"Two Minute Blues Wank" - Studies in Mixing

[Khompewtur]

Two Minute Blues Wank

Let's say during recording each individual track, the clip meter never goes on (meaning the visual representation of my sound wave never hits the top or the bottom of the track), then my track hasn't clipped. Correct?

So here I have 6 tracks and when each is played individually it does not activate the clip meter.

Then I begin to play them back all in conjunction and the combined levels begin to activate the red meter. I don't necessarily hear any distortion at this point but I'm assuming this is a bad thing (is this clipping or something different?), so I begin adjusting the faders and envelopes until playback does not activate a red light, but the signal is as close to red lighting withouat actually triggering it.

At this point I mix it down to a .wav, .mp3 or whatever. Is it done? Why do some mixes sound louder than others?

The signal graphic is hitting the top and bottom of the track, but could that be only in a certain frequency? Am I looking at this in purely two-dimensional terms? Is there any tool that will show you what sort of levels you are getting across the frequency spectrum so you can see "holes" where you can bring up your levels?

If I record more than one instrument with the exact same settings, what are methods I might want to use to separate them?

Ahh.. one other thing I forgot. I see people commenting on things activating the red meter when they are listening to mp3's. When a mix is converted to an mp3, shouldn't the conversion process not allow anything above the maximum gain level, therefore redlighting should not be possible? You would be able to hear distortion in the sound of things, but it shouldn't physically be able to clip. Am i thinking of this wrong?

 

[jake-owa]

Two Minute Blues Wank

Let's say during recording each individual track, the clip meter never goes on (meaning the visual representation of my sound wave never hits the top or the bottom of the track), then my track hasn't clipped. Correct?

Jake- not really so. Those tracks will combine to create a louder or quieter signal depending on how the frequency content of each track interacts with other tracks in the mix. It's usually louder...and 6 tracks will be a lot louder combined than any of them individually.


So here I have 6 tracks and when each is played individually it does not activate the clip meter.

Then I begin to play them back all in conjunction and the combined levels begin to activate the red meter. I don't necessarily hear any distortion at this point but I'm assuming this is a bad thing (is this clipping or something different?), so I begin adjusting the faders and envelopes until playback does not activate a red light, but the signal is as close to red lighting withouat actually triggering it.

Jake-that is clipping and it will cut the peak off in a square wave which can sound bad or worse depending on how much of a transient was cut off.-

At this point I mix it down to a .wav, .mp3 or whatever. Is it done? Why do some mixes sound louder than others?

Jake-some mixes have more mid content and sound louder because we humans are tuned to that frequency range. Some mixes will have a higher average loudness and some will peak higher without actually being louder overall. So percieved loudness has a lot to do with the time aspect of music. Is the song hitting all the way through? Is the mix "thick" and full of sound or sparse and mellower? Compression can cause the average loudness to be much higher without actually raising the peaks too. So think: the snare may be the loudest thing in the mix but after a compressor it may be the average. The mix is not "louder" as in peaks but overall the sound reaches a higher amplitude.

The signal graphic is hitting the top and bottom of the track, but could that be only in a certain frequency? Am I looking at this in purely two-dimensional terms? Is there any tool that will show you what sort of levels you are getting across the frequency spectrum so you can see "holes" where you can bring up your levels?

Jake- spectrum analizer

If I record more than one instrument with the exact same settings, what are methods I might want to use to separate them?

Jake- which settings? Compressor? DO NOT use all the same settings on a compressor unless it's purely used for a peak limiter. I've had the problem where a certain attack time will leave a blip at each note begining that gets really loud and shitty sounding when all theguitars, bass and drums are doing it exactly the same.

Ahh.. one other thing I forgot. I see people commenting on things activating the red meter when they are listening to mp3's. When a mix is converted to an mp3, shouldn't the conversion process not allow anything above the maximum gain level, therefore redlighting should not be possible? You would be able to hear distortion in the sound of things, but it shouldn't physically be able to clip. Am i thinking of this wrong?

Jake- yes you are. The clipped mix will not be magically fixed by mp3 conversion. It will actually get worse in the encoding. The levels may not play on a listeners stereo at red levels but the sound will still be cut off at the peaks in a square wave. Not musical at all.

 

[Khompewtur]

ahhh.. spectrum analyzer. Any suggestions where to find one of those?

So then maybe that earlier one you had problems with was a clipping issue. Why do things not always sound screwed up even though the red light is seizure inducing?

oh shit p.s. thanks JakE!!

 

[jake-owa]

Sometimes it's just a tip of a transient peak. Sometimes it's more but often it can go unnoticed in a thick mix. Also instruments will sometimes render or mix differently than they sound in your program. Your soundcards summing sounds different than the software calculations most times.

I try to get each track to hit the top once or twice barely and no more. When your mix looks like a thick solid line when you zoom out you have done something wrong usually. Try to get mixes that have shape as a waveform. Preserve dynamics by not slamming it all together. I'd rather lose a couple bits of possible sound than get bricks for mixes.

Turning down the master buss to recuce clipping is not really the right way to do it. You want to leave that at unity and try to reduce the offending signal from the tracks. I start with cutting EQ and move on to cut volumes bit by bit though I'm sure there are a million ways to do it.

 

[Llarion]

Well, let's see. This mix is VERY thick, and guitar heavy. Even so, the drums cut, but have not enough meat.

Try mixing from the bottom up. Here's a VERY basic boilerplate: Set the drums first. Get em nice and fat, and set em so they peak about about -6db. Then add the bass. Get IT nice and fat, and keep your aggregate level from moving too much. Then add the rhythm guitar, pull back on the EQ from 250HZ on down, the part will thin out a little, but still be throaty and present in the mix, without driving the master levels up too much. THEN add your lead, and play with levels and field placement.

That should get you more even-feeling.

Nice playing, by the way!!!