Knowledge Is Power

How To Web Cast PA Audio & Interface A Public Address System Amplifier To A Sound Card On A Streaming Audio Server

Overview

This page covers how to interface a PA (Public Address) amplifier with the sound card in a streaming audio server or PC running a software encoder. After searching the Internet in vain for this information I developed the audio interface circuit shown in the schematic below by modifying a circuit I found for connecting a cassette tape recorder to a telephone line. I present this information here in the hopes that it will help prevent others from having to re-invent this wheel.

Why not just use Line-Out to Line-In ? Using a patch cable to go directly from the PA amplifier Line Out jack to the sound card Line In jack isn't always practical. Older amps may not have a Line Out jack. Those that do may over-drive a sound card Line In because there is no signal attenuation with a patch cable. (Line Out impedances are typically around 600 ohms while the input impedance of most sound card Line In jacks can be as high as 20 Kohms.) In other cases the PA system amplifier(s) may be located in a meeting room where the addition of a PC isn't practical. However, you should give it a shot before going through all the effort of breadboarding and then building a circuit, this may be an easy way out. If the audio at the PC seems over-driven, try adjusting the Line In volume in the Windows Volume Control applet on the PC with the sound card. PA systems typically distribute audio signals to a lot of different areas by using higher-voltage (70 volt) signals. This higher voltage allows for a low current which means the signal can be sent over long distances using small wires. Some installations may even use 25-pair telephone cable as a means of distribution. Often, it is easier to gain access to an audio signal cable than to the PA system amplifier(s). (Because higher-voltage PA system speakers are generally connected in parallel, you may even want to experiment with tapping off one of the PA system speakers to feed the circuit featured on this page.) If your installation has multiple amplifiers in different locations (see the Multiple Sources section below), but the audio signals from all locations are available at some single location, you wouldn't need a streaming system for each amplifier. Naturally, if you do have easy access to the amplifier(s) and there is network connectivity in the same area, you'll want to try a straight patch cable (and use the Windows Volume Control applet to adjust the input level) before you go through the trouble of building an interface circuit.

This audio interface circuit is used to "netcast" (over an enterprise network) and "Webcast" (over the Internet) the audio from committee, board, and other public meetings of local government to lectures at schools and college campuses. Our streaming is done using Windows Media Services 9 (included with Windows 2003 Server) but any streaming audio server can be used. End users listen to the net/web-cast using Windows Media Player or similar application. Notes on setting up Windows Media Services/Encoder are given later in this page.

The sound systems in our meeting rooms consist of typical PA microphones and amplifiers with a 70 volt speaker output. The circuit shown below can easily be modified to interface with a PA system with 25 volt speaker outputs simply by using a different transformer. PA amplifier manufacturers include Dukane, Bogen, and Realistic (Radio Shack). We tested the circuit below using unbalanced 70-volt output from a Bogen C-35 amplifier.

If you do plan to feed the circuit by tapping off of an existing PA system speaker you'll want to verify that you're dealing with a 25-volt or 70-volt system and you'll want to tap off the correct point. Speakers in 25-volt and 70-volt systems will typically have a transformer (similar to the one used in the circuit below) attached to them. There may even be text printed on the transformer with the voltage rating. Be sure you tap off the primary side of the transformer (the one the cabling connects to), not the secondary side (the one that attaches to the speaker itself). If you don't see a transformer connected to the speaker, don't tap off of it. You likely don't have a 25-volt or 70-volt system.

This setup can also be used as an alternative to extending PA speakers to an out-building or other nearby facility which has network connectivity.

I M P O R T A N T ! ! ! I am NOT an electronic/electrical engineer. The circuit presented on this page is intended as a sample to be used in breadboarding and experimenting ONLY! A starting point if you will. This circuit has NOT been evaluated for shock hazard, fire hazard, component overheating, or other safety issues. Your final circuit (circuit design, component values, etc.) should be the result of your own experiementing and testing. Additional experimenting may yield a safer design with better audio reproduction and overall sound quality.

The Circuit

The circuit used for breadboarding includes a step-down transformer which is the primary device for adapting the 70-volt PA audio signal for use with a computer sound card. The transformer used has 5 leads (taps) of varying power levels on the primary side each with a different color insulation. The lowest-power brown lead (.625 watt) was used to present a minimal load to the PA amplifier. On the secondary side the white (8 ohm) lead was used as the output (a 4 ohm gray lead is also available). There are black leads on both the primary and secondary which indicate the common connection. The only reason we used a 10-watt transformer is because that's the smallest size we could find. Smaller wattage transformers could also be used.

