good ol' loudness...

andrushkiwt

andrushkiwt

Well-known member
hey guys, question here on achieving a key point of mastering...loudness.

Aside from the usual "add a limiter" response, what else do you guys do to add some loudness to your track? for example, i usually use a tape saturation plug-in (though very, very minimally) as it has the added benefit of increasing the volume by 0.5 - 1db (i can select "hi" "low" or "neutral" on this plug-in, if i want to avoid affecting a particular end of the spectrum). after all my EQ work (i plug directly into an interface and overlay 10 guitar tracks, so it's difficult getting the sound just right going in) i usually lose a lot of volume on the track. i go into the mastering session with my peak at about -5db, do EQ work, and then find I have to set my limiter to -0.04 threshold and +6 or 7 input to get to an average RMS (K14) of -9ish.

It feels like simply smashing the input level up on the limiter is squashing my tracks. The obvious piece here is the mastering fader. Do you add db's on this fader? If so, what type of guidelines do you go by when adding volume via the mastering fader? any at all? I don't use compression in mastering, typically. I let the limiter do any squashing for now.

of course, i have only a years experience here, so feel free to tell me anything I may be doing wrong. i can already hear "do more of the EQ work in the mix". ;)

i want to add a completely different question at the end here. nothing to do with the above: hypothetically, does mastering itself make the file more clear and/or of higher quality? Without adding any inserts, does simply hitting "make image" (my DAW's final step in mastering) make the file better off than the original wav that went in? again, no EQ, no anything...just hitting the button.

thnx
 
Last edited:
I'd like to know the answer too. I am very new to digital recording and have barely cracked the surface as regards mixing and mastering. I'm finding it difficult to achieve much loudness on the master without excessive squashing. Maybe somebody can tell me what I'm doing wrong. Working in Reaper. A lot of compression goes into the mixing stage but the results have been sounding okay to my ears. Mixing completed, I render the results to a wave file at a conservative volume level (-12 dB, maybe), then import that into a fresh instance of Reaper on a stereo track. I put an EQ on the stereo track and a compressor on the master. Then the game is to try to achieve the right balance of dynamics and loudness by balancing the track fader and master fader. More track fader gives me loudness but drives the compressor on the master bus harder. More master gives me loudness without adding compression. I haven't tried a limiter yet. Thoughts?
 
I don't think you should be adding gain to your master fader - it sounds like bad gain staging to me. Add gain earlier on (i.e. channel/track/clip gain) and leave the master fader where it is if you need more gain.

@Robus - if you haven't tried a limiter, try one. It might/might not do what you want and it's only a few clicks away. If you use it as the last stage in the mastering chain you can prevent clipping on your stereo output and maybe add some percieved loudness if you're gentle.

i want to add a completely different question at the end here. nothing to do with the above: hypothetically, does mastering itself make the file more clear and/or of higher quality? Without adding any inserts, does simply hitting "make image" (my DAW's final step in mastering) make the file better off than the original wav that went in? again, no EQ, no anything...just hitting the button.
Click to expand...

No. If you arent doing any processing then you are just reformatting the file without altering the content. If only it was so simple :-(
 
Jake_JW said:
I don't think you should be adding gain to your master fader - it sounds like bad gain staging to me. Add gain earlier on (i.e. channel/track/clip gain) and leave the master fader where it is if you need more gain
(
Click to expand...

i don't add any volume to master fader...though i was thinking about it. thanks for clearing that up. perhaps my tracks are loud enough...but of course, in comparison to popular tracks, they're short of volume. maybe what i'm getting is decent enough for internet radio...? does it sound like i'm squashing too much by those limiter settings?
 
Jake_JW said:
No. If you arent doing any processing then you are just reformatting the file without altering the content. If only it was so simple :-(
Click to expand...
There may be a bit depth change, and that usually involves dither, and those both do alter the content, but not technically for the better. It adds a tiny bit of noise.

If there is also a sample rate change, then there will be a pretty significant change in the data, and depending on the algorithm somewhere between a teeny and a tiny bit of added artifacts and spectral change.

Edit to add - I have a "secret trick" for final leveling that has been working well for me and for number of folks over at the Reaper forum who tried it and became addicted.

