Barix, EVDO remote Broadcasts

[ChiefOperator]

Reading the posts about Barix, I'm thinking these boxes would work well for those who wish to do remotes within EVDO A service areas.

I'm wondering whether the below set-up, comprising 4 boxes, would work. If so, it appears to be a professional solution that would cost less that $1,500.

Studio side: Place an instreamer and exstreamer at the studio. The instreamer would send audio (mix-minus program, cue) to the remote site. The exstreamer would recieve program from the remote. These boxes would network into the station router.

Remote side: The instreamer would send remote program audio to the studio and the exstreamer would receive studio audio. As for transmission, the two remote boxes could plug into a portable router, such as the new ones that accept an EVDO card (and have WiFi, I think). This set-up would be completely portable within the EVDO service area (or where WiFi exists, provided the proper router is used).

Would this not work--and work similiar to the Comrex Access that accepts an EVDO card?

Finally, the new Barix boxes (1000) are full-duplex, I believe, so my arrangement could eventually be reduced to two boxes.

Thoughts? Thanks!

 

[RolfTaylor]

To make this work well for live remotes you would probably need a low delay algorithm.

Telos and Comrex have algorithms with delays as low as about 90 msec. apt-X would drop this to nearly zero (under 2 msec).

Also note that these Barix units do not have the dynamic buffering, error concealment, and realtime scalable codecs, so you comparison is apple to pineapples.

LaGuy, what is the lowest delay algorithm on this boxes?

The above is algorithmic delay. Packetization and buffering add considerably to the total delay.

Rolf Taylor

Applications Support Engineer

APT North America

 

[BofH]

I just ran a test last night with the Instreamer at the studio and the Extreamer (Streaming Client Firmware) plugged into a laptop via it's Ethernet port (using ICS) via an HSDPA 3G card.

I found that using MPEG1/44.1kHz (MP3) at the highest encoding quality 7 (192k bitrate) worked with no glitches with the buffer set to 500ms in the Exstreamer. At the 100ms buffer setting there was the odd glitch here and there.

At this high bitrate a couple of dropped packets etc are barely inaudible - the Barix guys have done a great job on their RTP error concealment.

There was well under a second delay, probably 600ms which would be about right, the ping times over the Wireless connection were 80-150ms

Reducing the quality setting (and thus bitrate) will probably increase the latency slightly and may increase the audibility of any packet loss.

Am going to use this setup at a remote this weekend to pickup the Mix Minus from the studio to feed our PA system. Will be using our existing solution to send the talent audio to the studio from the remote for now - until I can buy another Instreamer/Exstreamer.

However I do expect sending audio from the remote back to the studio will be a stumbling block, as upload bandwidth is much more limited on these 3G wireless connections. So you could probably only safely use 64-96kbit upload bandwidth and I'm not sure how much that would blow out the delay ...

 

[Chuck]

However I do expect sending audio from the remote back to the studio will be a stumbling block, as upload bandwidth is much more limited on these 3G wireless connections. So you could probably only safely use 64-96kbit upload bandwidth and I'm not sure how much that would blow out the delay ...

I've noticed that in RTP mode, the Barix boxes use about twice the bandwidth as the sampling rate might indicate. It seems there is a lot of hand shaking going on. In other words, 64K seems to need a minimum of 128K to actually work without buffering. That might be a problem for your wireless connection. It will depend on how robust it actually is. Not all wireless connections are created equal. I understand the new version using AAC+ will use less bandwidth for comparable quality. These are amazing little boxes, but they are a work in progress.

 

[ChiefOperator]

Rolf, I've seen the APT boxes and I agree that they are awesome.

Chuck, I wonder whether they need 128K because each channel is 64k, yes, no?

 

[RolfTaylor]

However I do expect sending audio from the remote back to the studio will be a stumbling block, as upload bandwidth is much more limited on these 3G wireless connections. So you could probably only safely use 64-96kbit upload bandwidth and I'm not sure how much that would blow out the delay ...

I've noticed that in RTP mode, the Barix boxes use about twice the bandwidth as the sampling rate might indicate. It seems there is a lot of hand shaking going on. In other words, 64K seems to need a minimum of 128K to actually work without buffering. That might be a problem for your wireless connection. It will depend on how robust it actually is. Not all wireless connections are created equal. I understand the new version using AAC+ will use less bandwidth for comparable quality. These are amazing little boxes, but they are a work in progress.

Chuck,

"64 K" almost certainly refers to the audio bit rate. Unlike ISDN or T1 ,the actual bit rate on a T1 is higher, due to the fact that headers need to be added to each packet. The smaller the packet size, the higher the actual bandwidth is since the header information is the same regardless of the size of the packet. On the other hand the larger the packet size the more delay (since the packetizer must wait until all data for a packet arrives before it can be completed and sent.

For a chart the shows this relationship (plus lots of other educational information) see page 5 of our Guide to IP Audio here:

http://www.aptx.com/Admin/Editor/Assets/PDF/IP Audio Networking.pdf

I think you will find this worth a read. We have kept it light on the sales stuff and focused on useful real-world info.

Rolf Taylor
Applications/Support Engineer

APT North America

 

[Chuck]

"64 K" almost certainly refers to the audio bit rate.

It does, and as you note in your chart, the actual bandwidth used is quite a bit more than just what is used up by the audio. Although EVDO can provide some very good through-put it can also choke down to speeds that are little better than a good dial up connection. It just depends on where you are at a given moment. My experience doing this is it works very well.....sometimes. Obviously the less bandwidth you eat up, the more robust the connection.

Thanks for the link.

 

[LA_Guy]

Barix units work great for remotes-especially with the stock firmware. The trick is to set up the studio Exstreamer with a static IP address (and the firewall ports opened), then use raw UDP to send the Instreamer signal to the studio. The Instreamer can be set up to work with a dynamic IP address, so all the 'street team' has to do is plug it into any vanilla DHCP Internet connection. Even better, the Exstreamer can be set up to listen to two UDP ports at the same time and switch its audio output to one (the priority port) if audio appears there. This means that you can have a weekly club remote set up to always stream to the station-unless the street team's Instreamer (set to stream on the priority port) comes online. Then their audio automatically gets switched on, until they turn their Instreamer off. Even better, Raw UDP has very low latency.

One customer has their STL set up this way (he's a small station). He does lots of sporting events, and has his second Instreamer set to stream to the transmitter Exstreamer's priority port. When he gets to a sporting event, he sets up his gear and at airtime, turns on his Instreamer. At his transmitter, the Exstreamer sees the audio appear on the priority port and breaks the audio from his studio, replacing it with the game audio. At the end of the game, he shuts down his Instreamer and he's back on the air from his studio. This can also be used where studios are located in storm prone areas. It is a truly powerful Barix feature!

 

[ChiefOperator]

Thanks LA_Guy... Are these boxes successful with EVDO Rev A?? It seems it would work well.

 

[LA_Guy]

Yes.

Thanks LA_Guy... Are these boxes successful with EVDO Rev A?? It seems it would work well.

We have many working well with wireless. Obviously, the larger the buffer the better...