Increasing volume without causing clipping

[Prasa_U.]

It seems like everything I do is a third of the volume of the commercial stuff, as well as some remixes, I listen to. For example, this file:

http://www.omohide.com/qqqqq.mp3

Seems to be as loud as I can get it without the clipping red light going off in my sequencer. But it seems quiet compared to some other programs, and especially when I play it on my cellphone, compared to other stuff on there. So what is the secret here? Is it possible to increase the volume on this example? How?

[Lunahorum]

Louder sounds better. Unless the sound is so loud that there is no dynamic content. As an artist, it is at your discretion where you draw the line.

To make things louder without clipping the peaks, use a compressor with 0ms attack, a ratio of infinity:1 and a release time of whatever sounds good on your particular compressor (generally somewhere around 600ms. Set it around 20ms and hear what happens). A compressor with 0ms attack and a ratio of infinity:1 is also known as a limiter.

What the hell does that mean?

All a compressor does is change the volume or amplitude of the waveform.

It is very hard to hear compressors working so mess around with the extreme values of the compressor to get a feel for it.

Attack time - how fast the compressor lowers the volume once it goes over the threshold (or raises it if the ratio is less than 1:1)

Release time - how fast the compressor stops lowering the volume (ie it raises the volume) once it stops going over the threshold.

threshold - this is the point at which the compressor turns on (or partially turns on - see knee). Some compressors are activated by any frequency band amplitude going over the threshold, some are activated by the average amplitude of each frequency band going over the threshold, some are activated by any frequency band amplitude in a range of frequencies going over the threshold, some are activated by the average amplitude in a specified range of frequencies going over the threshold, and I am sure there are many other variations. The two latter designs I mentioned are known as multiband compressors. As you may have already guessed, EQ is useful before and after compression. The first EQ is used to change how the compressor reacts, and the second EQ is used normally, to shape the frequency spectrum.

knee - this defines the transition range going from 1:1 to your ratio starting at the threshold. I believe there is a good picture of it in zircon's guide.

ratio - pretty obvious

gain - also obvious

[Wiesty]

Make sure your original wave form is clean and dosent clip. Compressors work, but remember that they make the quiet stuff louder, and the louder stuff quieter, so use discretion when using compressors.

[DDRKirby(ISQ)]

turn your speakers up. =)

remember compressors are used to decrease dynamics, right? Use them accordingly.

[anosou]

use multiband compression and a limiter until your song is just a block instead of a waveform!

works for me ;3

[Tensei]

Yeah, what AnSo said. Chances are you got one little frequency peaking, so you can't boost it anymore before it starts clipping. I think it's better to go back to the mixing stage, find out which instrument/track is causing the peaking, and boost the volume on/eq the other ones accordingly.

[Fishy]

If you don't want to use compreesors I like to use the old cut and boost on whatever causing it. Cut frequencies until you find whats causing the clipping, then cut that chunk a bit and boost the whole thing by about the same. Use your ears to find the right ranges.

[Zephyr]

I'd recommend going back into the individual instruments and try to keep them all at about the same range of volume before they get mixed in, (also use compressors so that it's fairly consistent) then run a multiband compressor on the master track as well as some EQ, if you do that, then the output of your mix should be pretty steady and you could probably increase the overal volume quite a bit, but since it's still likely to go over a little bit every once in a while, use some sort of a limiter just in case.

Another tip, when trying to balance the instruments, quiet down the loud ones, don't make the quiet ones louder.

[Fray]

If you like how it sounds now but you just want it to sound louder, you want to use a limiter -- that will hopefully remove the peaks more transparently and allow you to turn the volume up more. If some parts are too damn quiet and others are too loud, use compression. No, there's not much difference between a limiter and a compressor with a really fast attack and a 1:infinity ratio. A dedicated limiter probably uses some kind of lookahead algorithm whereas the compressor probably won't.

As others have said, it's better if you can apply the limiting to the instrument that's responsible for the peaking. Especially if you don't have $2000 mastering plugins that are good at doing this stuff transparently!

Louder sounds better.

No. No no no no no no no no.

NO!

GRGRHAIFJWJAF!!!

Louder gets more attention, maybe. And it keeps your mix at an even level with the other stuff in a random playlist, but that's only because commercial recordings try to blare each other out to hide the fact that they have very little worthwhile content. Otherwise everyone could just turn their speakers up a bit and get better quality sound at the same volume.

There is plenty of dynamic headroom in 16-bit audio.

It makes me mad because audio actually sounds worse and more fatiguing because of this goddamn loudness war.

[Cerrax]

Honestly I see both sides of the loudness war. When I'm in my car and I want to blast the stereo, quieter songs are disappointing because they don't have the same punch as the loud ones, but the reverse is also true. The FF7:AC soundtrack is ridiculously louder (much noticeably louder than even "professional" recordings) and it borders on painful, because I already have the volume up to a comfortable level for regular music and then "Chase on the Highway comes blasting through the speakers and shatters my ear drums.

