Finally Able to Post on our Recent Webcast Experience

[rwgeorges-p40]

I've been reading the various threads waiting for the day when I could tell everyone about our webcast experience. Below, I've pasted in the contents of a Report I sent our visiting authority about our experience. He was specific about what he wanted it to contain. At the end, I've also added an Excel spreadsheet with info about the equipment we used to make it all happen...

Ramon Georges
Stake Technology Specialist


Report

On Sunday, April 12, 2009, Stake Conference was successfully webcast from the Stake Center to congregations at three other chapels within Stake boundaries. This was our third attempt at webcasting conference and the results were much better than expected. Prior to conference, we set a silent goal to improve technologies, and our production methodology, so they wouldn’t hinder the influence of the Spirit. The feedback after this conference was more positive, with several members commenting on how the quality of the webcast was similar to a satellite broadcast. This did not happen without inspiration from the Spirit and members willing to consecrate their time and talents. This report highlights changes we made to our webcast process (Capturing, Encoding, Delivering, and Presenting) so it can be repeated and yield the same positive results.

Capturing - Two microphones were used to capture audio from the choir. They were placed on boom stands, which were to the side of the choir, but extended the microphone over and in front of the choir. A third microphone was used to capture the organ sound reflecting off the wall behind the choir. The podium microphone was cleanly split into two feeds. One feed connected to the building’s sound system and the other connected to a channel on an audio mixer. All four microphones were fed into the audio mixer on separate channels, which allowed us to tune the audio before it was output to the Webcast Communicator. Video was captured using a Canon 3-CCD XL-1S camcorder. S-video baluns were connected via Cat 5 cable to extend the camcorder’s S-video output approximately 150 feet to the Webcast Communicator located in the technology closet.

Encoding & Delivering - We purchased a Webcast Communicator from the Church, and after much testing, used the 550Kbps setting on the Communicator. We modified the Communicator’s encoder profile for 550Kbps to output video at a resolution of 720 x 480 pixels and use Mono sound for audio encoding. Qwest DSL (1536Kbps downstream/846Kbps upstream), was activated at each Chapel and provided the mechanism for receiving the webcast stream. We sent the webcast stream from the Stake Center to Streaming Servers in Salt Lake, where they were sent to each Chapel. The Church’s new Communicator and streaming service worked very well.

Presenting - We purchased a Webcast Receiver from the Church for each Chapel. Video from the Receiver was sent to a newly purchased projector via a scan converter (borrowed from our local FM Group). We also purchased 80” collapsible projector screens for each chapel. Audio from the Receiver fed into a crab unit, which was connected to the building’s sound system. A video feed from the scan converter, and a second audio feed from the crab unit, were fed into an audio/video combiner unit (from local FM Group). The combiner unit output an RF feed to televisions in Relief Society rooms. The setup at each chapel required minimal monitoring.

[russellhltn]

Thanks for the report. That's quite a setup. I think we normally refer to the combiner as a RF Modulator.

While I'm sure the setup worked well for you, it does seem to suggest advanced knowledge of sound systems. I'm not sure how well it would work in the hands of an average STS.

Out of curiosity, what was the out-of-the-box encoder profile settings?

[rwgeorges-p40]

Sorry for another long post, but I’m a little passionate about this subject. I am glad there’s a forum for this because I know there are others that are struggling with this topic just like we were (and still are in some respects). It really isn’t easy to do this, especially if you want good results.

I admit, the setup has involved some expertise, but it’s taken a while to evolve to where it is today. The former STS, Mark Duff (now on the High Council and overseeing technology), kick started the webcast idea. About two years ago, when Elder Bednar came for Stake Conference, Mark was asked to come up with a method to “broadcast” conference from the Stake Center to all three buildings because there was only time for one session. I was one of the people that volunteered to help with the equipment and setup. After looking at a few options, Mark settled on webcasting conference over the Internet.

The original setup used only the podium microphone at the Stake Ctr. Video from the camcorder and audio from the building’s sound system were pumped into a laptop, which was connected to the Internet via DSL. Windows Media Encoder was used and SLC agreed to provide us a streaming reflector for the conference. At each chapel, there was a laptop, crab, projector, scan converter, and RF modulator. Except for the scan converters and modulators, which the FM Group agreed to purchase, everything was borrowed and nothing was the same. To make it work, several members were involved and it took a lot of testing. While the speaker ended up sounding pretty good, the choir and organ were awful. They sounded very robotic and indiscernible.

We tried again last year, but we didn’t have a reflector. We had to cut the encoding rates to accommodate sending three simultaneous streams, one to each building. In addition, we borrowed everything again and performed extensive testing. It obviously didn’t work as well. The video was ok, the speaker sounded ok, but the organ and choir sounded bad again. This time, we noticed that several members at each of the three buildings had a hard time paying attention. People didn’t mind pixilated video, but they found it very hard to concentrate when the audio was inadequate. It affected the Spirit people felt. That was a key revelation.

