1)volume lines, compression
2)sony noise reduction
Steve S.

could someone explain volume line and compression and how they relate to keeping the voice volume the same throughout a song?

also is there a way to regulate this while recording? a lot of times people say certain words louder then other parts in the song and they only way i can fix this is in sound forge by lowering volume manually and playing it by ear....which id really like to do more accurate

The ear is a pretty good tool, dont hesitate to use it. Theres tons of good stuff written on compression all over the web, I'd start here:

http://www.recordingproject.com/articles/article.php?article=4

what i'd like to be able to do is control the sound (voice) level coming in so it's not too loud (and not to quiet) so its the same level all through the track....that was a good article you sent me, but i dont know if the "make up gain" will fix only the low parts after compression or turn the whole track up or what? plus does vegas even have that ability?

Vegas has basic audio compression on each track. I would turn off the "auto gain compensation" or whatever name they have for auto-make up gain, and do it manually. The trick with a compressor, is once you squeeze down the highest levels of a track with it, you can turn the whole thing up. Right now the problem you are likely having, is if you turn up the track, many passages become WAY too loud right?

The compressor will turn down those really loud parts for you ( in theory). Once the average volume of the whole track is lower, but more uniform, you can turn it up to where you wanted it.

sorry for so many questions but I am new to this (kinda) and it's very fascinating...thanks for replying...

can the compression on a track be set and used while recording? or can it only be applied afterwards?

one of our group members voice tracks is always all over the place, with volume levels really high and really low at certain points within the song....i just wanna fix it up so all 3 of us have exactly the same level vocal track

Compression can be used during recording, and in days of yore, was kind of necessary, in order to keep recorded material above the noise floor, while being below the onset of undesireable distortion. This is usually the job of a hardware compressor, although some apps will let you track with plugs ( tho since you could apply the same plugs with the same results afterwards there may be little point). You might want to track with compression, if the singer you are working with knows how to "play" the compression, just as a guitar player would play a guitar, in the hands of a skilled singer, a compressor is an instrument in its own right, slap bass players many times have this skill as well. Another case may be where you have a "magic" compressor, like an LA-2A or something you REALLY like analog and do not want to undergo another DA/AD conversion to use. Best bet in your case I suspect is to leave it off during tracking, tho you could set up a parallel monitoring track and throw a compressor on that, just so you can hear it, tho not recording it.

Exactness is kind of opposite of artistic in many cases. I mean you should want a little bouncing around of levels and things to make it alive. Control is one thing, oppression is another. I would give this a try: Set one of vegas's compressors to 50ms attack, 50 ms release,ratio("amount" in vegas) 2.5:1, auto gain compensation off. Mess with the threshold so that about 10dB of reduction is showing around the loudest parts of the song and see what happens. From there start going to higher ratios, with higher thresholds, so that you can leave more of it alone, while doing more squashing on the really loud stuff.

If you find that attack times on the compressor seem to need to be less than about 35-40 ms, its probably time for a pop filter on that mic

thanks

we have a pop filter....so plosives and what not don't stick out...so should i forget about the attack part you mentioned? i dont want the beginning of words cut off...

but still sometimes certain parts of the song are louder, or verse 2 is louder than verse one etc...i mean i can play it by ear and get them to sound good but i just wondered if there was an automated way to do it

"we have a pop filter....so plosives and what not don't stick out...so should i forget about the attack part you mentioned? i dont want the beginning of words cut off..."

Dont confuse a compressor with an expander or gate. The attack controls how long before the compressor start turning DOWN sounds, opposite of a gate where it controls how long the sound plays before you actually hear it. Shorter attacks will turn it down earlier

Many times you will have to use Vegas' volume envelops and gain handles to make things sit right, volume wise. A compressor will get you close, but much has to be done by hand, unless the singer has a great degree of control. If there were an automatic way to do everything, the art of engineering wouldnt be a life's work, and all kinds of guys with Digi 007 in their bedrooms could say they have recording studios. Oh wait they already do

As I originally said, "Get educated and become an audio engineer." Before you can use a compressor to achieve what you want, you need to understand the controls on a compressor and what they do. These controls have a wide variety of settings to be able to set the compressor up to handle many types of sources which will have differenct characteristics within themselves. While Pipe's original suggested settings are nearly in the ball park, there is a couple things that may be quite wrong considering the source material, and may be farther off since he cannot hear the actual recording and dynamics within it. The suggested attack time is probably too slow for something like spoken word, and the release time is too short, that you may actually hear the compression. When using a compressor, you should actually NOT be able to hear the compressor. Here's some of controls and the description of what they do along with some other useful information.

Compressor- It is named a "compressor" because it's function is to "compress" the dynamics range. That is, it will bring the louder parts of the audio closer in level to the quiet parts. It does this by turning down the volume of the louder parts, while leaving the quieter parts at their current level.

CONTROLS-
1. THRESHOLD: This is the setting that determines at what level the compressor will start to work. Anything below this level will go unaffected by the compressor, except during the release time duration after the threshold has been exceeded.

2. COMPRESSION RATIO: This is the amount of dB the compressor will turn down the audio depending on how much louder the actual audio is over the Threshold. Example. If you have a 4:1 compression ratio. The actual audio level must go 4 dB above the set threshold for 1 dB of gain past the threshold to occur. So what use to be 4dB past the threshold value, now get's compressed into 1dB above the threshold. If the audio exceeded the threshold by 10dB, it would now exceed it by 2.5dB. What use to be 10dB of dynamics get's comressed into 2.5dB of dynamics.

