Audio Processing Advice?

[amos]

Hey,

I've been enjoying reading about the early Optimods, and different suggestions as to what to put in front of them. Honestly, most of the stuff is a bit over my head, but I still pick out some things I can understand and enjoy.

I was wondering if any of you guys would mind giving me a critique on the audio on my internet station. I'm streaming mp3-PRO at 64kbps/44khz. If you know anything about mp3-PRO, you are aware that the sampling rate cuts to 22khz on a non mp3-PRO player. So, if you'll tune in using an mp3-PRO player that would give you a more accurate sample. Now if I could just get the listeners to do that.

I'm using older, 80's vintage equipment. Before you say it, yes I have "flashing light syndrome." LOL

For the airchain, I'm using an Autogram Minimix 8A console, a Harris MSP90 AGC, a CRL SEP Multiband proccessor, and finally a Harris MSP90 FM Limiter. I also use a Valley 400 mic processor, and an RE20.

Here's the deal. I adjust the audio to what I think sounds good, but I am no CE. I suspect that much higher quality audio is possible from the equipment that I have if adjusted correctly. But when I try to do anything, I end up with something that sounds too squashed, so I revert to what I have now.

If someone would tune in and critique my audio, I suspect that your experienced ears would hear things that I just can't recognize. I want what most CHR stations want, bright, clean, clear, no pumping, nice low end. I'm not in a loudness war, but I do want presence. In short, I want to sound like I knew what I was doing when I set it up! LOL

Please remember my choice of format, CHR, when you reply. I want to sound clean and clear, but I don't want to sound like a classical music station.

Someday, maybe, I hope to get a pair of Audio Prisms, or a Compellor. But in the meantime, what I have is what I have to work with.

Thanks!!

http://www.live365.com/stations/wava2

-amos

 

[1q2w3e]

Just curious how you plan on using the optimod and what model.

FM Optimods typically have a composite output, and the AM versions are designed to offer asymetrical modulation and a NRSC roll off.

Hey,

I've been enjoying reading about the early Optimods, and different suggestions as to what to put in front of them. Honestly, most of the stuff is a bit over my head, but I still pick out some things I can understand and enjoy.

I was wondering if any of you guys would mind giving me a critique on the audio on my internet station. I'm streaming mp3-PRO at 64kbps/44khz. If you know anything about mp3-PRO, you are aware that the sampling rate cuts to 22khz on a non mp3-PRO player. So, if you'll tune in using an mp3-PRO player that would give you a more accurate sample. Now if I could just get the listeners to do that.

I'm using older, 80's vintage equipment. Before you say it, yes I have "flashing light syndrome." LOL

For the airchain, I'm using an Autogram Minimix 8A console, a Harris MSP90 AGC, a CRL SEP Multiband proccessor, and finally a Harris MSP90 FM Limiter. I also use a Valley 400 mic processor, and an RE20.

Here's the deal. I adjust the audio to what I think sounds good, but I am no CE. I suspect that much higher quality audio is possible from the equipment that I have if adjusted correctly. But when I try to do anything, I end up with something that sounds too squashed, so I revert to what I have now.

If someone would tune in and critique my audio, I suspect that your experienced ears would hear things that I just can't recognize. I want what most CHR stations want, bright, clean, clear, no pumping, nice low end. I'm not in a loudness war, but I do want presence. In short, I want to sound like I knew what I was doing when I set it up! LOL

Please remember my choice of format, CHR, when you reply. I want to sound clean and clear, but I don't want to sound like a classical music station.

Someday, maybe, I hope to get a pair of Audio Prisms, or a Compellor. But in the meantime, what I have is what I have to work with.

Thanks!!

http://www.live365.com/stations/wava2

-amos

 

[amos]

Just curious how you plan on using the optimod and what model.

FM Optimods typically have a composite output, and the AM versions are designed to offer asymetrical modulation and a NRSC roll off.

sorry, i didn't mean i was going to use an optimod. my airchain consists of the Harris MSP90 AGC, CRL SEP 800, and Harris MSP90 FM Limiter. All of this equipment has balanced inputs and outputs.