When members go to the Stake Center to watch a broadcast, they expect a certain level of quality and the Church goes to great lengths to ensure a consistent and pleasing experience during a broadcast. It dawned on us that members weren’t going to know the difference between a webcast and a broadcast. Most don’t care about the delivery mechanism; they just know what looks and sounds good. But when it’s not the same experience they’re used to, it’s very noticeable. We felt that it would be worse for investigators, who may not have any expectations.

In preparation for this last conference, we decided to overcome the audio problems. At the same time, we wanted to purchase equipment so there was a guarantee that future conferences produced similar results. Also, the setup at each of the three Chapels needed to be easy and shouldn’t require a techno geek to set up, monitor, or tear down. We also didn’t want to spend weeks testing everything. Reading the other threads, that seems to be the same philosophy that led to the creation of the Communicator and Receiver. We knew the next STS may not be very experienced and we didn’t want him to freak out at conference time.

Prior to any purchases this year, we met with the STS of Val Vista Stake. They had just used the Communicator/Receiver combo for a funeral and were figuring out how to improve the experience. Gary Smith, Director of Engineering at Bonneville Communications in Phoenix, brought a bunch of equipment for everyone to review. Mark and I already knew how to best solve the audio problems, but with Bro Smith’s equipment in hand, we were able to go over different options before we purchased anything.

With much of the equipment purchased, we scheduled a choir and organist to test the setup. It took two sessions for us to nail down the audio settings. Most of it involved trying to tone down the organ so it sounded good when encoded (switching to mono on the encoding profile eventually overcame this issue), while keeping it from overpowering the choir and not sounding hot on the receiving end. One thing to note, you need to mute the channel for the podium microphone when the choir is singing or you’ll get a weird sound effect. We did the opposite when the speaker was at the podium otherwise the microphones picked up any conversations by choir members. Lastly, bypass the sound system. This gives you the cleanest sound. You can use a microphone splitter, which I included in the list of equipment. I highly recommend investing in your audio setup. The microphones we purchased are solid performers, provide great sound and aren’t expensive. Microphone stands and booms are pretty inexpensive as well.

As for the encoder settings, I believe that 350Kbps and lower have resolution of 320 x 240, using 48-bit stereo audio. The higher encoding rates have the same audio settings, but the resolution was increased to 640 x 480. I’ll have to check the Communicator when I’m at the Stake next week. I can post actual info then. Also, the processor on the Communicator never moved about 35%, unless you moved the camera, then it usually jumped up to 60%. RAM usage kept below 809MB (the Communicator has 2GB of RAM). Slow pans or zooms caused the Communicator to hiccup on the encoding. We don’t recommend panning the camera. Figure out the framing of the speaker and, if your camera supports it, preset the shot. Do the same for a wide shot that shows your choir. Don’t try to pan and zoom in on your chorister. He or she will be captured in the wide shot. The fewer the camera moves the better. Quick zooms never caused the Communicator to hiccup.

My first post explains the results. It was awesome. People totally forgot about the technology and focused on the message. We’d achieved our primary objective. There were a lot of very positive comments from people at each of the buildings. Some thought it was an actual satellite broadcast. Members at the Stake Center weren’t bothered by the choir microphones. Neither was the choir. They were more interested in sounding good. People were distracted by the microphones and that was good. For a few, seeing the equipment on the stand made them feel a little like they were up at General Conference.

One building had a non-techie taking care of things and there were no difficulties. The other two buildings had volunteers from previous conferences who wanted to see what the difference would be more than anything. They don’t plan on helping next time because the setup is pretty simple now.

Even though it went well, we’re not finished. There are a few things we want to refine, and purchase (borrowed the mixer and camera), before the next conference coming up in September:

1. We plan to purchase our own Mackie Onyx 1220 mixer (w/o the firewire card), because the signal output is incredibly clean. It sounded excellent and when you’re encoding audio, the cleaner the source, the better. You can find it on Amazon for around $600. That’s less than one Receiver. Audio made a world of difference this time around.

2. We’re interested in the idea of a PTZ camera, so that video can be controlled from the technology closet, not just audio. Remember that the video needs to be clean as well. No decisions yet, but we’re looking.

3. We plan on purchasing our own scan converters & RF modulators. What we use belongs to the local FM Group and other stakes borrow it for their conferences. There have been scheduling conflicts. We won’t purchase the same gear. It’s commercial grade, which we really don’t need. No decisions yet.