Insert ReaComp on the master bus. Set attack and release to 0. Set the pre-comp to maximum (250ms) and set the RMS time to double that (500ms). Set the ratio to 1.1:1, and just turn the knee all the way up. Now push play and bring the threshold down to just about where the meters hit during an average level section of the track. I usually click the auto gain makeup on. Then listen to the whole thing and watch your meters and be amazed! ;)

This is one of the few times when I will advise specific numbers. You don't have to stick to them, but this effect is so subtle that it's almost impossible to "turn the knobs til it sounds good". Like, it kind of always sounds good until you hit that point on the knob where it sounds bad, but it's a sneaky kind of bad that sounds good until you step back for a minute. :/

You can try different times. Maybe do the math to figure out ms per beat or whatever, but I haven't seen where it matters, and a few people are complaining that the pre-comp's maximum is too low.

You can obviously experiment with the threshold and knee also. I like to have it doing a tiny bit most of the time but almost never actually applying all of its gain reduction. If you set the knee to 0, there should almost never be anything on the red GR meter. For whatever reason, I prefer to adjust the volume into this compressor rather than change the threshold. With the gain compensation, it's kind of a moot point.

The ratio is the one that will get you! This really is what's determining your crest factor, but a little bit goes a long way, and we're still not worried about the few stray peaks that might just be that much louder than everything else.

Most often all I need after that is a soft clipper to chop off those really loud "abberant peaks". I use my own JS plugin mostly because I wrote it, but ReaComp can do that too. Turn all of the time constants to 0, and ratio to infinity. Adjust the threshold to wherever you want the maximum level to be, and use the output gain on the previous ReaComp to drive your signal into it till you get what you want. Adjust the knee to taste, but this one usually doesn't want to be that deep. 6db or so max.

A slightly more "polite", but not quite as accurate way to catch those peaks would be to sort of combine both approaches, take the same settings from the clipper, but increase the pre-roll just a few ms, and set the RMS to double that. You'll probably have to increase the level going in just a bit, but you can get a little less distortion out of this.

When I'm feeling really whacky, I will chain three or four ReaComps in series with progressively smaller RMS/pre-comp times, higher thresholds and ratios, and narrower knees to really shape the overall envelope. It works quite well, but I'm not convinced it's really any better than the straight "level and chop" technique.
 
Last edited:
can you tell me, ashcat, is this close to what you described above...? i forgot to mention that i also apply a MBC to the master session. I set all bands to 1.1:1 with 0 attack and release, though each band has their own threshold levels. in fact, there are two thresholds for each on that MBC plug-in (a high threshold and low threshold, i believe.) anyways, i leave the thresholds at what the MBC has for its default. sometimes, i set the release much longer on certain bands. what do you think about using this technique but making the high and low bands 1.2:1 or 1.3:1?
 
I never use (need) multiband compression and have no opinion either way. I do quite often shelf down the low end a bit going into the long slow comp, and then boost it back on the other side, but this is much different.

Long attack and release are not the same as long RMS and pre-comp. They can have similar effects, but without pre-comp it is always too late to do any good. No attack and longish release is close, but can get uglier quicker.
 
andrushkiwt said:
can you tell me, ashcat, is this close to what you described above...? i forgot to mention that i also apply a MBC to the master session. I set all bands to 1.1:1 with 0 attack and release, though each band has their own threshold levels. in fact, there are two thresholds for each on that MBC plug-in (a high threshold and low threshold, i believe.) anyways, i leave the thresholds at what the MBC has for its default. sometimes, i set the release much longer on certain bands. what do you think about using this technique but making the high and low bands 1.2:1 or 1.3:1?
Click to expand...

I used to do this kind of thing a lot - but over a few years I've come to the conclusion that anything I can do with a multiband compressor can be done better at the mixing stage (though, obviously it took me years to get to this point and it's probably not helpful for your current projects). They are great tools sometimes but I normally see them as something for audio restoration/rescue than a solution for adding loudness.

Ashcat's trick with Reacomp is really worth a look - I havent used it on a project for anyone else yet but I was impressed with the results. It sounded to me kind of similar to using some Vari-Mu modelled compressors I have, but without the colouration and with generally a more subtle effect.

If you're looking for some other avenues to explore, parallel (new york) compression might be worth a look but it's easy to overdo somtimes and make things sound messy (though in a different way to hitting a limiter hard).
 
ashcat_lt said:
Insert ReaComp on the master bus. Set attack and release to 0. Set the pre-comp to maximum (250ms) and set the RMS time to double that (500ms). Set the ratio to 1.1:1, and just turn the knee all the way up. Now push play and bring the threshold down to just about where the meters hit during an average level section of the track. I usually click the auto gain makeup on. Then listen to the whole thing and watch your meters and be amazed! ;)
Click to expand...

