Sidebar Sidebar
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.

Compresser/Limiters?

manosaurus

manosaurus

macrumors 6502
Original poster
Aug 22, 2006
285
0
I am going into a second project of producing songs for a rock band.

Here is my very simple current setup:

2nd Generation MBP
Using Garageband 3 software.
Presonus FirePod as mic preamps and interface.

Here is what I want different for my current project:

I want a device that will keep my signal as hot as possible without clipping it in the computer. Something that will boost the low parts of a signal so I can get an overall more even signal. I guess that that is a compresser/limiter right?

If so, then does anyone have any recomendations for one of these somewhere around $200 or so?

TIA!

p.s.

HEY ZIM!!!!!!!!!!!!
 
manosaurus said:
I want a device that will keep my signal as hot as possible without clipping it in the computer.
Click to expand...

you don't know it yet, but you don't want that at all. your digital buss has only so much headroom, and you'll eat it up fast by running your signals hot. and everything will sound like crap.

what you really want is to record your signals shockingly low, like -36 dB. it'll feel un-natural at first, but it's really the way to go.

....

there are a couple ways to even out un-even performances, and compression / limiting is the 2nd choice. the 1st is to automate volume. with a DAW, it's tedious work but technically easy, and it produces a great result.

i don't know the procedure for automation in GB, but in PT it's straightforward: draw (with the mouse) the volume curve to lower the loud parts and raise the quiet. don't go overboard, you'll still want some dynamics in your parts. (btw -- i'm talking about automating each part, not the song as a whole, though you an do that, too).

after that time-consuming but effective job is done, then you can put some compressor plugs on the tracks. do so if you like the vibe of the plug for that part, or it evens it out in a desirable way, but don't do it for the sake of doing it. it's really easy to go overboard w/ compressor plugs, and ime it's easy to take a part and make it sound very small.

do that too much and the entire song sounds small and insignificant. it's a difficult discipline, but sometimes less really is more.
 
I agree with Zimy. You do NOT want this at all.

In recent years, this phenomenon, referred to by sound engineers as "pumping" has turned sound recordings from even the most elaborate studios into absolutely unrefined mush.

To make a recording that stands out, you have to have dynamic presence... Uncompressed PCM has a dynamic range of around 96-97dB. By pumping amplitude on every channel, and then compressing/limiting it to normalize it just under a clipping volume, completely wastes all that dynamic range.... your recording will lack nuances that keep a listener's ears excited and interested in your song.

Listen to some of the better recordings of the 1970's and 1980's with sound engineers who knew what they were doing... The process of manipulating the spatial and dynamic characteristics of the overall mix to produce a superb listening experience is referred to as "sweetening". Not sure if you've heard this expression... but sweetening is almost a lost art these days.

Don't come in at maximum amplitude... that's what people have a volume knob for. You want headroom so you can play around with stuff... make sounds jump out and grab you at important points in the song, and also without drowning out the more subtle melodies, layers and instruments that you've arranged together in one piece.

Pumping just makes your stuff sound indistinguishable from everyone else's crap. It was a practice that became more rampant as cuts were being prepared for FM whereby the act of analog pumping was used to mask the noise floor inherent to FM transmission... but with digital audio the noise floor is not a problem. Even in 16-bit dithered Linear PCM, the noise floor is well below audible levels and pumping actually creates more problems than it solves because you're introducing levels of amplitude that at 16-bits per sample are likely to induce quantization error to a noticeable degree.
 
Avatar74 said:
Listen to some of the better recordings of the 1970's and 1980's with sound engineers who knew what they were doing
Click to expand...

no kidding. Queen, Fleetwood Mac, Heart (from the Dreamboat Annie days, anyway), Lynyrd Skynyrd, the list goes on and on. heck, listening to the Carpenters is like getting a lesson in audio engineering.

listening to Heart's Magic Man right now. ohhhhh, man.
 
Avatar74 said:
I agree with Zimy.


"To make a recording that stands out, you have to have dynamic presence..."

Aren't there other ways to have dynamic presence than just considdering the sheer amplitude? Percieved dynamics?



"...your recording will lack nuances that keep a listener's ears excited and interested in your song."



Well that's what modern engineers are doing. So are people not interested and excited in pop recordings since the 80's.


"Pumping just makes your stuff sound indistinguishable from everyone else's crap."

Doesn't the way the the music is constructed composed count at all?
Click to expand...

