Has anyone had experience doing long form programming on the Comrex Access? ISDN is no longer available at my location. And I need more than a POTS codec. If you have experience doing long form (talk w/phone calls) programming on the Access, how was the delay time? I've heard that the longer you are connected to an IP address the longer the delay becomes between the remote location and the studio, thus frequent retraining of the codec must be done. Thanks.
-
Get involved.
We want your input!
Apply for Membership and join the conversations about everything related to broadcasting.
After we receive your registration, a moderator will review it. After your registration is approved, you will be permitted to post.
If you use a disposable or false email address, your registration will be rejected.
After your membership is approved, please take a minute to tell us a little bit about yourself.
https://www.radiodiscussions.com/forums/introduce-yourself.1088/
Thanks in advance and have fun!
RadioDiscussions Administrators
Comrex Access
- Thread starter shure
- Start date
C
callfm
Guest
Comrex offers a 10-day trial of their Access units. We are getting ready to try them out in the next few days, as soon as we receive the trial pair.
ISDN, for us, is getting too unreliable for the STL. Last time "it" went down (for 36 hours) AT&T told us they weren't under any obligation to fix it in less than 24 hours. They also indicated there are 4 "cards" along each side of our ISDN STL (so at least 8 places to fail).
After being up all night since it first failed at about 2AM, I took a laptop with a G3 wireless card to the transmitter site at 5:30AM and with all of the co-located wireless carriers, found that even with the laptop placed in an equipment rack between a 20kW and 30kW transmitter, the available wireless bandwidth was greater than 1300-1500 kbps. I picked up our Internet stream on the laptop and had a solid connection easily accepting a Windows Media stream as high as 128k. After more investigation on the technology I was told that Kyocera makes a router that takes a G3 wireless card and that some broadcasters are using such a configuration with the Comrex Access in some cases.
We're going to use a T-1 Internet connection at the transmitter site (no DSL available) and dump the ISDN. Given that a lot of tower sites have co-located wireless carriers you might want to consider checking out the wireless Internet functionality if you are out of options (no DSL available).
ISDN, for us, is getting too unreliable for the STL. Last time "it" went down (for 36 hours) AT&T told us they weren't under any obligation to fix it in less than 24 hours. They also indicated there are 4 "cards" along each side of our ISDN STL (so at least 8 places to fail).
After being up all night since it first failed at about 2AM, I took a laptop with a G3 wireless card to the transmitter site at 5:30AM and with all of the co-located wireless carriers, found that even with the laptop placed in an equipment rack between a 20kW and 30kW transmitter, the available wireless bandwidth was greater than 1300-1500 kbps. I picked up our Internet stream on the laptop and had a solid connection easily accepting a Windows Media stream as high as 128k. After more investigation on the technology I was told that Kyocera makes a router that takes a G3 wireless card and that some broadcasters are using such a configuration with the Comrex Access in some cases.
We're going to use a T-1 Internet connection at the transmitter site (no DSL available) and dump the ISDN. Given that a lot of tower sites have co-located wireless carriers you might want to consider checking out the wireless Internet functionality if you are out of options (no DSL available).
M
menotti1
Guest
call fm, you're going to use the t-1 internet for STL? i have a client whose old analog stl in on it's last leg and he's wanting to update.has great line of sight to a tower doing broadband wireless internet.i think it's the motorola canopy system.claims to have t-1 capability using 11mb ethernet wireless ethernet radios.that should be plenty of bandwidth .dedicated connection so internet traffic should not have a bearing.without having to do something like the comrex, what would be the best way to deliver the signal to the ethernet radio port on each end??ISDN is just
not cost feasible anymore,so he does not want to go there and really does not want to go back to the old analog 950 stl...any input from engs who are on top of this would be appreciated...wants to do uncompressed linear not mp2 or mp3.
not cost feasible anymore,so he does not want to go there and really does not want to go back to the old analog 950 stl...any input from engs who are on top of this would be appreciated...wants to do uncompressed linear not mp2 or mp3.