Louder is not always better. Consistency is more important. If everything has -10 RMS, your songs should not dip below or go too far over that, because then it will not match the volumes of other songs.

[Fray]

Lol, we're talking about compressing the individual instruments and such... The mp3 in the OP is solo piano.

Prasa -- are you comparing this to other solo piano mixes? The piano is a very peaky instrument, and on a solo piece anything but a very small amount of compression or limiting is going to obviously color the tone. As a result most solo piano music, even pro stuff, tends to sound a lot quieter than other mixes.

Cerrax -- I agree. It would be nice if everybody tried to keep to a sane median volume. The problem is that the median volume of current commercial music is much too high. I also agree that you do have to make a compromise on the dynamic range for people trying to listen in their cars; if parts of the music are too quiet, you can't hear it above the road music at all. Makes it kinda hard to listen to classical recordings

[Lunahorum]

It is completely true that louder does sound better. Turning up the volume on the speaker is a better solution, but since most people keep their volume locked, the loudest song will sound the best. Just read some of dave moultons stuff on his site. He has some great psychoacoustic articles on there as well as other learning gems (golden ears, platinum playback, ect)

[dannthr]

Also, there's too much air in the recording that is needlessly adding to the overall volume.

[Fishy]

It is completely true that louder does sound better. Turning up the volume on the speaker is a better solution, but since most people keep their volume locked, the loudest song will sound the best. Just read some of dave moultons stuff on his site. He has some great psychoacoustic articles on there as well as other learning gems (golden ears, platinum playback, ect)

It is completely untrue in terms of the loudness war. When you turn up the stereo, it isn't the same as compressing songs to make them as loud as possible in comparison to other tracks at the same master volume. What is DOES do is make you more likely to pick it out of a crowd as a result, as the the overall volume stands out. The sound quality in general suffers from being pumped louder then it normally could be without clipping via compression.

There is louder in terms of turning your player up, which is good, and relative loudness in terms of compression, which is NOT good for dynamic music at all.

[Fray]

It is completely true that louder does sound better. Turning up the volume on the speaker is a better solution, but since most people keep their volume locked, the loudest song will sound the best. Just read some of dave moultons stuff on his site. He has some great psychoacoustic articles on there as well as other learning gems (golden ears, platinum playback, ect)

Ok that I can mostly live with -- that's the reason behind the whole loudness war. While louder music is going to sound more full and energetic, people do have a limit where they say "ouch, that's too loud". So they keep their volume knobs relatively low to avoid being blasted by the obnoxiously hot recordings people are putting out. The ultimate consequence, though, is that everything sounds worse, not better.

That is a pretty neat site, I didn't know about it. For the Google-challenged, I believe this is what he's talking about: http://www.moultonlabs.com

[SnappleMan]

People tend to confuse properly mixed for louder. When you have a song that's mixed properly and all the levels are good and everything is even and clear, just a little bit of limiting and compression will make it sound big and powerful without being tiring.

[Lunahorum]

Sometimes when I EQ the tracks, I end up thinking I made a lot of difference in a clearer mix, but after comparing to the previous version, it's only louder. So yes, louder does sound better.

[DZComposer]

Moderation. Always use compressors in moderation. The volume knob on your sound system goes beyond 4 for a reason.

Here are two waveforms of the same song. One was mastered in 1981, the other in 2005 (pics from Wikipedia, licensed GNU FDL).

1981:

Notice how the compressor was used in moderation. None of the peaks hit the clip lines. Also, you can clearly see the dynamics in the bulk of the waveform.

2005:

BRICK! BAD! UGH! It may be loud, but it sacrifices sound quality and dynamic range to be so. I mean, yuck. That's one reason why I don't like today's music. The dynamics are stunted. There isn't as much dynamic range. Also, note now that almost all all of the peaks are right at the clip lines. Not to mention, things like the bass drum and bass get their "umph" stolen by the compressor. Also notice how the main song loses it's impact from the intro. The original difference between the loudest peak of the intro and the main song beginning is about a 50 db difference. The brick is only around 20 db difference.

2005 is louder, but sounds like crap.

Use your compressor/limiter reasonably, but don't overdo it. If the waveform looks like a brick, you did it wrong. If you want to listen loud, turn up the volume on your speaker system.

Great article on the subject for further reading: http://www.prorec.com/Articles/tabid/109/EntryID/247/Default.aspx

[Zephyr]

While music changes with the times, I'd say that one standing principle for me is that I'll take a well mixed track over a really loud track everytime, if it's mixed properly, then a little compression goes a long way and you should be able to make it loud enough to hear everything properly while maintaining dynamics and quality.