4. We didn’t have the audio cables permanently installed on the stand prior to conference. We wanted the liberty of finding the right placement and used long XLR cables. Now that we know where the microphones will go, we’re going to have the cable run under the stand and install in-floor audio jacks. We’ll run them back to the technology closet, which is North of the Chapel, and terminate on jacks in the wall. We’ll purchase short XLR cables to run from the jacks to either microphones or the audio mixer. That will eliminate a lot of set up time, cables all over the floor, and a lot of gaffer’s tape, which isn’t cheap.

5. We’re thinking about tying a translation receiver to the building sound system. We don’t have enough receivers for all the members that need them. We’ll be able to set up a TV in one of the rooms and use the Spanish modulator (channel 4) to feed an RF signal to the TV with the audio translation.

6. We’re getting some cases, with rollers, for the equipment that is setup at each building. Everything, except the screen, will go in the case. All the cables are labeled and I’m almost finished with the diagram. When I get the replacement converter/modulator, I’ll finish the instructions. We plan on having non-technical people take care of the set up, monitoring, and tear down at the buildings during conference in September. That will be the ultimate test.

Mark took photos of the final setup. When I get them, I’ll post them to the thread. I’ll also add the diagrams when they’re done. As we select other equipment, I'll add that to this thread as well. If you happen to be in Mesa during the weekend of Sept 19th, and want to see the arrangement in action, please let me know. My email is rwgeorges@me.com.

[rmrichesjr]

Here's an alternative idea for choir mikes that is less expensive and less visually distracting than large commercial booms and mikes. While I haven't used it for a webcast, I have used it for reinforcing a (quiet) stake youth choir into the cultural hall and overflow rooms within the building.

The microphone unit is a very inexpensive electret condenser mike unit sold at the neighborhood electronics retailer. I think I used RG-174 as the cable, because it is rather thin. The stake center where I used them had a motorized projection screen, so I hung the mikes on brown-colored string hung from PVC water pipe caps slipped over the ends of the tubing at the bottom of the screen. Of course, I was careful when raising the screen, which pulled the mikes up into position. If you weren't on the front few rows in the chapel, you had to know what you were looking for to see them.

The interface between the mike units and PA system is described in this image I posted to an earlier thread: http://tech.lds.org/forum/attachment.ph ... 1178582241 I'd be happy to try to clarify anything not readable in the schematic. (Sorry for it being such a poor hand-drawn image.) The transformer is the 600 ohm 1:1 audio transformer available at the same neighborhood electronics retailer. One interface unit can handle multiple mikes.

The condenser mikes are a little bright due to their resonance around 18-20kHz. If desired, that could be equalized out. The mikes are definitely sensitive enough for a choir. In fact, they will pick up paper rustling if not potted down between choir numbers.

[rwgeorges-p40]

There were a few mistakes on the spreadsheet, so I've uploaded a corrected attachment. I've also completed diagrams of the stake center and chapels. If you want a Visio version of the diagrams, let me know.

Feedback is always appreciated.

Ramon

[Mikerowaved]

Feedback is always appreciated.

Except in the middle of a broadcast, of course.

Seriously, nicely done documentation. Thanks for sharing it with the group.

[rwgeorges-p40]

There is one variation on the setup at each Chapel. Instead of running the RCA cable from the EJ-10 Crab to the AV Modulator, we've run the cable from the jack under the sacrament table to the AV Modulator. The jack under the sacrament table comes from the building's sound system. At a couple of buildings, this actually sounded better. We didn't do it for the last conference setup, but it can be done. For those that aren't aware, there's usually a jack under the Clerk or Sacrament table where a tape recorder can be plugged in to record a talk or something.

btw - The circles on the drawings are tables. At two buildings, they don't exactly fit, but they work. In the future, we plan on using either a small, portable table we purchase or make the tables ourselves and just leave them at each building.

Ramon

[DeeGardiner]

I thought I would piggyback on this thread to share our most recent web casting experience. We just had stake conference this last weekend (May 30-31). Last December we borrowed a prototype webcast kit from church headquarters, while this time we used our new purchased kit.

We did a test run last Wednesday evening that went very well. We then did a full dry run during the Saturday evening session which required some equipment adjustments. And then the real thing during the Sunday general session.

We have two cameras (one owned by the stake, one owned by the stake A/V expert) connected to a borrowed mixer. They are both on tripods in the chapel overflow. It would be nice to have taller tripods so it didn't catch people's heads as they enter/exit the chapel - throwing the camera out of focus.

Neither camera produces very good color in our chapel, but they are tolerable. One camera was set to a wide angle for the choir, and the other for close-ups of the speaker. Panning and zooming cause encoding problems, so this avoided that problem. Some day we would like better cameras.

I suspect the mixer can drive the long cable to the A/V equipment room better than a camera. The encoder is in the A/V room.

I purchased a Radio Shack video distribution amp so that I could drive the webcast encoder and the RF modulator with nice clean signals. This seemed to be over driving the signals as everything was washed out and saturated. Both the encoded image and the analog image distributed throughout the stake center were terrible. I ended up removing the distribution amp and replaced it with a piggyback RCA cable. This doesn't provide proper termination of the line, but the audio and video seemed clean.

