Need Assistance on Processing

[Mr70s]

Hi Gals and Guys

I am looking for assistance on processing, here is what I want, I want to take all my music and make it sound
like super processed Top 40 FM Station (compressed and at the same levels) at the MP3 level before it gets uploaded to the servers

So what settings in Adobe audition can I make that happen !!

Thanks
BUD

 

[elchupacabras]

Your best bet would be not to pre-process, but get the Breakaway Processor http://www.claessonedwards.com/winmore.htm. Trust me, I have seen lots of them and tried them, but this is really the best.

 

[Tom Wells]

I second that opinion on Breakaway. If you have no experience with audio processing, and are just starting out, you can
do no better for the money. Using it since late last year, I am still fully satisfied and not tempted to tweak the settings.
It is probably the only processor a non-professional would ever need. The slight delay is totally acceptable, given the amount of improvement and the processing required to make the magic happen.
There is a free demo available. Hear for yourself and you'll be quickly converted.

Thanks, again, Leif for adding the features that make my AM thunder like WLS and WCFL did in 1966.

 

[wgliradio]

Get Breakaway and set it to French Kiss (that is the preset in the desktop version, don't know if it's in the full version). Play with the controls towards 100... you will get very close...

 

[Goat Rodeo Cowboy]

Your best bet would be not to pre-process, but get the Breakaway Processor http://www.claessonedwards.com/winmore.htm. Trust me, I have seen lots of them and tried them, but this is really the best.

I went to the link. I got the idea the Windows version "works in real time"... as in while you are streaming? Maybe while you are recording something live?

If I have a WAV file on my hard drive, can I process the WAV file and send it back to the hard drive for later use?

The verbiage on their website did not clear this up for me.

 

[Mr70s]

Thanks Guys, But that is exactly what I need to process each song
does any know of software or settings in Audition or pro tools that will do it

Thanks
BUD

 

[Goat Rodeo Cowboy]

Play with the Multi-band processor.

Menu: EFFECTS -> AMPLITUDE -> MULTI-BAND PROCESSOR


I'm still trying to find what is optimum for my needs so I don't have settings to offer you.

Play with it. There is no extra charge for experimentation. :

 

[LA_Guy]

There is something about audio processing streams that some here don't understand. Audio streams reduce bandwidth by perceptual coding. Perceptual coding works by removing 'redundant' audio information based upon the algorithm of the encoder. In simpler terms, it has it's designer's model of how humans hear sounds-and removes information that it PERCIEVES we will not hear. Most coders remove over 80% of the original data.

Unfortunately, there's a problem....most perceptual encoders are designed to be used with unprocessed or minimally processed audio. When you process audio before a perceptual encoder too agressively, you turn the designer's algorithm upon it's ear (pun intended). You need to realize this fact when setting up an audio processor for a stream.

My streams (see below) are minimally processed-yet they sound MUCH louder then most of the streams out there-including some streams that employ much more agressive processing. I've tried to adjust the processing to compliment the perceptual coding going on. Yet, it too is a compromise-ideally I should be employing different processing settings for the two streams, because each uses a different encoder (MP3 and AAC+). This is the reason why one astute listener reported that the AAC+ stream sounded more 'compressed' then the MP3 one, even though both have the identical amount of processing. If I had set things up to make the AAC+ one sound its best, the MP3 one would have sounded quite thin. I know this because I tried things set that way.

In some ways, processing net audio is a lot like a dance. When everything's right it works great. If one things wrong-everyone knows it.

So if you want your streams to sound their best, realize that everything works as one big system, not a bunch of individual units thrown together.

 

[gtomas]

There is something about audio processing streams that some here don't understand. Audio streams reduce bandwidth by perceptual coding. Perceptual coding works by removing 'redundant' audio information based upon the algorithm of the encoder. In simpler terms, it has it's designer's model of how humans hear sounds-and removes information that it PERCIEVES we will not hear. Most coders remove over 80% of the original data.

