadn’t used it since the 10.4 days, but it worked fairly well back then.
It doesn’t double the workload, it doubles the upload.
The process I’m using is:
1. Record show with Skype and put audio into GarageBand, add metadata there. (about two-three hours)
2. Edit show in GarageBand (about one-three hours)
3. Export show in AAC/64 which retains the metadata, MP3 export does not. (about 10-30 minutes)
4. Use Max http://sbooth.org/Max/ to convert to MP3 and OGG at the same time. MP3 takes 3-6 minutes, OGG takes 10-30 minutes.
5. Upload–anything up to an hour depending on how cruel my ISP wants to be.
THIS IS A LONG PROCESS. I do wish people would stop simplifying just how much work this really is, considering this is on my day off when I really should be outside and if not outside, working on OSn5.
Edited 2010-02-16 19:02 UTC
I guess, its surprising there is that much editing. But also, couldn’t the conversion (to mp3 and ogg) and uploading be scripted? That would remove some of the pain.
Editing requires listening to the show once over sometimes, and whilst not all shows require much editing, some are a nightmare. It’s been getting better with the new recording method in Skype, but sometimes you’re hit by recording problems.
Next to no difference, it’s just drag and drop into Max. I could save 1 minute by scripting it.
Ubercaster for the Mac is a great all-in-one studio allowing you to record, cut, output and upload in one app. They don’t support OGG, but I would switch to that if it did because it would probably save an hour or more of time. I’ve mailed them to ask about OGG support, hopefully they’re game.
Ok, I think I understand except: why would Ubercaster save you an hour, but scripting would only save you a minute?
I appreciate the amount of time everyone at Osnews puts into the site. I don’t meant to be difficult, just trying to understand the process.
There’s a lot of time wasted by having multiple steps and apps in the chain. I’m using Skype, Skype Call Recorder, GarageBand, Max and Transmit. Using Ubercaster would reduce that to just Skype and Ubercaster and cut out a lot of inefficiencies (as long as editing is quick). Scripting can’t get over the fact that there’s lots of steps, it just removes a few drags and drops.
Ok, I guess I was just imagining you doing the editing in garage band and then running the script on the file garageband output which would convert it to mp3 & ogg then upload them. So when you were done with the edits you could walk away.
That is what I do, but whilst it’s uploading I’ve got to prep the article. There’s a lot of inbetween stuff that adds up. I have to split the skype recording into two channels, then put both of those two channels into GarageBand to begin editing (About 5-10 minutes in total). I have to export to AAC, and then convert that to MP3/OGG. All of that would be much quicker if I recorded in the same app I edited, and could export to the final formats, instead of using an intermediate and adding an extra step.
This is a real use case for open source software stepping up to the plate. None of this CLI chaining. People need a podcasting studio capable of making this task _easy_ and take as little time as possible. Don’t say Audacity or I’ll bite your head off Ubercaster have nailed the idea, but lack an OGG output.
I still don’t understand how it takes one hour of your time for uploading. Do you have to babysit the command line every second till the upload is finished?
Editing a podcast can take an awful lot of time, but for me converting it and uploading it in one format or ten doesn’t make any difference time-wise.
I’ve used self-produced audio for language lessons in the past. The recording takes a few hours a week (even though I grouped recordings together because getting a hold on native speakers isn’t easy), the editing takes between 10 to 20 hours (lots of dialogs, with each speaker recorded separately because it’s too hard to get them to commit to a schedule, etc…). And I also did that for free, since it was for an association, on top of a day job.
The conversion takes 2 seconds, time to open a command line and input #name_of_the_script file(s)
Same for the upload. Add a shutdown -h now at the end of the script and you can go to bed the second you’ve pressed enter, even with a dial-up connexion.
How can it take one hour of your time just for uploading?
And as said, if the bandwidth is an issue, converting the file on the server means you don’t use any more bandwidth offering 2 (or 20) formats instead of one (though I’d usually convert from OGG to MP3 instead of the reverse).
I can’t hit the publish button on the website until it’s uploaded–so it still ‘takes’ that time to upload because the podcast cannot go live until that step is complete. Yes, I can twiddle my thumbs inbetween, but just because the conversion has started and it will upload afterwards, it does not mean I can walk away and leave it and the article will magically write and publish itself.
It doesn’t always take that long to upload, but it depends on my ISP, they see a big upload as ‘pirating’ and cap the connection bad.