The 4.7 uF non-polarized capacitor on the primary side of the transformer is used to keep any DC current out of the primary transformer winding and ensure that the circuit doesn't represent any sort of DC load to the amplifier.

The .01 uF non-polarized capacitor on the secondary side of the transformer serves as a simple low-pass filter which filters out hiss and other high frequency white noise. While simple, the filtering effect is dramatic when viewing the output audio signal with an oscilloscope.

The 20 Kohm potentiometer is used to adjust the level of the output signal going into the sound card. Some sort of adjustment is recommended due to the variety of input impedances encountered with sound cards from different manufacturers. Average Line In jack input impedances range from 8 Kohms to 20 Kohms. A PCB (Printed Circuit Board) mount potentiometer was used instead of a conventional case mount type due to the fact that once it is adjusted to provide a proper output signal level, no future adjustments should ever be necessary. However, if the outputs from multiple amplifiers are going to be switched into one circuit (as covered in the Multiple Sources section below), a case-mount type of potentiometer may be called for.

The 470 ohm resistor is simply a current-limiting "safety" resistor to prevent a short across the secondary of the transformer (such as a shorted filter capacitor or output leads connected together at the same time the 20 Kohm potentiometer slider is adjusted all the way up).

Note that the two common leads of the transfomer are not tied together on a common ground. This is to ensure isolation between the PA system and the sound card. In addition, neither of the common leads is connected to a metal case (i.e. no chassis ground).

If the computer sound card offers both Line In and Microphone inputs, the output of the interface circuit should go to the Line In jack of the sound card, NOT the Microphone jack. The Microphone jack has additional amplifiers which can introduce noise and can more easily be over-driven.

Circuit design is simplified by the fact that no power supply is needed for this type of circuit. All necessary power is derived from the 70-volt AC PA audio input line.


Sample Breadboarding Parts List

Part	Vendor	Vendor Part No.	Price
Vectorboard Unclad Perf Board 4.5 x 6.5	circuitspecialists.com	64P44XXXP	5.95
20 Kohm High Rel Horizontal Mount Trimmer Potentiometer	circuitspecialists.com	32JR403	.85
470 ohm 1-watt Carbon Resistor	circuitspecialists.com	RG470	.14
.01 uF 100V Mylar Capacitor	circuitspecialists.com	23BK310	.12
4.7 uF 250V Dayton Metalized Polypropylene Capacitor	partsexpress.com	027-422	1.89
Speco 10-Watt 70V Line Transformer	altex.com	T-7010	3.95
2-position Screw-terminal Terminal ("Barrier") Strip	Radio Shack	274-656	2.29
3.5mm Panel-Mount Stereo Open-Circuit Female Jack	Radio Shack	274-248	2.99
20-guage Solid Hookup Wire	Radio Shack	278-1222	5.49
3.5mm Male-to-Male 6' Stereo Audio Cable	Radio Shack	42-2387	4.99
6/32 1/2" length bolts Quantity of 6	Local hardware store	n/a 2 used to mount the transformer 4 used to mount the input terminal strip	.06 ea.
6/32 1" length bolts Quantity of 2	Local hardware store	n/a Used to mount the circuit board	.08 ea.
Nylon Spacers 3/8" OD - 1/2" length (with center hole for 1" 6/32 bolts) Quantity of 2	Local hardware store	n/a Used to mount the circuit board	.19 ea.
6/32 nuts Quantity of 8	Local hardware store	n/a 2 used for the circuit board 2 used for the transformer 4 used for the input terminal strip	.06 ea.
No. 6 Flat Washers Quantity of 8	Local hardware store	n/a 4 used for the transformer 4 used for the circuit board (none used for the terminal strip)	.06 ea.
No. 6 Lock Washers Quantity of 8	Local hardware store	n/a 2 used for the transformer 2 used for the circuit board 4 used for the input terminal strip	.06 ea.
Rubber Grommet 3/8"	Local hardware store	n/a	.40 ea.
Small Wire Nuts (5 needed)	Local hardware store	n/a To cap unused transformer leads	$3.30/pkg.