Actually, the encoder will remove only as much as you tell it to - and you tell it with the bitrate parameter. From the crude mathematical point of view (and in reference to uncompressed CD-quality):

320 kbps will remove 77.4% of data
128 kbps will remove 90.9% of data
48 kbps will remove 96.6% of data

Unfortunately, there's a problem....most perceptual encoders are designed to be used with unprocessed or minimally processed audio. When you process audio before a perceptual encoder too agressively, you turn the designer's algorithm upon it's ear (pun intended). You need to realize this fact when setting up an audio processor for a stream.

True!

In some ways, processing net audio is a lot like a dance. When everything's right it works great. If one things wrong-everyone knows it.

So if you want your streams to sound their best, realize that everything works as one big system, not a bunch of individual units thrown together.

Nicely put


Regards,
Goran Tomas

 

[jesseg]

the reason why one astute listener reported that the AAC+ stream sounded more 'compressed' then the MP3 one, even though both have the identical amount of processing

For starters... AAC+ has much worse peak control than mp3, for relatively the same perceptive quality. When your input to the encoder is already next to 0dB, some of that lack of peak control is going into clipping on decode.

Unfortunately, there's a problem....most perceptual encoders are designed to be used with unprocessed or minimally processed audio.

This is true, yes. And it's a big problem for most codec tuners. Not many groups have bothered to "weed out" the problems presented by "unusual" material... for instance like the Lame project has. That's WHY it is such a top-shelf codec. Because they pushed it's R&D harder than anyone else.

 

[LA_Guy]

the reason why one astute listener reported that the AAC+ stream sounded more 'compressed' then the MP3 one, even though both have the identical amount of processing

For starters... AAC+ has much worse peak control than mp3, for relatively the same perceptive quality. When your input to the encoder is already next to 0dB, some of that lack of peak control is going into clipping on decode.

Unfortunately, there's a problem....most perceptual encoders are designed to be used with unprocessed or minimally processed audio.

This is true, yes. And it's a big problem for most codec tuners. Not many groups have bothered to "weed out" the problems presented by "unusual" material... for instance like the Lame project has. That's WHY it is such a top-shelf codec. Because they pushed it's R&D harder than anyone else.

You are correct, which is why I'm running the output lever (volume) of Breakaway at -2 db. Running it at zero made the AAC+ stream sound really clipped. Lowering the output 2 db made a huge difference in the AAC+ quality, and only affected the MP3 volume a bit.

That was just one of several compromises. 48 kbps AAC+ also doesn't sound as bright as high bit rate MP3 does-but if I had adjusted the EQ and/or multiband processing to make it sound better, the MP3 would have sounded thin and trebly. Another compromise.

 

[GeneSavage]

A good side-question has been brought up here:

Why CAN'T stations do most of their gain riding, equalization, & compression on a track by track basis in the production booth?

I understand there's lots of places for peaks to get out of hand, but it seems like you could create the "sound" (density, etc) on each song and element and then use a final peak limiter or safety clipper to control overshoots before hitting the transmitter.

You might say, "levels may be different," but if you're on an automation system using digital files, everything can be easily controlled.

Live voice could be an issue, but solid processing on the mic I would think would take care of that issue.

Let's say, at least for internet radio: why CAN'T processing be tweaked to each file and applied BEFORE being fed to the stream, & Sound Solution or Breakaway Live or anything else be bypassed?

 

[jesseg]

There's no reliable way to tell what something has been processed with, if it's been processed, etc... and the whole point of processing is for consistency (of varying degrees)

If you're processing tracks destructively then consistency goes right out the window. Processing in realtime you keep all consistency because you know you're working with original content.

If you wanna listen to what it sounds like for pre-processed audio, and why not to do it... Check out 3WK
http://3wk.com/

 

[gtomas]

For starters... AAC+ has much worse peak control than mp3, for relatively the same perceptive quality. When your input to the encoder is already next to 0dB, some of that lack of peak control is going into clipping on decode.

This is true for any SBR enabled codec - be it HE-AAC (aacPlus), MP3Pro, etc. Which is why you need to leave some headroom, when driving the codec...