But is the world really going to crumble if you don’t publish it as soon as it’s uploaded, or is it just a lack of planning?
Basically, what problem would arise if you publish it in the morning instead rather than at midnight? I mean you could complain having to wait half an hour for the sole MP3 version as much as waiting 1 hour for the MP3 and OGG versions… so what gives?
What is this, the Spanish inquisition!?
Ok, so basically you OSNEWS staff can both pat yourselves on the back for being such a good example providing an OGG version, while at the same time discouraging others sites to do so by complaining how it increases your workload by 100%… but when explained how it doesn’t have to be so you’re just calling that the Inquisition?
Offering 20 different files format doesn’t increase anybody’s workload by 100% nor 1900% – if you choose to look at the upload bar for one hour, it’s your personal choice. However, if you then whine about your martyrdom for free format’s cause, don’t you expect to be called out on it?
As for the lack of GUI tools for editing OGG, first it wouldn’t save you any time if they did exist (using Audacity I spend more time in the menus switching the format to export than just typing the few letters the command takes in a CLI — ever heard of auto completion?), second they do exist at least for Windows and Linux, and you probably just don’t know them yet for OSX.
Note: I could care less for OGG or MP3. When talking about lossy formats for music, it makes sense to use AAC or OGG, but for a podcast anything would actually do the job. In this day and age, where we’re supposed to have learnt how computers can save ourselves time for fastidious tasks though, I can’t fathom how one could pretend that exporting audio in different formats increases his workload by 100%.
Edited 2010-02-17 22:17 UTC
I didn’t say it increased my workload by 100%, that’s a misquote. Do not continue that fallacy.
Because of the change in format it has added extra steps to the process, as well as the additional upload time. Adding another format adds time, there is no escaping this truth, despite this lengthy thread’s persistence that I am doing something wrong. It still takes CPU time to transcode and it still takes time to upload, and it’s taking more time than before because I have to export an intermediate format because GarageBand is so crap, and I can’t really use anything else because it either doesn’t support OGG, or is just plain crap and won’t save me time (Don’t mention Audacity or I will bite your head off).
So few stuff works with OGG that I can’t get an integrated solution like Ubercaster. This is the fault of OGG developers for not making a compelling case for software to support OGG and for the OGG tools being so obtuse. The QuickTime component on Mac is a year out of date and completely stagnant and encoding video with OGG is such an alien experience to most video engineers (who are not programmers) that few have ever heard of the format.
There are no compelling tools for OGG. Where’s the open source Ubercaster? Podcasting has been mainstream since 2004 and been around since about 2000 and still I’m expected to do things with the command line.
The fact is that working with OGG is frustrating and hard for most people who _just want to get shite done_ and I’m only providing OGG for the podcast in the hope that with time I will be able to help improve the situation.
We could just use Windows to record the podcast. The Mac is horrible when it comes to decent codec support, whereas DirectShow codecs are maintained quite well.
It’s GarageBand, if anything. It won’t export MP3 with tags, nor OGG. I have to export to AAC then use that to put out the MP3 and OGG. Doing it on Windows won’t help, that’ll add extra steps as well for me as my netbook doesn’t have the disk space to record, and it’s too cramped to do the editing on, so I’d still have to transfer to Mac.
We can change from one set of steps to another all day long but unless we significantly reduce the number of steps with an integrated tool (like Ubercaster), there’s no point, it’s not really saving any time. We could also look into the doing the OGG conversion server side.
Your last podcast (OSnews Podcast #20.X: The Apple Tablet) is using 44100 Hz and 96 kbps for voice. “Episode 31: In Which We Discuss the Inevitable” is 32000 Hz, 48 kbps. I didn’t check the others.
For voice, it’s overkill. You could easily go down to 22050 Hz for the MP3 and OGG and keep the same quality, while going into the 30 kbps range. It’s human voice, which means 44100 Hz isn’t buying you any extra quality. Or, in other terms, somebody with a better microphone would still be able to produce better quality audio with 22050 Hz or less and a a third of the bitrate you’re using. Just compare with FM radio audio programs recorded in studio – even with better quality than most podcasters to start with, 22050Hz is more than enough for them to sound better than a podcast.
You could produce half the size (or less) while keeping the same exact quality. Basically, you’d keep the same upload time (or less) you had when you just uploaded MP3, but you’d be uploading 2 different formats.