Well, to play DA here, there are other dimensions to music that dynamics. There is such a thing as percieved dynamics, i.e. instument/layer density could cause the listener to "percieve" things as being loud or relatively soft according to how many "things" are going on in the mix. All the while everything could be in or close to the red.

I totally see what you guys are saying. I come from a classical music backround where dynamic range, and an extremely wide one at that, is crucial and indispensible to say the least.

I am though trying to see things from the larger part of audiences, which doesn't include audiophiles, point of view. While a listener can experience a piece of pop music that is very well composed, fresh, innovative, very dynamic and so on, they can also experiance Fall Out Boy or some such band whose recordings would, from a modern engineering stand point, kick the ******* out of my piece. The point is the the Fall Out Boy, or whatever, will always sound better, bigger etc because at every moment the engineers have worked it out so that even some half-wit soft guitar intro sounds LIKE THE REAL THING.

Remember I am kinda playing devil's advocate here but how do you guys respnd to this?
 
manosaurus said:
Well, to play DA here, there are other dimensions to music that dynamics.
Click to expand...
absolutely. king in all this, imho, is songwriting and arrangement. as mix engineer, i see it as my job to ensure that the artistic intent is met: certain parts are exciting, vibrant, coming forward, what have you, while others take a step back, bring it down, prepare to be built back up, etc.

at the same time, i also have to decide which instruments will take focus during any given moment, and which ones will take a back seat. (that's why i mix alone :)

getting back to automation, that also plays a key role in adding dimensions. at any point during one of my mixes, there's a darn good chance something is moving / changing with automation. could be volume, could be panning, could be an effects send, could be something as wacky as a compression ratio or EQ value.

that's what i think is the most useful aspect of the DAW, and is a real gift vs analog. what we get in software, regarding automation, is either quite expensive or just not available in analog. or it's an "all hands on deck" situation (which is difficult when mixing alone).

All the while everything could be in or close to the red.
Click to expand...

if that's anyone's job, it's that of the mastering engineer. i fully believe the mix engineer's job is to turn over something that still has dynamic range.

there are a lot of theories why modern music, when viewed as a waveform, is basically a rectangle. in addition to the loudness wars, i think a lot of it has to do with keeping a song audible in a wide variety of listening environments: in the car with the window down, piped into a store or restaurant, as a bed under video, fighting street noise when played from an ipod. with some older material, there are times where the quieter parts just disappear. i guess they figure it's too much trouble for us to reach for a volume knob :)
 
manosaurus said:
Well, to play DA here, there are other dimensions to music that dynamics. There is such a thing as percieved dynamics, i.e. instument/layer density could cause the listener to "percieve" things as being loud or relatively soft according to how many "things" are going on in the mix. All the while everything could be in or close to the red.
Click to expand...

You have this backwards... the frequency saturation in such examples as the THX "deep note" which consists of multiple octaves introduced sequentially produces the illusion of greater amplitude while maintaining an actual average loudness around -27dBFS... well below clipping levels.

I totally see what you guys are saying. I come from a classical music backround where dynamic range, and an extremely wide one at that, is crucial and indispensible to say the least.

I am though trying to see things from the larger part of audiences, which doesn't include audiophiles, point of view. While a listener can experience a piece of pop music that is very well composed, fresh, innovative, very dynamic and so on, they can also experiance Fall Out Boy or some such band whose recordings would, from a modern engineering stand point, kick the ******* out of my piece. The point is the the Fall Out Boy, or whatever, will always sound better, bigger etc because at every moment the engineers have worked it out so that even some half-wit soft guitar intro sounds LIKE THE REAL THING.

Remember I am kinda playing devil's advocate here but how do you guys respnd to this?[/QUOTE]

I am both a lifelong fan of music and an experienced sound engineer (not full-time but I've done projects for national releases as well as for Dolby Digital certified soundtracks). I don't consider mixes like Fall Out Boy to be technnically proficient in the least.

The last example of a dynamic mix outside of jazz, film soundtracks and classical recordings that comes to mind, is the 1994 release Counterparts by Rush. Produced by Peter Collins, it has a very live, but managed and crafted sound in which all the instrumental nuances can be heard and nothing ever gets muddy.