R
RolfTaylor
Guest
Both the Access and the Telos Zephyr/IP have dynamic buffers so the delay will vary, in some cases considerably, during a connection. This is very suitable for remotes, but can be problematic
If you need the shortest delay possible, and want a delay that is fixed you should look at the APT codecs. The codec itself is very low delay, packet size is fully adjustable (to allow you to tweak bandwidth versus delay) and the fact apt-x is sample based rather than frame based means that dropouts are nearly inaudible, particularly with small packet sizes.
We have a nice utility that can be used to qualify UDP connections - if you ask really nicely I'll email you a copy.
We too have free demos, if you are interested, ask your dealer or email me and I'll have someone contact you. We have 4 different offerings, so we cover a range of applications.
Oh, and one more thing, with apt-x you also get a much less lossy compression than psychoacoustic algorithms (about 6% of info lost at 4:1 compression versus 25% for psychacoustic algorithms such as MPEG.
Rolf Taylor
Applications/Support Engineer
APT North America
www.aptx.com
If you need the shortest delay possible, and want a delay that is fixed you should look at the APT codecs. The codec itself is very low delay, packet size is fully adjustable (to allow you to tweak bandwidth versus delay) and the fact apt-x is sample based rather than frame based means that dropouts are nearly inaudible, particularly with small packet sizes.
We have a nice utility that can be used to qualify UDP connections - if you ask really nicely I'll email you a copy.
We too have free demos, if you are interested, ask your dealer or email me and I'll have someone contact you. We have 4 different offerings, so we cover a range of applications.
Oh, and one more thing, with apt-x you also get a much less lossy compression than psychoacoustic algorithms (about 6% of info lost at 4:1 compression versus 25% for psychacoustic algorithms such as MPEG.
Rolf Taylor
Applications/Support Engineer
APT North America
www.aptx.com
Rolf,
You should have your IT person set the aptx.com domain name to resolve "http://aptx.com" or just "aptx.com" to the website as well. I see a lot of great websites lose out on traffic because the site does not resolve without the www. A lot of people just type the domain name without the www part. Otherwise a great website.
You should have your IT person set the aptx.com domain name to resolve "http://aptx.com" or just "aptx.com" to the website as well. I see a lot of great websites lose out on traffic because the site does not resolve without the www. A lot of people just type the domain name without the www part. Otherwise a great website.
I'll put in a word for the Access...exellent tech support as well!
Wow, I only just found this board by googling our company name. Thanks for the plug, FatPunk and Dudefan. I think it's only fair to add my angle to my good friend Rolf on IP codecs. The APT algorithm indeed has a clear advantage over others in terms of coding delay. But that's one of the only advantages it does have, and that diminishes somewhat in an Internet environment, especially over challenging networks where the only way to provide stability is to add decode jitter buffering. It has one big disadvantage in that it's a relatively simple scheme that uses more bandwidth than more modern choices. Overall success rate of IP coding on the Internet involves a blend of low delay, reduced data rate and intelligence in the part that "catches" the packets at the decoder and aligns them.
And to answer the original question, the only environment where the delay would be constantly increasing over usage time would be one where the network was degrading over that time. On wired IP on each end like DSL or Cable, the ACCESS experience really rivals ISDN. The wireless networks like 3G are very useful for drop-in remotes or news, but I'd certainly do a site survey before relying on them for a four hour game.
Best to All
Tom Hartnett
Tech Director
Comrex
And to answer the original question, the only environment where the delay would be constantly increasing over usage time would be one where the network was degrading over that time. On wired IP on each end like DSL or Cable, the ACCESS experience really rivals ISDN. The wireless networks like 3G are very useful for drop-in remotes or news, but I'd certainly do a site survey before relying on them for a four hour game.
Best to All
Tom Hartnett
Tech Director
Comrex
R
RolfTaylor
Guest
Tom,
good to see you here. Comrex deserves a lot of credit for introducing adaptive technology. And there is no doubt that it is solid in theory and practice, but by its nature it is "variable" in terms of delay.
Here at APT we were recommending that our IP products be used on links with assured Quality of Service. However customers did not listen, and we began to get surprisingly favorable reports from them! That made us go back and examine why it works as well as it does.