As for an audio editing tool in OGG, you don’t see one because any donkey who ever had the idea to produce one has long ago been taken behind the barn and shot. Nobody in their right mind would use OGG for editing – audio programs use a non-destructive format, and for good reasons.
GarageBand and Max don’t allow me to choose the Hz, rather annoyingly. I may have to script LAME and the MPEG2Theora to do the encode so I can get precise export options.
You should file a bug report at Xiph.org then, since it’s probably their fault. Or Adobe’s.
Number of mobile devices I own with audio capability: 5
2 Phones (Nokia and Motorola), 1 iPod, 2 Laptops (1 Windows, 1 Mac).
Number of devices that can play OGG out of the box: 0
Number of devices that can play AAC and MP3 out of the box: 5
OGG isn’t relevant to me.
Also: my CD-Player doesn’t play OGG, but plays MP3.
Edited 2010-02-16 15:24 UTC
What makes you think anyone cares?
Number of devices you own that are free and/or open: 0
Edited 2010-02-16 18:17 UTC
Dunno, maybe he stole them from the shop or something.
I did mean free-as-in-speech
Total number of devices I own with audio capability: 2
2 computers (Linux)
Number of devices that can play OGG out of the box: 2
Number of devices that can play AAC and MP3 out of the box or even legally: 0
MP3 isn’t relevant to me.
Thanks for the OGG support!
FFMPEGX is a great encoding utility for OSX. I wouldn’t be surprised if it has OGG plugins.
In spirit of Audio for Everyone, would you pretty please provide plaintext as alternative format for me?
I’m an OS News regular. Your podcasts may be spectacular but I wouldn’t know because I won’t waste my time listening instead of reading. Reading is faster and enables skimming which you can’t do with audio.
Ergo I don’t get why anyone would prefer audio to text (except as an alternative) so popularity of podcasts is beyond my understanding.
do you know of any tool that can produce text from audio feed? .. I think it is asking a bit too much to ask for somebody to listen to audio and then manually typing all words in the audio and proof reading it later on.
Most podcasts are usually given with timelines and you can skip only to sections you are interest in and they usually come with a summary so its easy to know if the podcast will interest you or not.
Then there’s the browser back button, isn’t there? .
Since I’ve studied conversation analysis in great detail at university (as in, in VERY great detail), I know what an absolute horror transcribing is. In case you’re wondering what it’s like, let me paint a picture for you.
You’ve found the girl or boy (or both, I’m Dutch) of your dreams. On your romantic getaway to Praha, you’re strolling along the Vltava (the Moldau for the language-impaired), and you have it all planned out.
The ring is burning a hole in your pocket, and with the moon shining bright, stars dancing in the night’s sky, your hand slightly shaking, you go down on one knee. You take the ring out, and you pop the question. Her eyes well up, and visibly trembling, she says yes.
*Just as you’re about to get up, she gets hit by Britney Spears’ tour bus, gets hurled to a lampost, and dies, her dead, shattered body curled around it. Right at that moment, your phone rings, you pick up, and it’s the police telling you your entire family died in a horrible car crash in South Carolina – he tells you they had a slow, agonising death by petrol fire.*
Transcribing is experiencing that section between asterisks over and over and over and over and over and over, while a huge smelly Russian guy is prodding you in your left ear with an uncooked hotdog sausage every ten seconds.
But hey, feel free to provide a transcription. We’d totally appreciate it.
Thom, I’d +10 that if I could. Having done transcription myself I agree with you how horrible it is. Well, actually, I don’t think you painted a horrible enough picture, let me add:
*After the call about the fire, you are picked up by the local Mafia to whom you owe gambling debts. They take you somewhere dark and pull out a pair of pruning sheers, proceeding to cut something off for every grand you owe them. As it happens, you owe them $21,000…
There. Now the picture is horrible enough.
I’ve always had good luck with speech recognition, the problem for controlling a computer has always been speed of input.
Worst case scenario, I’ll end up benchmarking open source speech recognition programs with charts with little indicators labelled “Bad”, “Horrible”, “Tax Fraud”, “Felony”, “Crime against Humanity (execution pending)”,”Deceased (executed during testing)”, “Ashes scattered to prevent zombification”.
Wow, didn’t realize the state of the transcribing software. Still very much oriented around research. I just looked at some of those benefiting from the voxforge project. Pretty much all of the software is still in the research stage where the most important part of the output is measuring how many utterances were matched to words in the library. The actual text output itself is almost an after thought. It will take longer than I thought to get a horrible result. No one hold their breath for a result.
