Hello.. I'm looking for a little feedback on those who have experienced and used the USB sound device and realtime CPU requirements for the DSP routines.. Reading the documentation suggests that I MUST use at least a 1.3Ghz CPU for usb_chan applications. The dilemma I have is that I have a beautiful rackmount AMD Duron 1.1Ghz CPU server sitting right here that would be awesome for the application if it will keep up. It's going to be a very light use system (testing only) and single user at a time type scenario.. not multiple connections or linking to other repeaters etc. This will not be the production multiuser system. * I will be using non compressed codecs (G.711/WAV/PCM) wherever possible for voice quality Internet bandwidth is not a concern on this system. Do you think I'm still going to run into a CPU bottleneck with a 1.1Ghz CPU for this use? (I know just try it out and see... and I'm going to) haha. What part of the code eats the most CPU? I ahve a lot of experience using Asterisk 2.4 on older P2/400Mhz class systems and I have had a lot of surprising performance out of them in terms of multiple users compressed codecs (GSM/G729) trascoding meetme conferences and IAX trunks all going on at the same time without issues and seeing relatively low CPU usage considering all the activity and bridging transcoding goign on. I must admit I've never used asterisks built in inband DTMF detection because I've never been able to get it to work reliably even on a lighning fast 3.5Ghz Xeon. Might just be that I've not taken the time to work with it because RFC2833 works flawlessly for VOIP. (getting a little side tracked there sorry). Take care and THANKS for sharing your thoughts & experiences! Steve -- This message has been scanned for viruses and dangerous content by MailScanner, and is believed to be clean.