I'm not using Reaper. Aside from "pre-comp" and "RMS time", the settings are telling me this is a basic compressor? How can I manage this on another DAW?
 
A lot of the plugins that come with reaper are available as free vst's

REAPER | ReaPlugs

Most of them are really well thought out and are well worth a look - possibly some of the more full featured plugins that come stock with a DAWs. I use Cubase and use for ReaComp as a general purpose compressor reasonably often.
 
Jake_JW said:
A lot of the plugins that come with reaper are available as free vst's

REAPER | ReaPlugs

Most of them are really well thought out and are well worth a look - possibly some of the more full featured plugins that come stock with a DAWs. I use Cubase and use for ReaComp as a general purpose compressor reasonably often.
Click to expand...

sweet. i'll download a few when i get home. If you have time, what is RMS size and pre-comp? i'll google it if i don't hear back from ya. thanks a lot, i'm excited for new (and free) plugs!
 
The RMS size is the window the compressor samples to determine the amplitude of the waveform and wether it exceeds the threshold, as far as I understand it.

So an RMS Size of 0ms uses the instantaneous amplitude of the waveform to determine wether the compressor should apply gain reduction if the signal exceeds the threshold. Higher values will make the compressor use the average amplitude of the waveform over the given time period (in ms) - in this case if the average amplitude over the set time period exceeds the threshold, the compressor will act and apply gain reduction. It can smooth out the response of a compressor when adressing complex sources (like whole mixes) and I generally find it a useful thing to have on a compressor.

The same goes for most compressors with an RMS option/dial/switch, but ReaComp has a bit more control than most stock plugins (afaik), letting you set the time period in ms rather than using a dial.
 
so by setting it to 0, you are pretty much bypassing that feature and just letting it act constantly at the set ratio rather than by looking ahead and using averages for short periods of time?
 
Sort of. If all of the time constants are 0, then the amount of gain reduction will be exactly what the ratio says it should be for the given each sample's individual value. That is a waveshaper, and sounds like distortion.

Attack, Release, and RMS time all do a sort of averaging of the detector so that the gain reduction happening now is actually based on the average of a given number of samples in the past. With any of these greater than 0, the compressor never actually applies the gain reduction that the ratio would dictate for the given sample being processed (except maybe by accident), but instead is always heading towards where it thinks it should have been a couple ms ago.

Since we're always looking backward to know where we want to go, we will most definitely miss any sudden changes. We might try to correct for it a little late, but the damage is already done.

Pre-comp tells the detector circuit to look into the future before any of the averaging so that we can see the big changes coming and start heading there in time to be where we need to be when it gets here. When it's right in the middle of the RMS window, and a peak comes up, it will actually start turning down a little early so that you have the most attenuation you're going to get exactly when the peak hits.

With short windows, it will have less overshoots than a limiter without pre-comp/lookahead, and less distortion than a straight clipper.

With longer windows it will start to level a performance on a note-to-note basis so that loud hits are not quite so much louder than quiet hits, but the hits themselves maintain their own attack/decay/sustain proportions. I use it on vocals, basses that I haven't played (and this have dynamic issues ;) ), drum buses, anything that needs a little help in the consistency department or which just has too wide a dynamic range from soft hits to hard. On a mix, it will sort of make sure that your "big chorus" doesn't get quite so much bigger than the verse. In a way it forces everything to kind of gently (with low ratios) nudge everybody around just enough to fit into a bit tighter dynamic box as a whole. It very much is that whole "glue" thing that people talk about.

It feels like cheating, and is kind of addictive, but it's been working for me.
 
By setting a value of 0 for RMS size you are telling the compressor to use the current peak volume of the audio as the trigger for compression rather than using an average over the specified time period.
So, basically yes? :D

Lookahead is a different thing (the pre-comp slider), but I get your meaning ;-)
 
thanks guys. i'm going to try this out tonight. i don't know if my noob ears will be able to tell the difference initially. i'll make an image file with and without it and compare in my car. thanks a lot
 
Ideally, you won't hear any difference, but your meters will show a tighter crest factor, and it will feel louder.
 
I didn't want to start a whole new thread for this one question, although it's not exactly related to the post...

i want to reference my collection of cd's in comparing my tracks. how can i actually get the tracks from the cd into my mastering session? i don't have itunes or another downloading program...can you just use good ol cd's somehow?
 
You must log in or register to reply here.
Back
Top