processed audio over comrex bric link

[stephend2]

I'm trying to get the station owner back happy again, he refuses to deal with the ~30ms delay introduced by the bric link over a t1. He HAS to hear processed audio in his headphones or it is not acceptable. My idea for this is to get a good stereo gen with composite clipper to put at the far end and bring the processor back to the studio, run aes into the comrex box and back out to the stereo gen with aes as well then run the left/right outputs to the console for him to monitor from. My only concern is how is the digital link going to handle the processed audio. I'm using FLAC which is supposed to be lossless and as far as i can tell it is. PCM is not an option as 44.1khz would not fit in the 1.5mb I have available and I have to make sure I have enough bandwidth left for control and monitoring as well as a 160k mp3 stream that feeds our AM station.

My second option was to obtain a backup processor and use that to feed the monitor. If I bring the main processor back to the studio then I'm going to run its composite output to the old STL and use that for backup. The reason we are still not on that full time is there is no line of sight at all. the signal varies all over the place and when it gets windy it can get quite noisy from the trees bouncing the signal around.

I'm looking at the BW Broadcast DSPmpx for the stereo gen, using DSPXTRA for processing now.

 

[OttoM]

Get your hands on a computer with XP (even an old one with a 1GHz processor will work), put in a cheap ASIO-compatible sound card (do NOT use ASIO4ALL) and buy Breakaway Live. There's so little delay with ASIO that even the most persnickety jocks won't notice.

 

[HowardMBurgers]

I wouldn't feed my processed program audio into a digital codec like that for two reasons:

1. Depending on the sample rates, etc., you may have some interesting artifacts appearing over the air.
2. Assuming you're using a all-in-one processor from Omnia or Orban, you won't be able to provide the peak control once it comes out at the other end.

I would place a decent analog AGC like a Compellor or Orban 424 ahead of the codec to keep the levels in check and give somewhat of a compressed sound. Assuming you run an average of 5dB of compression at the studio with an AGC, then either shut off the AGC on your processor located at the transmitter, or back off the AGC level 5dB to compensate.

 

[dspxscott]

I wouldn't feed my processed program audio into a digital codec like that for two reasons:

1. Depending on the sample rates, etc., you may have some interesting artifacts appearing over the air.
2. Assuming you're using a all-in-one processor from Omnia or Orban, you won't be able to provide the peak control once it comes out at the other end.

You're right about being careful about using codecs with preprocessed material.

The Original poster mentioned using the DSPmpX stereo generator and i thought i would chime in because this unit is just perfect for protecting peaks and dealing with overshoots from studio to transmitter links, both analogue and digial.
It has distortion controlling audio clippers with protection limiting to prevent clipper abuse. If that's not enough the pilot and sca protected composite clipper will help you turbo-charge things further should you decide to.
However if the link is transparent the encoder won't color the sound or add any distotion at all. It typically has 70db of stereo seperation and even has a built in silence sensing switching circuit to switch from anlog to digital or vice versa if your program feed should die.
Regards

 

[gtomas]

I'm trying to get the station owner back happy again, he refuses to deal with the ~30ms delay introduced by the bric link over a t1. He HAS to hear processed audio in his headphones or it is not acceptable. My idea for this is to get a good stereo gen with composite clipper to put at the far end and bring the processor back to the studio, run aes into the comrex box and back out to the stereo gen with aes as well then run the left/right outputs to the console for him to monitor from. My only concern is how is the digital link going to handle the processed audio. I'm using FLAC which is supposed to be lossless and as far as i can tell it is. PCM is not an option as 44.1khz would not fit in the 1.5mb I have available and I have to make sure I have enough bandwidth left for control and monitoring as well as a 160k mp3 stream that feeds our AM station.

It's a solid plan. FLAC is lossless and completely transparent, so in terms of audio quality it's the same as the PCM, apart from the encoding/decoding delay. Watch to leave enough headroom when feeding the encoder so that possible peaks don't get clipped in the link.

The DSPXtra and DSPmpX will go well together as the clipper in the DSPmpX is the same as in flagship DSPXtreme. In essence you would only be inserting a (transparent) link between the multi-band clipper and the main clipper. Use FM for the AES/EBU output (and set the de-emphasis to off) but make sure the clipper drive is set to low numbers, so that you don't clip twice (in the Xtra and then mpX again).


Regards,
Goran Tomas

 

[stephend2]

that is exactly how I was planning to do it, I'm not running the clippers hard at all, its an oldies format with sound quality as the highest priority.

I already knew to turn pre-emphasis off at the studio end and turn it on at the stereo gen.

I'm using the AC-ST preset just slightly modified.

 

[LA_Guy]

FLAC is transparent. What comes out is identical to what goes in-bit for bit.

 

[gtomas]

I already knew to turn pre-emphasis off at the studio end and turn it on at the stereo gen.

Actually, it would be better to let the processor work with the pre-emphasis turned on in the input menu. Leave the de-emphasis option turned off for the AES/EBU output, so that you don't unnecessarily de-emphasize audio at the output of the processor (which would require that you pre-emphasize it again at the stereo generator). By applying pre-emphasis in the processor and leaving it pre-emphasized at the output, you can the leave pre-emphasis turned off at the generator and just clip.

This way the multi-band limiters in the processor will work on pre-emphasized audio which provide more control on the high-end and would sound better than if you left the limiters operating on flat (non pre-emphasized) audio and then pre-emphasised audio in the stereo generator just before the clipper. Specifically with the BW DSPXtra, you can still have the control of the amount of high-end that will be controlled by the limiters vs the clipper by using the HF Clipping control.

As the codec you are using (FLAC) is lossless and non psycho-acoustical, it won't have any problem with encoding/decoding pre-emphasized audio (unlike other perceptual codecs, that would) so you are fine there as well.

So to summarize, turn the pre-emphasis in the processor input menu on. Set the source for digital output to FM (alternatively you could probably use MON as well). Set the de-emphasis in the digital output menu to off. Leave the pre-emphasis in the stereo generator off. Reduce the clipping drive at the processor so that you don't clip in the processor. Make sure you leave enough headroom when feeding the Comrex so that you don't clip in the link. Leave the pre-emphasis off in the DSPmpX. Adjust the input and clipper drive controls in the DSPmpX to achieve the desired sound/amount of clipping.


Regards,
Goran Tomas

 

[stephend2]

alright that sounds good, one last question:

can i still turn on the de emphasis on the analog l/r outputs for monitoring in the studio without affecting the digital output?

 

[gtomas]

can i still turn on the de emphasis on the analog l/r outputs for monitoring in the studio without affecting the digital output?

Yes. Each output has separate de-emphasis and source (FM/MON/DR) control.


Regards,
Goran Tomas