[GNHLUG] [DLSLUG-Announce] DLSLUG Podcast Now Available

[Bill McGonigle]

Thanks for all the comments.  I can answer a few:

On Feb 17, 2008, at 12:28, Ben Scott wrote:

>   When people asked questions, I could barely hear that someone was
> speaking, let alone make out what they were saying.  As has been
> discussed before, source quality is key.  You can either mic the
> questioners, or have the MC or speaker repeat questions for others to
> hear.  This often benefits the live audience, too -- audience members
> behind a questioner sometimes cannot hear comments directed at the
> stage.
...
>> I'm still looking for some open source software that has a  
>> compressor which can operate on
>> an audio file based on a moving window, rather than the whole file.
>
>   I don't think I understand what you're asking for.  Are you talking
> about VBR (variable bit-rate)?  LAME can do that.

So, this is an interesting one.  Compression is a signal processing  
technique which can make the dynamic range of an audio waveform more  
'consistent', in this case the volume.  Details:

   http://en.wikipedia.org/wiki/Audio_level_compression

Now, in Audacity, if I select the questioner's question and a bit of  
the speaker's response (a small snippet), and run the compressor at  
-60 db, then both are quite intelligible.  However, if I run it on  
the entire file, Audacity seems to consider the entire large waveform  
as its analysis target, and basically does nothing to the already- 
normalized audio.  What I want is something that's going to look at,  
say, a 2-second window and do a running compression on it.  This  
appears to be commonly available in commercial tools but I haven't  
found an OSS version.  People have asked for it on the Audacity list,  
and IIRC, were told to go learn to program.  I expect this will show  
up eventually.

>  So I spend lots of time wondering what I was missing.  :)  I suspect
> coupling the audio recording with the original slides/movie/whatever
> the presenter was using would be a huge improvement, without wasting
> resources on a video recording of a slide presentation or talking
> head.

Right, I tried doing video once and it was majorly labor intensive to  
get decent results.  Volunteers welcome.  :)

Would the slides be worthwhile without an obvious clue about when the  
'turn the page'?  Anybody know about a wireless USB presentation  
device that makes a chirp when you click?  Or some better idea?  One  
not better idea: me listening to the whole lecture and doing anything.

>   As I'm sure you know, you should explicitly identify your copyright
> policy.  Can this be saved for later replay by downloaders?
> Redistributed to others?  Used for derivative works?  Redistributed
> via commercial services?  Keep in mind that you should (as a matter of
> good taste, if not legal requirement) ask each speaker for their
> desires regarding such.

Ah, good point.  I did one of the CC licenses last go around.  I'll  
pull that out and add it to the template.  I guess as-is, it's Berne- 
convention'ed, so you violated copyright law by downloading it. ;)

I will add this to the standard set of requirements for a DLSLUG  
presentation.  I don't want to deal with different licenses for each  
presenter, but will not record at the speaker's request.

>   An introductory audio lead-in should be included, to identify above
> copyright policy if nothing else.  It would also be nice to know what
> the heck this podcast is about.  :-)  If the "live" introduction is
> not suitable for podcast, record one separately.

Agreed - in this case the presenter was running 30 minutes late, so  
we skipped the intros and threw him into the den as soon as he walked  
in the door so we could get out on time, ordinarily I read the  
presentation from the announcement.  Most people haven't actually  
heard it before. ;)

>   It's also good form
> to have a lead-out that concludes the recording, so listeners know
> they got the whole thing.  This is also the perfect place for a
> trailer -- identify next month's presentation, so people are
> encouraged to "tune in next time, same time, same URL".

Good idea, this can be roughly canned too and strung together with  
sox or something.  I'm sure it's tremendously obvious, but I'm trying  
to make this not a major chore for me, so each little bit of canned  
piece or automated script makes it that much more likely that it'll  
actually get done.  And hopefully we can then `cp -a` for other  
LUG's.  I've got a decent-sized patch to send back to podcastamatic,  
e.g.

>   Given that this was basically a spoken word recording, have you
> checked out the SPEEX codec?  In my limited tests, it yielded
> *significantly* smaller files vs MP3's of the same perceived audio
> quality.

Hmm, very interesting.  Any idea how well distributed the codec is?   
I went with MP3 over ogg because there's a 100% chance everybody's  
digital music player can handle it.

To answer another question I got, this was a Nady-brand wireless mic  
that I picked up from DAK Industries.

Ah, here (sunglasses advised on this markup):
   http://www.dak.com/Reviews/2030Story.cfm

And it was recorded into the Mic-In on my Macbook Pro straight into  
Audacity.  This was recorded under OSX - I have a friend who has  
tried all the studio-oriented linux distros and came away with  
recording gaps on all of them.  I haven't had a chance to play with  
the RTL patchset yet to see if that would help.  And maybe his HP  
hardware sucks, but experimentation can come later, leaving me with  
an imperfect but working solution for now.

-Bill

-----
Bill McGonigle, Owner           Work: 603.448.4440
BFC Computing, LLC              Home: 603.448.1668
bill at bfccomputing.com           Cell: 603.252.2606
http://www.bfccomputing.com/    Page: 603.442.1833
Blog: http://blog.bfccomputing.com/
VCard: http://bfccomputing.com/vcard/bill.vcf