3. ATTACK TIME: This is the reaction time for the compressor to start working (ie start turning the volume down) once the THRESHOLD has been exceeded. Example: Attack Time=10mS. With this setting the audio exceeds the threshold value, but this level get's unaffected by the compressor until 10 milliseconds later, where at which point it will then start to turn the volume down.

4. RELEASE TIME: This is the time the compressor stops turning down the level once the audio level goes below the threshold value. EXAMPLE: Release Time=100 mS. The Threshold has been exceeded and starts to turn down the level of the audio after the initial attack time setting. The input audio then goes below the threshold value, the compressor will still turn down the audio so that dynamics are maintained for the next 100mS. After the threshold has remained below the threshold for 100mS, then the compressor will stop turning down the level.

5. GAIN: This is the level you adjust to bring the level back up once it's been processed by the above compression settings. EXAMPLE: If the "compression reduction" meter the majority of the time is reading "-6dB", then you might want to raise the GAIN to +6dB to bring the output back up to compensate for the reduction done by the compression. This brings the peaks that where turned down during the compression back to where they originally where. Therefore, what the compressor did was reduce the level of the loud parts, making them closer in level to the quieter parts, then you boasted the overall gain back up making the recording louder in level, thus the quieter parts are now louder.

AUTOGAIN COMPENSATION: Automatically increases the GAIN, taking into consideration the gain reduction amount.


Now for Voice type of work, you usually want a fast attack time, along with a slow release time. If the release time is too fast you will be able to hear the compressor working causing a pumping effect. That's what I meant by NOT hearing the compressor work. 50mS is along the lines of being considered fast. Probably better settings would be a 5-10mS attack, with a release around 150mS-300mS, but it will vary from here depending on the audio and if you hear any pumping going on. As a reference the human ear can usually detect a delay in the order of around 20mS or two discreet sounds. That's why the suggested attack time may be too slow, you can actually hear 50mS or there being a difference in level, depending on the compression ratio. A 2.5:1 compression ratio also, may be not agressive enough if the dynamics are wide like you are saying and you are trying to make everything the same level. Try something like a 8:1 ratio, and adjust the threshold so that the majority of the time it's reading around -10dB. Then adjust the GAIN to +6dB to compensate.

Hopefully this makes sense. My recommended settings are considering the type of audio (voice) and the problem description (ie too much dynamics wanting everything to sound the same in level) and may vary, but should get you in the ballpark for this particular use of a compressor. Hopefully, you can figure out the rest with the descriptions of the controls I gave. These setting should vary quite a bit depending on the source material, thus in your original question you ask how to make "music", and "voiceover" at the same level. Thus the frequency and transient content of music is quite different from voice, and will thus have to be different. So now you need to be able to understand the differences between different sounds and know how to adjust the compressor appropriately, so that you don't hear it working.

As far as background noise, using a compressor will actually bring the noise level up. Hopefully that makes sense now.

Steve S. response is probably the best and most concise, but then he already knows about the settings in a compressor and knows better to give suggested settings without hearing the source material. There's a lot of experience in working with audio that comes into play and why we all have jobs. Don't expect there to be any one button push automatic levelers and noise reducers that will be developed to substitue for that experience.

(Always good to read and re-read the basics...this triggered a question about attack time and release time:)

So the compressor kicks in (or kicks out) after waiting the specified amount of attack time (release time). Is this kick "instantaneous" or "gradual"? Is this something that varies depending on the quality of the compressor?

holy shit what a reply!

thanks for taking the time to type all that i'm printing it out and taking it home to play around tonight...

so when you say "hear the compressor" are you saying that im going to notice it for sure? as in, it's unmistakable?

i know that attack means the time it takes for something to start...but if a guy says the world "player" like this "PLAYer" will the compressor fix that PLAY part?

"So the compressor kicks in (or kicks out) after waiting the specified amount of attack time (release time). Is this kick "instantaneous" or "gradual"? Is this something that varies depending on the quality of the compressor?"

In a word "yes". The adjustments I gave are the basics of a compressor. There are other factors which will effect the sound quality. Many compressors use a selectable "knee" type or a selectable no knee, which makes the compression ratio variable at different levels on the threshold exceeding points. Many software/digital hardware compressors give a visual of the knee type, which is similar to common attack/release envelope.

"so when you say "hear the compressor" are you saying that im going to notice it for sure? as in, it's unmistakable?"

It may not be very obvious especially if you are not famaliar with the sound. Engineers with a good ear can usually notice the effect of the compressors pumping sound. A good exercise for getting accustomed to the pumping sound is to make extreme adjustments to the parameters (ie quick attack/quick release, high compression ratio on something that has large dynmics, like a snare hit which has a high transient with a tail). On a song that has not been overcompressed, the kicks and snare hits usually have the highest peaks. If a compressor is overly used you can easily hear the song ducking everytime the kick and snare hit. The major difference between a good compressor and a bad compressor, is that with a good compressor you can't hear the compression happening. You just notice it sounds better and even warmer as without it. A bad compressor you can hear pumping and noises.


"i know that attack means the time it takes for something to start...but if a guy says the world "player" like this "PLAYer" will the compressor fix that PLAY part?"

Yes, in theory you should be able to do this with a compressor if your attack, release, compression ratio, and threshold are perfect for this one word and you're processing only that one word on the track. Probably a much quicker way to do it, is to do an edit in Vegas and split the word between "PLAY" and "er", then adjust the volumes of each so they match with a slight crossfade.

I was answering reign who appeared to be in a band and not spoken word, but your point is well taken, no blanket settings, but I bet mine get him close :)