Harris CD Link/RDS/Stereo Generator

[chriscollins]

Ok... So I have worked out repurposing an uncompressed Harris CD Link to replace my old 6010 Analog STL. I have a couple of questions. Obviously, I could go AES into the STL, set the gain low, so it wouldn't clip and put my 8300 at the transmitter. Option 2 is buying a stereo generator (I saw Omnia makes one, the irony), and leaving the 8300 at the studio, with AES L/R going into the STL and generating the stereo signal itself at the transmitter site. At any rate, I will need to pass some data over the data channels to get RDS to my site.

Do any of you guys have any suggestions on which of these two routes might be better and why?

 

[GeneSavage]

Actually, my vote would be for a "Plan C," using a gentle AGC or simple peak limiter at the studio just to catch overshoots and setting the processor at the transmitter.

With an uncompressed digital line you can set the levels where somebody's got to hit it pretty hard to get any gain action at all and not have to worry about noise floor.

Others may have better ideas, though.

 

[fm-engineer]

I have a station with the exact same scenario. They use an Ariane Sequel to provide AES into the CD link and the 8300 is located at the transmitter (always best place for it), then composite into the transmitter. The Ariane protects the STL from overshoot. On the 8300, I have the AGC gate as far up as it will go, -15 I believe, and it sounds great. You can try to disable the AGC in the 8300, but for some reason gating the AGC out sounds better to my ears. The station has great ratings and is a country format.

Ok... So I have worked out repurposing an uncompressed Harris CD Link to replace my old 6010 Analog STL. I have a couple of questions. Obviously, I could go AES into the STL, set the gain low, so it wouldn't clip and put my 8300 at the transmitter. Option 2 is buying a stereo generator (I saw Omnia makes one, the irony), and leaving the 8300 at the studio, with AES L/R going into the STL and generating the stereo signal itself at the transmitter site. At any rate, I will need to pass some data over the data channels to get RDS to my site.

Do any of you guys have any suggestions on which of these two routes might be better and why?

 

[chriscollins]

Thanks guys. I had thought about the Arianne unit, although it seems hard to find a price on it. All of their 'retailers' on the site don't show it in the online catalogs.

I that thought about Plan C (with the Arianne). Did you have another unit in mind?

I was also thinking that since I had no noise floor, I could just have the AES pulled way back going into the cd link, giving me lots of headroom for when the mic or a caller hits.

 

[fm-engineer]

If I was taking that route I would make sure not to use the on-board A to D converters of the CD link, (same goes for moseley), and would rely on a high quality converter-24 bit or better with 16 bit shaping. I have a few stations using lucent A to D's and they work very well. If the AES is already created by the consoles or station switching then you won't have to worry about it. I have one station using a compellor D with has AES out. They seem to work ok for protecting STL's. Good luck

Thanks guys. I had thought about the Arianne unit, although it seems hard to find a price on it. All of their 'retailers' on the site don't show it in the online catalogs.

I that thought about Plan C (with the Arianne). Did you have another unit in mind?

I was also thinking that since I had no noise floor, I could just have the AES pulled way back going into the cd link, giving me lots of headroom for when the mic or a caller hits.

 

[chriscollins]

If I was taking that route I would make sure not to use the on-board A to D converters of the CD link, (same goes for moseley), and would rely on a high quality converter-24 bit or better with 16 bit shaping. I have a few stations using lucent A to D's and they work very well. If the AES is already created by the consoles or station switching then you won't have to worry about it. I have one station using a compellor D with has AES out. They seem to work ok for protecting STL's. Good luck

Thanks guys. I had thought about the Arianne unit, although it seems hard to find a price on it. All of their 'retailers' on the site don't show it in the online catalogs.

I that thought about Plan C (with the Arianne). Did you have another unit in mind?

I was also thinking that since I had no noise floor, I could just have the AES pulled way back going into the cd link, giving me lots of headroom for when the mic or a caller hits.

I come out of the board AES, there will be no conversion to analog. I think I am just going to go board to CD Link and not worry about leveling until it gets out to the transmitter and optimod.