For those interested,
http://voxforge.org/
Sorry, didn’t scroll down to read all the comments before posting. I’m looking into it. Ogg was the first step. Now, we can do it ourselves without waiting for Osnews staff to spend more of their time doing it.
Unless they script their podcasts, that would require a ridiculous amount of effort. Unless someone who works on the site is a professional stenographer, transcribing each podcast would probably take around 6-8 hours (for someone who can type 80-90 WPM).
What I like about the podcasts is that you can listen to them while doing something else, and still get the gist of what was discussed. It’s kinda hard to read while you’re washing dishes.
For some reason the ogg.php is not working for me, is it possible to add direct URL pointers to the OGG and MP3 files?
At present I am running the Chrome Browser on Windows 7 Starter.
Later in the week I will try this with Haiku/Firefox to see it it works better.
I am guessing PHP is not a standard component of Windows 7, but I am pretty sure it is now part of Haiku, so maybe it will work better there.
Edited 2010-02-16 16:12 UTC
PHP is entirely server-side scripting, whether the client os has it or not is irrelevant. PHP can be used to generate pretty much anything, in this case it generates a Podcast XML. You should point your Podcast aggregator of choice at the URL, it isn’t meant for browsing.
Btw, I think that’s the first time I’ve actually seen anyone say they were running Windows Starter of any version.
http://www.osnews.com/files/oggcast.php
It’s an XML file, it has nothing to do with Windows or PHP.
Then why does it have a .php extension if it’s not a php script? If it’s straight xml, give it a .xml extension.
What difference does it make? Extensions on a server are irrelevant: I could tell my server to use the extension .osnews if we wanted. The file’s content type is XML, that’s all that matters. This happens to be a php script that generates XML.
They’re probably using PHP to dynamically generate the XML file from a database. Most XML/RSS readers & clients don’t care about the extension, just as long as the correct content/mimetype is specified.
And that’s usually easier than the reverse (putting PHP inside a .XML file), because you then you don’t have to configure the server to execute the XML file as if it were a PHP script.
Yes, that was my assumption as well but I’m apparently wrong, as Adam said it has *nothing whatsoever* to do with PHP and says the file itself is an XML file. That’s why I was curious about the .php extension, I know it doesn’t matter but at the same time it’s generally a good idea to keep the extension relevant to the file’s content for your own reference if nothing else and that way you don’t at some later date go digging through files you don’t have to looking for code to update.
Clicking that link in Chrome on Windows just produces a jumble of text, but it looks good in IE8 and Firefox… I guess that’s a Chrome problem :/
You’re right, but it’s an RSS file! It’s not meant for you, it’s meant for a RSS reader.
Given that Earl’s a software developer ( http://www.bebits.com/devprofile/760 ), I suspect he meant “OGG” and wrote “PHP” as a typo/brain-fart.
I click on the link, and all I get is a bunch of runtogethertext that looks like it must be the XML you are talking about but I am seeing no links to audio, nor am I hearing anything.
My Chrome bowser says it is up to date. Could I have the wrong association to the PHP extension, if so how do I change/delete it. And if it is running on the server side why am I seeing all this text?
It is NOT for you, it’s for an application that understands it. It’s RSS. It’s not meant to be read by you.
If you want the Ogg files, you’ll need something like VLC and you can open the URL inside the app.
Well, that did not work.
Used Chrome’s copy link function, openned VLC, pasted the URL in the network option than after hitting return press play button. Still no sound.
Again where do I find associations in Chrome?
http://www.fsf.org/resources/formats/playogg/how
If you can’t figure it, I’d recommend you stick to the normal MP3 podcast in iTunes.
Edit: original came off too douchey, so I revised.
Edited 2010-02-17 19:38 UTC
Chrome simply refuses to display that RSS feed in a usable manner.
You’re better off going to a feed reader, like iGoogle or otherwise, and pasting the URL there. That should show the feed with clickable links to the audio files for each podcast.
If you can’t handle that, view the source for the page that Chrome displays and find the URLs directly…
Worst-case, IE8 and FF both seem to display the feed reasonbly well, just not Chrome (hint: complain to Google if you care).
If you want to see the link inline with the article for each podcast, you gotta go to the podcast article itself, for example:
http://www.osnews.com/story/22714/Episode_31_In_Which_We_Discuss_th…
Where you’ll see the HTML5 audio playback widget already embedded!
