I was kicking around the idea of doing a station. That led me to working out music programming and genre/market positioning. That pushed me toward the potential of running a syndication network providing after hours content strictly for part 15s. That got tossed as more licensing hassle than what it's worth.
On the fun side of things, I've been experimenting with low bandwidth streaming and free software. The stuff is damned impressive. I'm still novice with all this.
My stream starts as .ogg files specially prepared. I use Q zero in vbr mode which sound to me to be as good or better than cbr mp3s encoded at 128.
Before encoding, source files are run through an AGC DSP to normalize volume between songs, plus lift quiet parts so the don't get lost in noise during playback.
After DSP, I run songs through a trimmer that cuts intro and outro silence at a threshold of -21 db.
Bah, browser screwed up.
The reason for all the preprocessing is there is no recoding of the stream when it is sent to the server. A program called ezstream reads a playlist and streams the source files directly to the server. There is no fade on segue, just a blunt cut. With volumes and mix point level stanardized, the cuts are amazingly clean. The listener hearswhat sounds like automation, but the overhead is very low. You could easily run ten streams T once
Sorry again...
What I was trying to say is that given low overhead, a modern pc could easily run ten streams at once.
So what the setup looks like is easysream sends to icecast. Icecast sends it to the listener. And it sounds like automation except for not having liners and promos.
Fun stuff to play with.
If you have an icecast ready player, here's a link:
http://radio.xap.net:8000/chilldown.ogg
Stream is very low bandwidth. 56k to peaks around 70k.
Listening now, Ken. Interesting stuff.
If you can come up with a way to segue/overlap the tunes and not have the tails cut off, I think you're on to something.
Cuts aren't doing as well as I hoped. I think I messed up the trim because I trimmed then did the agc. Even if I cut 21 db down, the agc might be pulling the volume up which blows the seg. Recoding a bunch now to test.
Ogg sound quality blows me away. It's not aac, but much better than mp3 at low bitrate.
I've discussed this elsewhere, but here's what I do.
I use Virtual Audio Cable to transfer the sound stream to various applications - it's not free (a nominal cost) but it's invaluable at hooking stuff together.
I use VST (free) as a wrapper, with various plug-ins, to process the audio. Right now I'm only using a compressor, to normalize audio and give it some 'punch'. I then generate an MP3 stream using standalone Edcast (again free), at 40kbps mono. Icecast (free) is used to serve out the stream. It sounds pretty good to me.
Currently, everything is on one box (running Vista - definitely not free), but shortly I'm going to move IceCast over to its own box running Windows 2003 Server (again, definitely not free). I find the Windows Server products to run applications such as IceCast (as well as IIS) much more efficiently and effectively. Plus you can add IIS services such as FTP without worrying about bogging down the computer.
A.R., part of your message might have gotten chewed up there. What VST wrapper are you using?
On the topic of compressors, I encourage you to check out the free "Limiter No. 6" ( http://vladgsound.wordpress.com/plugins/limiter6/)
Artisan turned me onto Virtual Cable, and it allows sending the local monitoring output of my streaming playlist to both USB Speaker and HD Audio Output 2 at the same time. Without Virtual Cable those outputs can only be used one or the other, but not both.
USB Speaker Output sends a signal by Wi Fi, using an MCM Wi Fi Audio device (no longer available), a distance of 35-feet to the AMT5000 transmitter for re-transmission on 1640 kHz.
The HD Audio Output 2 sends audio over a 25-foot cable to the AMT3000/loading coil transmitter on 1550 kHz.
It's called VSTHost. I use something a plug-in called Simple Compressor. I've also used a stereo to mono converter - it's supposed to eliminate phasing issues but I didn't notice any difference between using it and just simply mixing the stereo output to mono at the transmitter end.
I use streaming mp3 as virtually every media player can play it natively (with no additional codecs required). Ogg might have slightly better quality, but like I said, I'm pretty happy with the way 40kbps mono sounds with the material that I broadcast and stream. A lot of it is mastered from 78's, 45's and vinyl, and certainly doesn't have the dynamic range of, say, digitally recorded classical music.
At the end of the day, I use the simplest stuff to get decent sound from my particular programming. KISS in action.
Quick follow up. I did have pre-processing screwed up. I recoded a smaller music collection and it segs properly.
It just goes to show that "good segs" can take a lot of effort. 🙂
Here's a stream with non overlapping segs fixed. The music isn't going to be of anyone's liking, but you can hear the segs are much better.
http://radio.xap.net:8000/ethnic.ogg