1) The very low delay of the algorithm gives you some "free" delay at the buffer. For example I think Access has a number of codecs in the ~75 msec range. Since apt-x has a coding delay of 5 msec this means we get 70 msec of buffering for free.
2) Most codecs use audio frames. Each frame consists of multiple samples. With such an algorithm, there is no real advantage (in terms of quality) to use a packet size smaller than the frame size, since the loss of a packet smaller than the frame size would cause the entire frame to be invalid in any case (note that smaller packets sizes would still reduce delay). With one of these algorithms, the minimum dropout would be one frame, which amounts to to 20 msec or more. We have all heard the loss of an MPEG frame and know that such a dropout is disruptive and easily noticed.
In contrast, apt-x does not use frames, only audio samples. This permits smaller packet sizes to give improved error resiliency as well as reduced delay.
For example with packet size of 512 Bytes a dropout would be only 2 msecs. These are audible, but hard to detect. I generally describe them as "micro dropouts"). At a packet size of 64 Bytes the dropout is only 249 micro-seconds.
Now to be fair, small packet sizes means more bandwidth. As does the relatively small 4:1 compression ratio of enhanced apt-x. For this reason, while the jury is still out (testing is ongoing), it is probable adaptive technology (a la Comrex) will work better for very low bit rates on highly contended links such as wifi.
For use on xDSL and Wireless broadband, where bandwidth is less limited, apt-x is proving itself. Even without QoS it can be configured such that any problems are minimal while simultaneously taking advantage of the far less destructive nature of the adpcm type of coding.
I would like to invite Tom, and the group, to attend the papers our staff will be giving at NAB.
cheers
Rolf Taylor
APT North America
good to see you here. Comrex deserves a lot of credit for introducing adaptive technology. And there is no doubt that it is solid in theory and practice, but by its nature it is "variable" in terms of delay.
Here at APT we were recommending that our IP products be used on links with assured Quality of Service. However customers did not listen, and we began to get surprisingly favorable reports from them! That made us go back and examine why it works as well as it does.
1) The very low delay of the algorithm gives you some "free" delay at the buffer. For example I think Access has a number of codecs in the ~75 msec range. Since apt-x has a coding delay of 5 msec this means we get 70 msec of buffering for free.
2) Most codecs use audio frames. Each frame consists of multiple samples. With such an algorithm, there is no real advantage (in terms of quality) to use a packet size smaller than the frame size, since the loss of a packet smaller than the frame size would cause the entire frame to be invalid in any case (note that smaller packets sizes would still reduce delay). With one of these algorithms, the minimum dropout would be one frame, which amounts to to 20 msec or more. We have all heard the loss of an MPEG frame and know that such a dropout is disruptive and easily noticed.
In contrast, apt-x does not use frames, only audio samples. This permits smaller packet sizes to give improved error resiliency as well as reduced delay.
For example with packet size of 512 Bytes a dropout would be only 2 msecs. These are audible, but hard to detect. I generally describe them as "micro dropouts"). At a packet size of 64 Bytes the dropout is only 249 micro-seconds.
Now to be fair, small packet sizes means more bandwidth. As does the relatively small 4:1 compression ratio of enhanced apt-x. For this reason, while the jury is still out (testing is ongoing), it is probable adaptive technology (a la Comrex) will work better for very low bit rates on highly contended links such as wifi.
For use on xDSL and Wireless broadband, where bandwidth is less limited, apt-x is proving itself. Even without QoS it can be configured such that any problems are minimal while simultaneously taking advantage of the far less destructive nature of the adpcm type of coding.
I would like to invite Tom, and the group, to attend the papers our staff will be giving at NAB.
cheers
Rolf Taylor
APT North America
comrex said:
Yeah, I'm terrible with remembering names, but I beleive the cat that's has been so helpful was named Andy. Another cool thing about Access, and myabe other boxes are this way these days, the "brains" are on a simple flash media card, which I thought was pretty cool. We have two units, one of them was acting a lil flaky, and a new flash card with the latest and greatest firmware was sent to me lickety-split!
W
WTIC
Guest
WTIC has employed the COMREX ACCESS units in the 2008/2008 season to backup its UCONN Radio Network programming for all home and away games(mostly in pots mode).