Edited 2010-02-17 22:07 UTC
To the team at OSNews,
Thank you for providing your podcast in an unencumbered format! Your decision to release the podcast in a non-proprietary format provides a positive example for other content providers who have not yet recognized the benefits–not only for themselves but for their users–of using open, patent-free technology. Kudos!
Firstly, thank you for providing the podcasts, it makes good listening (regardless of it being in mp3 or vorbis).
Please could you outline the podcast creation process. I like the idea of using open formats, but not the 20 hours extra work!
I’ve detailed the process here http://www.osnews.com/thread?409652
How can I listen this podcast on my Android phone?
I have tried with ACast and BeyondPod but it’s not working.
And here I was, thinking providing .ogg would shut everybody up.
I’ve been online since 1998, wtf was I thinking. Of course it wouldn’t shut anybody up.
why would you think that?
mp3 has its strength and weaknesses
h.254 has its strength and weaknesses
flash has its strength and weakness
ogg vorbis/theora has its strength and weaknesses
<insert audio/video format> has its strength and weaknesses
there will always be discussions and arguments about what format to use where and why. You should worry when people stop talking completely, and you should be if people start talking pass each other in strong and threatening tones.
You’re dealing with nerds here.
Nerds NEVER shut up.
Perhaps it would have shut everyone up if your own staff didnt bitch about it.
I doubt it, but thanks for making a veiled and passive-aggressive comment about me
While I wasn’t one of the people clamoring for OGG, I did point out that it’s inconsistent to provide a set of justifications for offering the podcast in MP3 only – after having dismissed (or refused to acknowledge) those exact same justifications when they were given as reasons why web developers use Flash for video.
I do think Kroc’s original justifications were valid – it sounds like a lot of extra effort was required in order to accommodate a relatively small portion of OSNews’ audience. And I don’t really see the practical reasons for demanding an OGG version of the podcast – it seems to be primarily an issue of principle more than anything else. It’s at least much less-dire than the h.264 licensing situation, I doubt Fraunhoffer will be sending any invoices or C&D letters to OSNews.
And on the same topic: it might just be my misunderstanding, but I didn’t get the impression that people were necessarily demanding that an Ogg version of the podcast be given equal standing with the mp3 version (although that could just be my misunderstanding). Personally, If I were in a position where I couldn’t listen to the MP3 version of the podcast, I’d be perfectly happy with just a plain vanilla download link (in the newspost for the podcast) for an Ogg version. I certainly wouldn’t expect the OSNews to go the effort of creating a separate XML/RSS feed for the Ogg version.
So kudos for going above and beyond.
Thank you for posting a reasonable and well thought out response. I agree, I see little practical need for an OGG feed, since I expect 99.9% of those demanding it have an MP3 decoder, and therefore, want us to go out of our way for principle. But we did it, and now some of those same people seem unhappy that I am saying so publicly.
That said, just so you know, the site generates little money, little enough that we can legally produce MP3 files. And H264 is currently royalty free as well, so we actually have no fear of a C&D coming our way.
Thoughtful and reasonable – now there are two things I’m rarely accused of
Agreed, some of the more vocal advocates of Ogg (Vorbis and Theora) apparently believe that they can “win” by ignoring current reality – and instead focusing solely on the way things “ought to be.” There’s a term that a friend of mine coined a few years ago: “Cause-itis,” defined as the state of having devoted oneself to a cause – and concluding that everyone else is subsequently required to devote all of their energy to the same cause (otherwise, they’re the “enemy”).
My understanding is that the licensing status of the encoder plays a much bigger role with h.264 than it does with MP3. In other words, you’re probably not going to incur any financial liability for distributing MP3s encoded with LAME – but the same isn’t necessarily true for MP4 files encoded using x264 (or so the MPEG-LA contends).
And to go OT/on a related note: have you ever considered cutting out the middle-man and selling OSNews ad space directly? It would require more effort, of course, but you’d presumably get a larger cut of the ad revenue – and you’d have more control over the content of ads that appear on the site. That approach seems to have worked fairly well for the Penny-Arcade guys for the last few years.
For making this available, but also for not being like the clueless driods at Engadget or AOL.
There isn’t such thing as “OGG”, since it’s not the abbreviation and it doesn’t follow “MP3” name logic:
http://en.wikipedia.org/wiki/Ogg#Naming
Edited 2010-02-18 15:45 UTC