Note that Radio Shack uses both the terms "3.5 mm" and "1/8 inch" for the jacks and plugs that are the size typically found with sound cards.

Once you have your circuit design finalized and built you can mount it all (using the hardware listed above) in a case like the LA-3 ($4.20) from circuitspecialists.com or similar from Radio Shack.


Circuit Construction Notes

As mentioned earlier, the primary and secondary of the transformer have unused leads. These leads however have their ends stripped and tinned. These ends should be cut off and the leads capped with small wire nuts so they don't short against any other component leads or a metal case which could place it at a potential above ground causing a shock hazard.

The unclad perf board specified above that the components are mounted to is larger than is actually needed (but is the smallest available from this vendor). However, it can easily but cut to size with a pair of scissors.

Here's the front of a finished audio interface circuit mounted inside an LA-3 case.

Here's the back of the LA-3 case with the terminal strip mounted externally. Note the wires going through the grommet. One is the black lead of the primary of the transformer. The other is a red wire going to one side of the 4.7 uF audio pass-through capacitor. The two 70-volt audio signal wires from the PA amplifier would attach to the top two terminal screws.

A side view of the LA-3 case shows the transformer and perf board mounted. Note the wire nuts capping the unused leads of the transformer and how the nylon spacers keep the solder side of the perf board safely above the bottom of the case. The white secondary lead of the transformer is soldered to one side of the 470 ohm resistor on the perf board. (The 470 ohm resistor is not visible in the photos.)

In this top view of the case you can see the red wire attached to the top of the large, black audio pass-through capacitor. This is the same wire that goes out of the grommet and connects to the terminal strip.

The single green output wire from the wiper of the potentiometer on the perf board is soldered to both of the channel terminals of the 3.5 mm output jack. This is so the same audio signal is supplied to both channels of the sound card Line In jack. The black wire attaches to the common terminal of the jack.




The Server

The biggest problem with using Windows Server 2003 as a streaming audio server is that Microsoft hasn't provided any drivers for stand-alone sound cards and sound card manufacturers have no interest in developing drivers for 2003 because you typically don't put sound cards in servers. As shown in the Server 2003 HCL, most supported sound devices are audio chipsets only which normally get integrated into motherboards. The problem with that is that most server-class systems don't have audio integrated into the motherboard. However, I've had no trouble getting Soundblaster Live cards to work with 2003. It seems that, while not listed in the HCL, there are drivers in 2003 for these older SB cards.

Windows 2000 Server and Windows 2003 Server both include "Windows Media Services" software to set the server up as a streaming media server. When you set up a Windows 2000 Server as a streaming media server the "Windows Encoder" software is also installed. With Windows 2003 Server (Windows Media Services 9) you have to download and install the Windows Media Encoder 9 separately.

The most important point in setting up the encoder for a live stream is to use the mono setting with a low bit rate. By default the encoder uses a stereo, 135 Kb bit rate which isn't needed for voice reproduction encountered with meetings. Use the 19 Kb or 24 Kb bit-rate settings. Using higher bit rates will dramatically increase the bandwidth utilization of your network links. Note that the above selectable bit rates include 3 to 4 Kb of overhead which means the actual bit rates of the voice signal is 16 Kb and 20 Kb respectively.

In our tests, using the default stereo, 135 Kb bit rate caused a T1 to be saturated with only 12 people pulling the stream using Windows Media Player at a remote site. As a result, all 12 users experienced frequent rebuffering. When using the 19 Kb setting the same 12 listeners only caused a 20% increase in the T1 bandwidth utilization and no rebuffering occurred.

Another benefit of a lower bit rate is that archiving the streams to a file for later playback (see below) takes less storage space because the files are considerably smaller (about 10 megabytes per hour of audio).

Setting up a session to stream live video is pretty easy using Windows Media Encoder 9. When you start the encoder several wizards are available for selection. Select the "Broadcast a live event" wizard.

Select the audio device you'll be connecting the audio interface circuit to.

Select "Push to server".

Enter a server name. This would normally be name of the same system that the encoder is installed on. However, you could set up separate encoder systems (which could be PCs with sound cards) and have them all push to a single server running Windows Media Services. Also enter a "Publishing Point". A publishing point is merely a logical designation for this encoder stream. The name you enter will be appended to the server name in the URL (see the "URL to server" box a little lower down).