You have a volume knob for a reason... The reproducing system at any volume level can still support the dynamic range of the input signal. Translation: Crank the volume knob during playback, not the input gain during recording. Doing the former will bring the low end of the dynamic range into audibility without distorting the high end provided your equipment is well matched and you're not cranking the volume knob beyond the power handling of the speakers... even so, any attempt at distorted playback will not damage the source recording. A quick adjustment and you're back to undistorted dynamic range.

Doing the latter, however, will produce irreparable damage to the recording, unrecoverable during playback. It's damaged for good... Garbage in, garbage out.

Here's the easiest way to test a mix... as you're playing back the mix, drop the volume down... if the louder instruments completely mask the quieter instruments at any point on the way down, you need to tweak the mix by bringing DOWN the louder instruments... NOT by bringing up the quieter ones... the only exception being if the quiet sounds are below the noise floor of the final format.

I'm a big fan of well-balanced speaker systems... Getting a couple tweeters and then two or four gigantic woofers with tons of power behind them is going to produce a lopsided image. However, having lots of smaller drivers to create a "big" image has the effect of saturating the room with sound without it ever sounding too bombastic. It sounds and feels loud, but without being painful. Combined with the right mix that takes dynamic range seriously, it's an absolutely blissful result in which even at the lowest levels the deep, bass frequencies aren't lost on the listener nor are the mids or highs.

That's what you want... at lower input signal levels, contrary to your original assertion, you can add MORE sounds and make it sound big, full and loud WITHOUT cranking the input signal to near-clipping levels. And you have a lot of overhead for sweetening (panning, fading, etc.) on each channel to give it an evenly distributed image where you can close your eyes and very distinctly imagine where in the room each instrument is sitting.
 
Avatar74 said:
Here's the easiest way to test a mix... as you're playing back the mix, drop the volume down... if the louder instruments completely mask the quieter instruments at any point on the way down, you need to tweak the mix by bringing DOWN the louder instruments... NOT by bringing up the quieter ones...
Click to expand...

this is an important point that i think is worth re-iterating. i always find that when i need to listen back at a higher volume for my mix to sound good, that i haven't done a good job. once it sounds good at lower volumes, it will still sound good at higher volumes.
 
Sympathy for the Devil

I'm no sound engineer, so everything is IMHO here. Music with good dynamic range is lovely and non-fatiguing. Some say that Radiohead OK Computer's over-compression prevented it from being an unequivocal classic. Given how little I listen to that album now, despite the song-writing genius, may give weight to that. However, recordings with levels running hot are exciting and have some instant appeal. When I digitise vinyl from 60s, 70s and 80 and play it along side contemporary masterings, the classics often sound meek. I've actually taken to running masterfully mastered classics through a multiband compressor then a brick-wall limiter so that they sit nicely in my itunes play list - awful, I know.

Back to the post, even though I sympathise with squashing the life out of a mix for a hot pumping sound, I'd never do it at the front end - I would have thought getting a fairly flat clean signal with lots of headroom would be the starting point, then do wicked things.

Given this is no audiophile project, I would have though doing it "in-the-box" with the available compressors, amp simulators and so-forth would serve you as well as some external equipment. Alternatively, for $200 get Logic Express which has some really excellent mixing and mastering plug-ins. Sure, it lacks organicity, but is cheaper and gives you lots of control.
 
The easiest way to get correct input gain is to have whoever is playing (yourself?) play the loudest part of the song and then adjust the gain to fall just below the red, usually around -2dB. That way dynamic range is preserved and the band can play the crap out of the song without hitting the ceiling. This works on both analog (I learned on an SSL to 1/4" multitrack tape) and DAW's.
 
QuarterSwede said:
This works on both analog (I learned on an SSL to 1/4" multitrack tape) and DAW's.
Click to expand...

i'm going to disagree with the "working with DAW" part. which should be easy to guess given what i wrote above.

find a meter that does inter-sample peaks and have a look at what you've recorded in the DAW. i think you may be surprised by how often and how far you go past zero.
 
zimv20 said:
i'm going to disagree with the "working with DAW" part. which should be easy to guess given what i wrote above.

find a meter that does inter-sample peaks and have a look at what you've recorded in the DAW. i think you may be surprised by how often and how far you go past zero.
Click to expand...
Now that you mention it I do remember being taught to be careful with the ceiling on a DAW and to keep it no higher than -6dB. I don't do much recording to digital but I would never record at such a low level because you simply can't get that information back. Amplification of any sort is going to be noisy especially when pumping that much amplitude back into it. Maybe I'm reading what you said wrong. I'm under the impression we're talking mainly about recording input and NOT mixing volume.
 