-amos

 

[amos]

ok, if nobody has time to listen, could somebody that is famaliar with the processors i'm using give me some hints or maybe some settings that might sound good on a Top 40 format?

Harris MSP-90 AGC
CRL SEP-800
Harris MSP-90 FM Limiter

Thanks guys, any help at all is REALLY appreciated.

-amos

 

[TPO]

Since most if not all of your music is already processed, and you are running a mic processor, I think you will get the best sound by using a minimum amount of processing unless for some reason your levels vary greatly. How much gain reduction do you see on the Harris AGC meters with normal programming levels?

 

[amos]

Since most if not all of your music is already processed, and you are running a mic processor, I think you will get the best sound by using a minimum amount of processing unless for some reason your levels vary greatly. How much gain reduction do you see on the Harris AGC meters with normal programming levels?

i wouldn't say the levels vary greatly but they do vary more than i'd like. and the station is in automation mode most of the time, too.

gain reduction on the harris AGC is ave 6db, peaks 9db, valleys 3db.
output level on agc 0db, rare +3 peaks
crl gain reduction 6db, operation medium, compress, multiband. output eq, low 2o'clock, med low 11o'clock, med high 11o'clock high 2o'clock
harris limiter gain reduction ave 1db, peaks 3db, rare peaks 6db. output 95-100% peaks, fm mode soft.

harris agc settings: compression attack time: 1ms, compression recovery time: 2 sec.
expansion attack time: .3 secs (3/10) expansion recovery time: 1 sec.
compression slope: 6:1 expansion threshold: 3db

harris limiter settings: high freq protection threshold: 400hz
limiter recovery time: dual mode / fast
pre-emphasis: 0u sec, de-emphasis: 0u sec
maximum multiplier gain: -10 db

thank you!!

 

[Turntable]

Amos:

I have some experience with processing for internet streams. It's a totally different animal than processing to feed an on-air signal.

First, you are dealing with a very lossy codec designed for speed of delivery. Retention of signal integrity is more of an afterthought. Do not make the mistake of listening to the output of your airchain while making adjustments. You *must* deal with the delay and listen on a separate computer to your stream.

Here's what I would start with:

Process hard with the multi-band CRL. a slow leveller before and after the CRL is key. Processing for streaming is like trying to stuff 10 pounds of *** in a 5-pound bag. I'd use the Limiter in front and the AGC after. Make sure FM pre-emphasis-- if any-- is disabled. It will drive the codec nuts. I'd also suggest some eq in line before the final feed to the host box. It will allow you to kill frequencies that cause the codec to get nasty. Don't try to pass anything above about 10k with your current settings. It'll just sound nasty.

One major point of contention with the major streaming programs is the auto leveling function built into them which in most cases cannot be disabled. Try to peak for -3 in the software to mimimise its effect.

Feel free to email me if you have any more questions.

proaudio 5 at hotmail dot com...

-A

 

[wgliradio]

It sounds OK. I think you can loosen it up a little. It's a little squashed. Most of what you are playing is mashed already ebfore it hits and airchain.

I would ditch the Harris FM limiter and use the SEP-800 in the limit mode and drive the codec with that, watching for peaks. Nothing should crest above -5. Make sure you test with voice and music. Listen to the final product off a PC. Use the Medium setting on the CRL with the G/R at -9. Remember, when the CRL shows the 0 LED's lit, that means there is approx 9 db of gain reduction going on when selected for -9 (that's what the switch and LED's mean.. it would -12 if selected etc).

 

[amos]

hey, i tried to follow both of your suggestions somewhat. i'm currently patching around the limiter. in a way it sounds less squashed, and i can certainly see that it IS less squashed on the codec mod meter.

i'm not sure exactly what i'm hearing. i was pretty used to the way it sounded, so i'm not sure if this is what i want or not.

by the way, i took this opportunity to go to 96/32 (non mp3PRO). i may end up at something different. to me, 96/44 mp3 has too many artifacts, 96/44 mp3pro means if they don't have an mp3pro player they only get 96/22. (ugh) and 128/44 isn't available to me on live365.

 

[wgliradio]

You have to remember, nobody is dialing around from preset to preset fast enough to realize loudness like the Fm dial. I think you can do yourself a favor and allow a little more dynamic range, especially when you consider that the program material is already squashed. You don't need (and shouldn't be) balls to the walls here.