Note:  The publishing point that you enter in Windows Media Encoder is set up in Windows Media Services server software. In other words, the publishing point has to exist in Windows Media Services before the encoder session pushing to that publishing point will work. Typically you'll have Windows Media Encoder and Windows Media Services both installed on the same system. However, this may not always be the case. See the Multiple Sources section below.

On the "Bit Rate" window de-select the 135 Kb setting by clicking on it to un-check it. Scroll down and check the 24 Kb checkbox.

At this time you have the option of archiving the live stream to a .wma file which can be played by most audio players. This is useful for what Microsoft calls "On-demand" streaming. i.e. you set up a Web page with links to the .wma files so people can listen to them whenever they want.

One the final screen you can enter the Title and Author (aka Artist) information which will appear in the Windows Media Player display while the stream is playing.

The completed session is now ready to go. You can then save this session to a file by clicking on the disk icon. To start encoding the live audio stream just click on the "Start Encoding" button.

A sound level indicator appears on the left side of the window. Make sure that there is only one indicator bar. If there are two, you're encoding in stereo which takes up unnecessary processor resources and bandwidth. Change it to mono by clicking on the "Properties" button and then on the "Compression" tab.

Since live meetings don't happen all the time, you can set your streaming properties by connecting a radio or CD player with a 3.5 mm headphone jack to the server sound card in place of your audio interface circuit to provide test audio.

Once you're streaming mono live audio from a meeting, adjust the 20 Kohm potentiometer in the interface circuit so that the green indicator doesn't go much past the half-way point otherwise you'll start to over-drive the Line In of your sound card (and the green bar will become yellow and then red as the level increases).

You can see how many users are connected to your server and listening to the stream by clicking on the "Server" tab of the Monitor window.

Now just go to a client system and open Windows Media Player. Click on "File" on the menu bar and select "Open URL". In the URL box type in the name of your server followed by the publishing point name like so:
http://saserver/boardmeeting
and press Enter and the player should connect to the server and start playing the audio being streamed. There'll be about a 15 second delay due to the encoding and streaming processes.

Note:  Theoretically, you could just leave the encoder running all the time so that you wouldn't have to manually start and stop the encoder before and after each meeting. However, if you set the encoder to archive the stream to a file you couldn't do this because you'd end up with a humungeous file containing mostly "dead air" which would eventually fill up the hard-drive.

If users start to experience rebuffering and drop-outs with Windows Media Player, it's an indication that the network link between them and the server is becoming congested due to the number of users pulling the stream. In our tests, about 300K of bandwidth was consumed with 10 users pulling a 24 Kb stream from a server on the other end of a T1. This comes out to 30K of bandwidth per user which would account for any overhead in the streaming process.


Multiple Sources

What if you want to make streams available which are coming from multiple sources such as with multiple meeting or committee rooms? There are two ways to handle this. But first let's be clear that each room would have its own PA amplifier (70-volt output). We are not talking here about a situation where multiple rooms all feed into a single PA amplifier. This would constitute a single source because with a single amplifier there is only one source for an audio signal.

If the 70-volt PA audio signals from all of the sources are available in one location, such as where all amplifiers for all of the sources are in a single room, you could still handle it as a single-source situation by modifying the audio interface circuit to replace the input terminal strip with a rotary switch. The deciding factor is if there would ever be multiple meetings going on at the same time and you wanted to give the users the ability of chosing which meeting they want to listen to. If there are no simultaneous meetings, you would simply select the active source (room) on the rotary switch when you start the encoder before the meeting starts. Handling multiple rooms with their own amplifiers as a single-source situation, you would only have one audio circuit (with a rotary switch), one publishing point, and one encoder session to deal with.

If there is a possibility of simultaneous meetings and your decision is to make them all available for users to listen to, the question of amplifier location is again a concern. If all of the amplifiers are in a single location, you could still only use a single system. You'd just install multiple sound cards in the server, set up a separate encoder sessions for each one of them, and create a corresponding publishing point in Windows Media Services for each one of them.

As for the audio interface circuit, you would need to build a circuit for each of the sound sources (amplifers) also. (The number of amplifiers would be the number of sound cards you'd put in the server.) However, the multiple audio interface circuits could be put in a single, larger case. When putting multiple audio circuits in a single case, be mindful that you do not want to tie their grounds together or tie the grounds to the chassis. Keep them totally (electrically) isolated from each other.