We used it for the first time in internet mode as a PRIMARY BACKHAUL from the Paradise Jam in USVI this past november because ISDN is not available on the island. after crossing fingers, we got to the gamesite, hit connect and it performed flawlessly for the rest of the 3 game tournement.
since then it has been put to the test at a variety of other locations with our news department in WiFi mode and ethernet mode and has done nothing but exceed expectations.
if anyone has serious questions about this product and considering purchase i would be happy to discuss our experiences with it at WTIC
my desk 860 284-9288
Jeff Hugabonne
We used it for the first time in internet mode as a PRIMARY BACKHAUL from the Paradise Jam in USVI this past november because ISDN is not available on the island. after crossing fingers, we got to the gamesite, hit connect and it performed flawlessly for the rest of the 3 game tournement.
since then it has been put to the test at a variety of other locations with our news department in WiFi mode and ethernet mode and has done nothing but exceed expectations.
if anyone has serious questions about this product and considering purchase i would be happy to discuss our experiences with it at WTIC
my desk 860 284-9288
Jeff Hugabonne
Rolf,
I cannot find where you are presenting the papers at NAB. Care to shed some light on this, such as time, place, date? Also, What is the bandwidth max for this unit. Intraplex (Netexress) has a 7.5 meg capapcity, what is the cap. on the Worldnet Oslo?
G
I cannot find where you are presenting the papers at NAB. Care to shed some light on this, such as time, place, date? Also, What is the bandwidth max for this unit. Intraplex (Netexress) has a 7.5 meg capapcity, what is the cap. on the Worldnet Oslo?
G
Broke,
thanks for asking
Here are the NAB sessions we will be presenting:
Radio Technology Advancements
April 13 • 1:00 p.m. - 5:30 p.m.
Las Vegas Convention Center S226/227
2:30PM - 3:00PM
Can the Public Internet be Used for Broadcast Applications?
Presenter: Simon Daniels, Sales Manager - Europe, APT
------------------------
Alternative STL Technologies
April 14 • 1:00 p.m. - 2:30 p.m.
Las Vegas Convention Center S228
1:30PM - 2:00PM
The Best of Synchronous with the Best of IP
Presenter: Guy Gampell, Sales Manager - Asia, APT
--------------------------
Regarding bandwidth on the Oslo. Of course this mainly comes into play when multiple unicast streams are used, since a frame handles 6 audio cards (6 bidirectional stereo pairs or 12 unidirectional (send or receive) stereo pairs)
The biggest project we have done so far was for the BBC with 24 routes at 576kbps/256byte packet with 10ms jitter buffer. This tops out at 13.824MBps (simplex) with total bandwidth of 16.7 MBps (duplex).
I am sorry that I cannot find the spec for max, but you can see from this example that Oslo is no light-weight.
Hope to see you all in Vegas.
Rolf Taylor
APT North America
thanks for asking
Here are the NAB sessions we will be presenting:
Radio Technology Advancements
April 13 • 1:00 p.m. - 5:30 p.m.
Las Vegas Convention Center S226/227
2:30PM - 3:00PM
Can the Public Internet be Used for Broadcast Applications?
Presenter: Simon Daniels, Sales Manager - Europe, APT
------------------------
Alternative STL Technologies
April 14 • 1:00 p.m. - 2:30 p.m.
Las Vegas Convention Center S228
1:30PM - 2:00PM
The Best of Synchronous with the Best of IP
Presenter: Guy Gampell, Sales Manager - Asia, APT
--------------------------
Regarding bandwidth on the Oslo. Of course this mainly comes into play when multiple unicast streams are used, since a frame handles 6 audio cards (6 bidirectional stereo pairs or 12 unidirectional (send or receive) stereo pairs)
The biggest project we have done so far was for the BBC with 24 routes at 576kbps/256byte packet with 10ms jitter buffer. This tops out at 13.824MBps (simplex) with total bandwidth of 16.7 MBps (duplex).
I am sorry that I cannot find the spec for max, but you can see from this example that Oslo is no light-weight.
Hope to see you all in Vegas.
Rolf Taylor
APT North America
- Status