Thursday, March 25, 2010

An Open Letter to Skype on Behalf of Audio Engineers

Dear Skype,

I'm writing to you on behalf of audio engineers, voice over performers, and project studio owners the world over.
Techies like me were very excited about your January 2010 announcement regarding HD video streaming, and having played with it, I was quite surprised by how well it worked on even my modest home internet connection (I only support a 1mbps up stream).

As well as the service worked, however, it still left me feeling a little hollow. As recording engineers, we regularly have to record talent remotely, sometimes even up to half a world away. In this day and age of high speed business broadband, what do you think is still the primary way that this recording is facilitated?

ISDN

We use phone lines to send "high quality" audio to other studios across the country and the world. Plus, this "high quality" usually maxes out around 128K AAC (most recordings are actually 64k). European cellphone video chat likely sends as high a quality audio stream (though lacking our studio and gear of course).

There are of course expensive and proprietary solutions that use data networks to send audio, but adoption has been slow, either do to initial teething pains in the services or high costs (even if those costs are lower than ISDN).

It got me thinking. If only there was a service in place that could create a point to point (or peer to peer) connection capable of sending a higher quality audio stream than ISDN.

And that's where you come in.

See, if you're capable of linking our computers to engage in real time 720p video calling at 30 frames per second, SURELY you'd be capable of granting us a 256kbps or higher audio only connection.

You would become the hero of studios nation and world wide.
You could single-handedly lift the entire recording industry out of the data dark ages.
And We would gladly pay for a stable service.

I hope you take this proposal under advisement.
Love,
SomeAudioGuy

18 comments:

  1. Let's cross our fingers the skype engineers see this. We can all dream right?

    ReplyDelete
  2. [Fingers Crossed]
    "PleaseOhPleaseOhPleaseOhPlease!!!"

    ReplyDelete
  3. Is that true? Does the quality of audio from ISDN only get up to 128K? What about clients who only deal in high quality wavs and do ISDN sessions?

    ReplyDelete
  4. Hey Darren,
    Yup 128K is the sad truth of dedicated phone line ISDN. Many Audio Broadcast manufacturers are incorporating broadband data into their codecs now too (even Telos), but as mentioned they're quite expensive.

    Now this 128K isn't DIRECTLY comparable to say a 128K MP3, as the methodology of data compression is a little different (as major a difference between MP3 and AAC), so "quality" can be something of a subjective term, but yeah data on a traditional ISDN hook up is heavily compressed.

    It's just a matter of bandwidth, and phone lines don't have a lot...

    ReplyDelete
  5. Wow, interesting. I didn't realise that the quality of the audio would be so low, I just presumed it would be top notch from ISDN.

    If I gave my agent a 128K MP3, she'd hit the roof! So lord knows why it's standard when doing sessions via ISDN.

    You're right about Skype, if we were able to send at least 128 down the line that would almost end ISDN as we know it!!

    Out of interest, am I right in thinking that at present, we're unable to record audio & video conversations on Skype?

    ReplyDelete
  6. And like I said before. Quality is rather subjective. 128Kbps actually isn't too bad on the human voice. We're not a radically dynamic "instrument", so removing info from a recording of a speaking human voice tends not to be TOO destructive (though it is why low bitrate recordings of strings or pianos sound atrocious).

    For auditions 128K is actually pretty reasonable. I think one of the strenghts of ISDN is it's typically too expensive to be utilized OUTSIDE of a proper studio environment. Going from a well treated/equipped booth to a well treated/equipped control room can do wonders for an audio stream even if it is being compressed rather harshly.

    As for Skype recording, there are plugins that work with varying degrees of success. I've found the best success just using Skype with a small mixer and routing the audio out into another chanel of my sound card. That has given me the best case results, and takes some of the load off of your computer (running Skype, a DAW, and an interceptor can be taxing even to sturdy workstations).

    Though of course, Skype isn't ready for prime time audio recording...

    ReplyDelete
  7. On a mac I've found Audio Hijack Pro to work very well. I'm not sure that it would capture audio in both directions, i.e. a whole conversation, but you could certainly grab the sound coming out of your sound-card.

    ReplyDelete
  8. I think what would be rad, is just having the higher quality audio stream, being able to run a couple outs to my mixer, just like I do for ISDN, to record output.

    I'll have to check out audio hijack though, thanks!

    ReplyDelete
  9. Great Letter! Lets hope skype listens!

    ReplyDelete
  10. I understood Skype to be somewhat akin to a torrent, but sharing "pieces" of a stream rather than pieces of a file.

    Part of the "deal" with getting Skype for free is that you "agree" to allow other Skype users' stream to bounce off your node. Your PC acts as a relay for other Skype callers.

    At lease, that's the way it used to be. Not sure if this is still their model. If it is, I'm not sure how effective that will be for lossless audio.

    ReplyDelete
  11. Hey Chip, thanks for the comment!
    As far as I understood it, Skype was built on P2P tech, but ONLY creates a two point connection. While you are signed on, you are NOT trafficking other people's packets. It would be pretty easy to check by signing into Skype and checking your network traffic.
    I think skype to skype is free because your computers are doing all the heavy lifting, and SKype out is a pay service because THEN you have to use one of Skype's switchboxes.

    I have no illusions that the service would be capable of lossless audio, but we should be able to match (and exceed) ISDN compression online by now! Especially from a service that can stream 720p video...

    ReplyDelete
  12. Skype - a great phone company which (unlike Paypal for example) has NO PHONE NUMBER! Therefore the ONLY way to get something through to them is through a forum such as this. I really do hope they are "listening" because it would sure be a boon to thousands of us "home studio" VO artists!
    Gray

    ReplyDelete
  13. Wow Tito, a phone company without a phone number, that's pretty meta.

    Skype = the sound of one hand clapping...

    ReplyDelete
  14. I've started a new blog carnival for all things audio and I wanted drop a line to encourage you to submit a post, or just keep us in mind and maybe check in and read what I hope will end up being a good sampling of posts from bloggers interested in all types of audio and audio production. If you're interested you can submit posts here

    Jennifer
    NCH Software Home
    NCH Software Blog

    ReplyDelete
  15. Great idea guys.
    re. auditions, I record all my auditions at 128k mono. Thats as high as SoundForge goes. You don't need any higher for auditions. In fact V123 only goes to 96k, so 128 is fine. ISDN keeps on hanging in there!

    ReplyDelete
  16. Integrity and stability is what ISDN gives you, and that is why it has stuck around. TCP/IP can't really work in real-time either.

    If 128k is an issue, why not use ISDN BRI/PRI? You could then transmit 320kb/s audio for example, enough for any type of speech.

    Another option may be for the remote studio to send a buzz track to sync your timecode and for you to then send them the audio files in BWAV format.

    ReplyDelete
    Replies
    1. I've heard this stability argument before. Yet I still encounter connection issues, delays of up to a second or more, and occasional degradation of signal. I know that IP solutions have issues with dropping packets, but if we can suffer a half second delay in ISDN, why can't we use a half second buffer on a digital solution? Coupled with a peer to peer connection with both nodes connected to >1Mbps upload, there shouldn't be any issues delivering a production quality 256Kbps audio stream.

      ISDN BRI seems to have the same issues web platforms seem to have, namely lack of adoption. I don't know any project or VO based studios using it. If I'm going o roll people over to something, I'd rather it be something as ubiquitous as having the internet. Especially with how reluctant phone companies seem to be to support ISDN copper...

      Delete