We did not hook up a crab box to a phone line for audio backup. We tried last time but never got it working. We assumed we would be fine without it.

I also noticed that the encoder gets very hot. I juggled things around to give it more breathing room in the cabinet (it is permanently installed now).

Last time we had problems with the audio getting out of sync with the video at about 1 hr 15 minutes in to the broadcast (on three of three tests). So we dropped the bit rate to medium (200 kbps I think). The video quality was fair, but we still had the sync problem.

This time I used the custom settings at 600 kbps. We have Comcast at the stake center with 1.5 Mbps upload and DSL at the destination with 1.5 Mbps download. So I figured 600 kbps should be fine. Image quality was much better.

At the destination chapel we had the church provided receiver. This worked well and was much easier to set up than a laptop. Last time we used a laptop and it took a long time to get everything configured correctly. The receiver was easy to use. It does require a keyboard, but we did not need a mouse. We could navigate via the tab key. The receiver drove the stake's projector and the crab box for audio.

We just used the podium microphone for our audio source and it seemed to work well for speakers and the choir. The only issue is prelude and postlude music require the microphone to be on, which also passes general irreverence noise.

Audio and video quality were very good - not distracting at all. The weak link in our system is in the camera and projector. They are not really good enough for large chapel settings.

We did have some technical issues though. Just prior to our intermediate hymn one hour in to the conference the receiver lost signal. It reaquired after a minute or so. It lost it again a few minutes later, and again reaquired. After the hymn it lost it again.

I was watching the stream in the stake office during that third outage. That stream was out of sync from the destination building, but it also locked up. The encoder was still running. On the stake computer I got an error message indicating a network error.

We didn't know the cause, so we stopped and restarted the encoder and the receivers. Everything worked fine after than.

All in all I feel it was a success. We lost a total of about 5 minutes during the middle, but otherwise the quality was good and not distracting. The congregation was very tolerant of the outage.

If anyone has clues on what caused our outage, I would appreciate hearing about it.

[rmrichesjr]

...

I suspect the mixer can drive the long cable to the A/V equipment room better than a camera. The encoder is in the A/V room.

I purchased a Radio Shack video distribution amp so that I could drive the webcast encoder and the RF modulator with nice clean signals. This seemed to be over driving the signals as everything was washed out and saturated. Both the encoded image and the analog image distributed throughout the stake center were terrible. I ended up removing the distribution amp and replaced it with a piggyback RCA cable. This doesn't provide proper termination of the line, but the audio and video seemed clean.

...

Glad to hear things went pretty well. Sorry, but I can't help with the dropouts and such.

I wouldn't worry about the camera driving a long cable. Even with some fairly inexpensive consumer equipment, baseband/composite video signals are driven as into a 75-ohm transmission line, no matter what length the cable is. At baseband frequencies of a few MHz, attenuation over even several hundred feet of cable is going to be negligible. However, this does presuppose proper termination at the receiving end of the cable.

Sometimes, equipment from the electronics retailer you mentioned is good. Other times, well ... If you have access to an oscilloscope, you might look at the waveforms coming out of it, with proper termination, of course. If I remember correctly from KBYU-CC days, the sync pulse is 0.4V below zero and whitest white is 1.0V above zero. Much consumer equipment is AC-coupled, which loses the zero reference, but you can compensate for that when measuring. One possibility with the video DA is somebody might have got cute and used a single smaller-value series resistor on the outputs rather than one resistor per output connector. For that case, the easiest solution is to terminate all outputs, even the ones you aren't using.

[DeeGardiner]

If you have access to an oscilloscope, you might look at the waveforms coming out of it, with proper termination, of course. If I remember correctly from KBYU-CC days, the sync pulse is 0.4V below zero and whitest white is 1.0V above zero. Much consumer equipment is AC-coupled, which loses the zero reference, but you can compensate for that when measuring. One possibility with the video DA is somebody might have got cute and used a single smaller-value series resistor on the outputs rather than one resistor per output connector. For that case, the easiest solution is to terminate all outputs, even the ones you aren't using.

My personal oscilloscope is really old and not calibrated, but I do have access at work. I will run some experiments and see what I learn.

I didn't think of terminating the unused outputs. I assumed they were totally isolated, but perhaps they are not.

Thanks for the suggestions.

But on the other hand, things seem to work quite well with my piggyback RCA patch cord. I realize it messes up the termination, but the encoded video and the video distributed throughout the stake center look pretty good. The patch cord places about a 3' stub with a second 75 ohm termination on the end of the line. If that is causing problems, what exactly should I look for in the image? I assume it would cause some kind of ghosting or something. If anyone has specific thoughts on what to look for, I would appreciate it.