Regards,
Goran Tomas

 

[elchupacabras]

There's no reliable way to tell what something has been processed with, if it's been processed, etc... and the whole point of processing is for consistency (of varying degrees)

If you're processing tracks destructively then consistency goes right out the window. Processing in realtime you keep all consistency because you know you're working with original content.

If you wanna listen to what it sounds like for pre-processed audio, and why not to do it... Check out 3WK
http://3wk.com/

THAT IS THE MOST GOD AWFUL AUDIO I HAVE EVER HEARD ON THE NET. I like processed audio, but that SUCKS! No dynamics, pumped and slammed. I think I'd rather commit suicide rather than listen to that again.

 

[GeneSavage]

Ugh... no, I'm not a fan of that, either. I would question how you know that's a result of pre-processing, though, rather than a poorly set up compressor. I've heard even the best processor designs be set up to destroy audio.

In the scenario I was describing, you would NOT compress the song or commercial in the production booth, only to send it back through another aggressive box coming out of the mixer... I was thinking you could assure each song had approximately the same density and eq in the production, then have just a safety clipper between the board and the transmitter.

In this situation you would HAVE to have some consistent rules, not just "til it sounds good" because you're right to suggest you run the risk of every song coming out different. I'm thinking things like each track needs to have an average loudness of Xdb when normalized to Xdb, each song should have an approximate frequency plot of X-curve, etc.

Every single piece of audio that went on the station would have to go through this process, which may sound impossible but with greater and greater automation, it's really not that hard to imagine.

This way, if you had a very dense song, you could compress it less in the production booth, while a very "open" track could be crushed to match it up. That sounds to me like what we've been working to do with real-time processing for decades, through AGCs and gates and compressors and expanders and limiters and clippers...

It seems to me real-time processing is the "quick and dirty" way to accomplish this (always a compromise), while what I'm suggesting would take longer, and more attention to detail, but would result in a better sounding station...

...Wouldn't it?

BTW, I just figured out what the highs on this station reminds me of... ever had an audio file that was downsampled, then upsampled again improperly, and you get that fingers-on-a-blackboard high frequency trash?

Yeah...

 

[The F Mister]

Well to certain degree I can follow you on this. But remember that maybe the current PD likes it to sound this way and then another PD comes along and wants it totally different. What do you do? Do you start to pre-process your entire library over? The great benefit of "real-time" processing is that you can bypass/adjust it when you need/want. With pre-processing in the extend you describe it's pretty defining for the future. I'd rather pre-process to make hypercompressed material as clean as possible before saving it for playback.

 

[rrsounds]

Well to certain degree I can follow you on this. But remember that maybe the current PD likes it to sound this way and then another PD comes along and wants it totally different. What do you do? Do you start to pre-process your entire library over? The great benefit of "real-time" processing is that you can bypass/adjust it when you need/want. With pre-processing in the extend you describe it's pretty defining for the future. I'd rather pre-process to make hypercompressed material as clean as possible before saving it for playback.

That's my feeling, too. Once you've processed all your files, you are locked into that processing. At least until you do the whole library over again.

How about if the processing is not done on actual audio, but rather in the form of metadata associated with the file....?

Kind Regards,
David

 

[The F Mister]

That's pretty forward thinking. But how would it work? If you have a playback system that plays the audio file and reads the metadata file and starts to process according to this metadata? That's basically the same as real-time processing right? OTOH with that meta data you can see in the future sort of speak and have the software pre-process in real-time. That would be nice.

 

[rrsounds]

That's pretty forward thinking. But how would it work? If you have a playback system that plays the audio file and reads the metadata file and starts to process according to this metadata? That's basically the same as real-time processing right? OTOH with that meta data you can see in the future sort of speak and have the software pre-process in real-time. That would be nice.

Honestly, I was very slow to attribute any usefulness to the idea of metadata.

But as time goes on, I can't help but think that ANY source of additional information we can feed to the processor can potentially be useful. And it doesn't involve modifying the file itself, so there is no destructive factor to it.

One remaining concern (there may be others) is whether metadata could be manipulated unscrupulously, for example giving an advertiser a loudness edge for his own commercials. (yeah, I know, put it in the rate card!)


Kind Regards,
David