QuarterSwede said:
I don't do much recording to digital but I would never record at such a low level because you simply can't get that information back.
Click to expand...

while you're talking theory, i'm talking from some years of experience recording digital. that's after years of experience recording to tape. they're very different things, and guess what: not only can i "get that information back" while tracking somewhere between -18 and -36, but it sounds (much) better than when tracking at -6.

24 bits is a *lot* of headroom, and signal/noise ratio w/ digital tends to be pretty darn quiet. i'm here to tell you that even -6, according to your broken, lying-to-you DAW meters, is not only too high, it's a lot higher than -6. those peaks are really at +12 or so. you just don't know it.
 
zimv20 said:
while you're talking theory, i'm talking from some years of experience recording digital. that's after years of experience recording to tape. they're very different things, and guess what: not only can i "get that information back" while tracking somewhere between -18 and -36, but it sounds (much) better than when tracking at -6.

24 bits is a *lot* of headroom, and signal/noise ratio w/ digital tends to be pretty darn quiet. i'm here to tell you that even -6, according to your broken, lying-to-you DAW meters, is not only too high, it's a lot higher than -6. those peaks are really at +12 or so. you just don't know it.
Click to expand...
Sigh. You don't need to get defensive. I'm not challenging your experience.

Oh and stand alone top of the line meters aren't going to be lying to me especially since the results are perfectly good hot signals with no clipping. I'm not dumb enough to be using the meters in any software program or a console.
 
not defensive, you just sounded like you needed some guidance. if you're happy with your results, fantastic, and good luck to you.
 
Just a comment on meters.

You have to be sure where they are in the signal chain particularly when using stand alone meters on DAW outputs.

It's very true that a good meter on the output of my 192 or Prisms will accurately indicate signal levels at the output, but the damage is done.

Having those meters patchable at the input is the key, and will indicate transient overshoot.

Now I know what to listen for from DAW headroom collapse, I can go back to using my ears to judge wether an input signal is too hot.

-15 seems about right for most stuff, but drums and vocals (and strings oddly) are getting -20 currently.
 
zimv20 said:
while you're talking theory, i'm talking from some years of experience recording digital. that's after years of experience recording to tape. they're very different things, and guess what: not only can i "get that information back" while tracking somewhere between -18 and -36, but it sounds (much) better than when tracking at -6.

24 bits is a *lot* of headroom, and signal/noise ratio w/ digital tends to be pretty darn quiet. i'm here to tell you that even -6, according to your broken, lying-to-you DAW meters, is not only too high, it's a lot higher than -6. those peaks are really at +12 or so. you just don't know it.
Click to expand...

You ain't kidding.

2^24 = 16.78 million amplitude levels per sample

Now consider what happens when you inevitably mix down to 16-bit LPCM for CD mastering:

2^16 = 65,536 amplitude levels per sample

The difference in dynamic range from 24-bit to 16-bit LPCM is ~50dB. That's tremendous. If you mix for 24-bit dynamic range, you're going to have a hell of a time mastering to 16-bit... Unless you release all your material on 24-bit DVD Audio.

EDIT: By the way, theatrical Dolby Digital mixes (a format which has ~103dB of dynamic range) are recommended to be maintained at an A-weighted average loudness, Leq(A), level of -27 dBFS. In the home theater systems they use some other tricks like Dynamic Range Compression and dialogue normalization to balance out the mix in playback, but when the source material is mixed to -27 dBFS, you can actually enjoy the entire mix... cheesy dialogue, explosions and all, without having your ears assaulted with distorted crap.

I've beta tested AudioLeak from ChannlD, the production release of which can be downloaded here and it seems to be very reliable for calculating Leq(A) of a given source track.
 
Tarkovsky said:
That's really news to me. Is there any easy way to downmix everything to 16 bit without bouncing the whole thing in PTLE then?
Click to expand...

No. Nor do you want to. A bounce is a process of resampling rather than transcoding... it's a lot more accurate because it gives the sample & hold buffer much more time to work the conversion of each sample accurately.

For this reason, the sample & hold function is the most critical component of a digital-to-digital or digital-to-analog conversion/converter.
 
You must log in or register to reply here.
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.
Top Bottom
Back