What if the amplifiers aren't in one central location? This will increase your hardware costs because you'll need to put a system at each amplifier location running Windows Media Encoder and have one encoder session on each which pushes to a central Windows Media Services server which has a separate publishing point for each of encoder systems. Note that the encoder systems at the amplifier locations do not have to be servers. They can be Windows XP workstations. You'd have to build a separate audio interface circuit for each location also.

Note that we are actually dealing with two different streams here. There's a stream between the encoder software and the server software, and there's a stream which is sent out to users. When you're running the encoder and server software on the same system the first stream stays internal to the system. Only the stream pulled by the end users running Windows Media Player ends up on the network wire.

This presents an important bandwidth consideration. If your server is located in Building A and your encoder system is in Building B where the meeting is going on, once users in Building B start listening to the meeting you'll have streams traveling in both directions on the wire between the two buidlings. If all you have between the buildings is a 1.5 Mb T1 line, and a lot of users in Building B want to listen to the meeting, the T1 could become saturated. This could disrupt the "source" stream traveling from the encoder system to the server which would affect all users no matter their location.

If you want to make streams from simultaneous meetings available, you'll have to have a way for the end users to pick which stream (meeting) they want to listen to. You'd have to create a central "Meeting Audio" Web page on one of your enterprise Web servers and then create links to the publishing points on the Windows Media Services server for each of the rooms. For example:

<a href="http://saserver/boardmeeting/">Board Room</a> <a href="http://saserver/committeea/">Committee Room A</a> <a href="http://saserver/committeeb/">Committee Room B</a>

Note that you will run into difficulties if you try to use your Windows Media Services server as a Web server also (to host the above Web page). That's because both IIS and Windows Media Services use port 80 by default.



Keith's Home Page



Contents, diagrams, and images    Copyright © 2004-2023    Keith Parkansky    All rights reserved.
Certain graphics, symbols, and terms used on this site and in its documents are registered trademarks
of their respective owners and are contained herein for identification purposes only.
No endorsement of this site, its contents, or its documents by these owners is expressed or implied.

LIABILITY

IN NO EVENT WILL KEITH PARKANSKY BE LIABLE TO ANY PARTY (i) FOR ANY DIRECT, INDIRECT, SPECIAL, PUNITIVE OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, DAMAGES FOR LOSS OF BUSINESS PROFITS, BUSINESS INTERRUPTION, LOSS OF PROGRAMS OR INFORMATION, AND THE LIKE), OR ANY OTHER DAMAGES ARISING IN ANY WAY OUT OF THE AVAILABILITY, USE, RELIANCE ON, OR INABILITY TO USE THE INFORMATION, METHODS, HTML OR COMPUTER CODE, OR "KNOWLEDGE" PROVIDED ON OR THROUGH THIS WEBSITE OR ANY OF ITS' ASSOCIATED DOCUMENTS, DIAGRAMS, IMAGES, REPRODUCTIONS, COMPUTER EXECUTED CODE, OR ELECTRONICALLY STORED OR TRANSMITTED FILES OR GENERATED COMMUNICATIONS OR DATA EVEN IF KEITH PARKANSKY SHALL HAVE BEEN ADVISED OF THE POSSIBILITY OF SUCH DAMAGES, AND REGARDLESS OF THE FORM OF ACTION, WHETHER IN CONTRACT, TORT, OR OTHERWISE; OR (ii) FOR ANY CLAIM ATTRIBUTABLE TO ERRORS, OMISSIONS, OR OTHER INACCURACIES IN, OR DESTRUCTIVE PROPERTIES OF ANY INFORMATION, METHODS, HTML OR COMPUTER CODE, OR "KNOWLEDGE" PROVIDED ON OR THROUGH THIS WEBSITE OR ANY OF ITS' ASSOCIATED DOCUMENTS, DIAGRAMS, IMAGES, REPRODUCTIONS, COMPUTER EXECUTED CODE, OR ELECTRONICALLY STORED, TRANSMITTED, OR GENERATED FILES, COMMUNICATIONS, OR DATA. USE OF THIS SITE CONSTITUTES ACCEPTANCE OF ALL STATED TERMS AND CONDITIONS.