<br><br><div class="gmail_quote">On Tue, Jun 28, 2011 at 5:47 PM, Bryan Smart <span dir="ltr"><<a href="mailto:bryansmart@bryansmart.com">bryansmart@bryansmart.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;">
Thanks, Anthony.<br>
<br>
I downloaded<br>
<br>
<a href="http://files.freeswitch.org/freeswitch-sounds-en-us-callie-48000-1.0.16.tar.gz" target="_blank">http://files.freeswitch.org/freeswitch-sounds-en-us-callie-48000-1.0.16.tar.gz</a><br>
<br>
and<br>
<br>
<a href="http://files.freeswitch.org/freeswitch-sounds-en-us-callie-32000-1.0.16.tar.gz" target="_blank">http://files.freeswitch.org/freeswitch-sounds-en-us-callie-32000-1.0.16.tar.gz</a><br>
<br>
I examined them with the statistical analysis tool in Sound Forge. Regardless of what DB scale is involved, both versions register in Sound Forge with identical max sample value and average RMS. This outcome doesn't seem possible if the 48000 set is sent through sox with "-v 0.2" to produce the other sets. Wouldn't the results of that processing be that the sets for other rates would be 20% as loud as the 48000 set?<br>
<br>
The 48000 set that I discovered must not be the correct set.<br></blockquote><br></div>I'll email you offlist and see if I can't get you read-only access to our sounds repo so you can see the raw files.<br>-MC<br>