 

[WNTIRadio]

Well, I wouldn't say that there's "no noise floor". Even digital has a noise floor, usually about the same as high quality professional analog gear, maybe a couple dB better. Most likely -96dBfs or close to it if it's 16 bit audio. So the lower you run your average levels, the higher your noise floor.

If you run your levels peaking mainly at -20dB, that now becomes your reference level. So, your noise floor is now 76dB below your audio, instead of close to 93dB (allowing for 3dB of headroom after the leveler/limiter). If you ran at -30dB to have headroom, then your noise floor is now only 66dB below your audio, about that of a reel to reel running 15ips or a Dolby C CrO2 cassette.

You really should put a leveler or limiter of some sort in front of the STL. The digital Compellor will work fine if you don't have the cash for the Ariane (that would be my #1, 2 and 3 choice). Clipping the digital input of the STL will sound NASTY on the air. Running your average levels too low will just bring the noise floor back up and negate the advantage of the quieter digital path.

 

[WNTIRadio]

Expect to pay around $2400 for the Ariane, by the way... $1,100 for the digital Compellor.

 

[chriscollins]

Well, I wouldn't say that there's "no noise floor". Even digital has a noise floor, usually about the same as high quality professional analog gear, maybe a couple dB better. Most likely -96dBfs or close to it if it's 16 bit audio. So the lower you run your average levels, the higher your noise floor.

If you run your levels peaking mainly at -20dB, that now becomes your reference level. So, your noise floor is now 76dB below your audio, instead of close to 93dB (allowing for 3dB of headroom after the leveler/limiter). If you ran at -30dB to have headroom, then your noise floor is now only 66dB below your audio, about that of a reel to reel running 15ips or a Dolby C CrO2 cassette.

You really should put a leveler or limiter of some sort in front of the STL. The digital Compellor will work fine if you don't have the cash for the Ariane (that would be my #1, 2 and 3 choice). Clipping the digital input of the STL will sound NASTY on the air. Running your average levels too low will just bring the noise floor back up and negate the advantage of the quieter digital path.

Great post... Thanks. I can probably get the Arianne.

 

[RolfTaylor]

While not directly relevant to this discussion, one nice advantage to the Worldcast: APT codecs is they offer apt-X at 24 bit resolution (assuming you have the bandwidth). 24 bits with non-destructive apt-X allows you to have the processing at the TX site without the compromises of multiple chained processors. In many cases this is a better option than 16 bit linear.

Rolf

 

[gtomas]

While not directly relevant to this discussion, one nice advantage to the Worldcast: APT codecs is they offer apt-X at 24 bit resolution (assuming you have the bandwidth). 24 bits with non-destructive apt-X allows you to have the processing at the TX site without the compromises of multiple chained processors. In many cases this is a better option than 16 bit linear.

Very good point! Indeed, this would be another significant advantage of using APT codecs for STL links. Having 144 dB of SNR allows you to leave 30 dB of headroom for peaks above normal program level without any issues with noise - no pre STL processing needed.


Regards,
Goran Tomas

 

[chriscollins]

While not directly relevant to this discussion, one nice advantage to the Worldcast: APT codecs is they offer apt-X at 24 bit resolution (assuming you have the bandwidth). 24 bits with non-destructive apt-X allows you to have the processing at the TX site without the compromises of multiple chained processors. In many cases this is a better option than 16 bit linear.

Very good point! Indeed, this would be another significant advantage of using APT codecs for STL links. Having 144 dB of SNR allows you to leave 30 dB of headroom for peaks above normal program level without any issues with noise - no pre STL processing needed.


Regards,
Goran Tomas

This is one of the reasons I come here. I didn't even know APT was still around. I only thought about them in regards to lossy compression on automation (of which, I use NO compression). I went to their site and had no idea they were still around. Thanks for the tips, guys.

FWIW, I think I decided on Ariane at the studio, CD Link, 8300 at the TX. I'll just use the serial pass through to get the RDS out there and put the encoder at the site with the 8300.