 

[amos]

yes, you're exactly right on that. the a/b comparison isn't nearly as good when you have to wait several seconds for a stream to re-buffer. still, i don't want to be really soft, either.

i had to go back to 64/44 mp3pro. 96 will come in a few days, i think. gotta get my ducks in a row first.

do i still sound too squashed? like i said, i'm so used to the way it sounded before, i'm not sure how to hear this new sound.

thanks a BUNCH guys...
amos

 

[wgliradio]

I'll have a chance to listen again later when I get home tonight.

What it sounded like before was that you were taking the output of an FM tuner and feeding it to the stream. The way you sound now is probably better for you, I will listen later. Alot of internet radio I have heard is very lightly processed. Usually the over processed stuff comes from off-air feeds fed to a codec... a no-no IMHO.

Also, even if you have defeated pre-emphasis, there may still be clipping going on in that Harris limiter even at low settings, which is a bad mix with streaming.

 

[amos]

gosh, in a way that sounded like a compliment. if i was taking the output of an fm tuner, and if the station was processed in a good way, then it seems like that would be a great sound.

however, i went ahead and took the harris limiter out of the chain. patched around it.

the level indicators on the codec sure do bounce around more now. it was easier to set the levels before.

i still have to take the program content off, and put a tone generator (well, actually cool edit's tone generator) on it to set the levels correctly. i'll do that after midnight... during the experimental period. haha, ok i'm old.

so you guys think i should have the CRL SEP800 on limit, instead of AGC? that's what i have it on now (i had it on agc up to tonight).

 

[Turntable]

Amos:

WGLI makes good points here, and I fully recognise where he is coming from with the dynamics and letting it 'breathe' a little. In a perfect world, he makes perfect sense. However, In my experience, the codecs used for internet streaming do not allow for very much latitude as far as dynamics are concerned. Going 'balls to the wall' as he puts it, is not what you want to do, but a happy medium must be found. The best way of acomplishing that is to keep your levels it a certain ballpark. You can spice that up a bit by adding a "signature sound". But remember, you are still trying to stuff ten pounds of crap into a five-pound bag and there is no way around it. You're still at the mercy of the stream provider.

All in all, I would say that the best way to handle audio for a stream is to keep the density high and the artifacts low. Also, you must keep in mind that many of your listeners are not audio-savvy, and their plastic speakers need all the help they can get. So another angle of techique is to strike a balance between good equipment and cheap plastic computer speakers..

Regards,

-A

 

[amos]

alan, wgli... how does it sound now? i'm on 64/44 mp3pro...

how do i obtain that signature sound, alan?

-amos

 

[Turntable]

Amos I'm listening now...

I'm sending you a PM with my phone number...

-A

 

[Turntable]

Processing is pretty much in the ballpark. Tonal balance is way off. If you call me I can walk you through some settings.

First observations:

1) No high end whatsoever. Very disturbing, to me at least.
2) Sounds like a reasonably good AM station
3) Keeps changing... call me please....

-A

 

[amos]

the lack of highs are because of the mp3pro codec, i'm using that now, but i will soon be changing.

)

amos

 

[wgliradio]

It doesn't sound nearly as mashed as yesterday. The FM tuner statement is not *really* a compliment because one processes much different for FM broadcast than for the internet.

This is not a shot at anyone in particular, because I have seen this everywhere, but alot of people process for HD and the internet with the exact same mindset as they do for AM or FM. You can't. You're dealing with a whole new ballgame. There are major differences. While you have lost pre-emphasis and clipping and deviation and button pushers that can percieve loudness differences instantly, you have gained other factors like codec artifacts and the absolute 0 that is digital audio (no running the speed limit here).

I don't believe processing for the internet should create as much of a signature as it does on the broadcast band. Processing should be an aid only to help keep levels in check and allow the audio to come over the artifacts. That's it IMHO. It should be as dynamic as you can make it without it falling apart.

 

[Turntable]

I think he's got it sounding pretty damn good now. You can hear the processing, but not enough to be fatiguing.

Nice job Amos

-A