Apparent lack of overall volume on final mix

A lot of the "loudness" we perceive in moden mixes is due to the use of compression, effectively reducing dynamic range and raising the RMS value (in audio, this is the average "power" or volume level).  I don't know your background or experience level, so I don't want to assume too much.  If you haven't done so already, you may want to do some searches and some reading on compression and the "loudness wars".  There's a lot of history here and different philosophies about what the "right" level of compression should be for each mix.  There are a lot of folks on this forum who can give you very good advice on addressing this issue.

My only advice here would be to avoid extremes (no compression or high compression). 

Good luck.

Dont talk to me about this, only in the last year have I managed to bring my levels up to somewhere near competitive, I still cant quite manage it yet.

I think I would be sure in saying that, compression is a MUST if you are to reach the levels your referring too and that probably means all along tracking aswell.

If you are doing that and your levels are on the verge of clipping then you should be getting something decent out of X1 unless some setting on the export is reducing it?!?!? make sure everything is directed to master? export as wav? if you could get a plug in (ozone for example) to gve you the rms value of your mix you would know wether it is loud in X1 for a start.

Cian

Take a look at the waveform of your mix.  If you see transient spikes sticking way out here and there, you need to tame some of your dynamics.  This comes down to determining which tracks/instruments are the offenders and applying some compression on them.   Over-do it though, and you will suck the life out of your mix.  Once the transients are under control, run the mix through a limiter with (probably) a max of 4 to 6dB gain reduction and you should be getting near the levels of pro mixes.  You'll never get it as loud as the latest Top 4o hit (without killing your mix) but you can get close. 

volume, for the sake of final levels to compete with pro releases, is all about peak control.

you do this with:

1. a good balanced mix, that has control of it's peaks without ruining dynamics.
2. compression.
3. brickwall limiting.



numbers 2 and 3 really only work if you KNOW what you're doing!
heheh

not to flippant here, but i hear more ruined music by home mastering, than i hear really good home mastering.


one reason is knowledge and skill.
another reason, is most mastering engineers use really high end gear to get their results.

Having a compressor is one thing... using it is a totally different level...... using it correctly is yet another level.

UNderstanding what it is doing and working to get your mastering chops up will take time.

I've been at it for +3 years.

I have a lot yet to learn but the difference in the mixes over those years are noticeable.

One thing I do is run Ozone in the final master..... Ozone is good but again, it takes time to get the fine points to move the mix to the next level..... I also use Normalize after I export to bring the highest peak up to 0db.... My music is now at what is considered broadcast quality and levels.  But it took me almost 4 years to get here and that was a lot of time recording music.

 the mastering process starts right at the beginning with the tracks and when you realize that, and start with that in mind, you will see a big jump in quality.  When you actually get to the "mastering stage" very little should need to be done. Just a bit of polish and shine.....

There is RMS value and Peak value. Read up on these. The loud mixes are getting very high RMS value and using parallel processing to layer some transients in on top. In some cases more than parallel processing to achieve the desired goal. This is all done with various stages of transparent compression or the occasional colored compression depending on the goal. The level of RMS against transient peak, without totally trashing the sound, is the goal for loud mixes.  

EQ can be used to tame excessive peaks and is safer sometimes than limiting. Pulling the excessive peaks down with EQ and then coming back with a brick wall limiter can preserve tone. Generally, compress to create a stable RMS sound or value then use make up gain to get it louder. Use another copy of the track and EQ the peaks down then limit it. Layer these two tracks for the final result. Simplistic but hopefully this gets you going.

I think some of the problem lies in that a lot of the song's volume is in the lower frequencies. This will tend to muddy things up. Below is a frequency analysis I did of the song with Adobe Audition. Your loudest sound is just above 100 Hz. You also have a lot going on below 40 Hz.

If it were me I would start with rolling off the frequencies of anything that wasn't a bass guitar or kick drum at around 100 Hz. Compression may make things worse if it ends up increasing the volume of the lower frequencies.

  Thanks for all of your help! Chappel,the waveform thing is very helpful. Thanks for taking the time to send that! I wonder if the voxengo is better than the cakewalk one.

Bitflipper- I haven't compared my material to a commercial mix yet other than subjectively hearing mine compared to theirs. Mine sounds less loud and theirs sounds better musically lol. I did use the Cakewalk Analyzer though.I am shooting for a volume that when the track is first heard,it is not necessary to turn up the volume on the average computer to hear the track.

Guitarhacker- I have been dabbeling for longer than you have, but thats about the extent of it. I'm trying to seriously do this and I know that will take some trial and error :) I have seriously considered Ozone...If only I hadn't re-ordered that variax.

Batsbrew- I can't afford the high end stuff and I don't entirely know what I'm doing. I am having a lot of fun trying to learn though.I think its a dangerous place to be in when an individual thinks they know everything about anything .  A man who admits he doesn't know what he is doing is far batter than a man who tries to cover up his lack of skill and BS everybody into thinking he does.A man that thinks he knows what he is doing either better be right or have a friend nice enough to tell him the truth. Agree?
Dave- I think you hit right on the problem.

Bubba- Loudness wars.I understand. I don't like competition but I want to be heard.Something I'll look into.

 This whole dilemma started as the result of me viewing the graphic meters in Sonar X-1 and in seeing those meters reading in the red or over the 0db threshold. 

 In order to get the levels I needed it appeared that I had to lower volumes to acceptable levels but still retain the percieved volume. The only method I knew to achieve this result was to compress on those tracks that were still low in volume after I had driven them up as far as I dared and still remain safe.

 So this is what I did. The drums were compressed slightly. The guitars were also compressed and I added a brickwall limiter and a compressor in the main bus. First off I notched the EQ of each respective  instrument at its best place in the mix,next I added compression and lastly panned.

 I suspect that a common thing to do is to add several layers of compression to address both peaks  to increase volume . I probably made a very common mistake. I added far too much compression which squeezed the life right out of the mix. You could not hear the EQ or discern between parts as well  for this reason.

 I deleted all channel compressors with the exception of the one in the main bus. The mix immediately opened up and got preceptibly louder. I used the Cakewalk analyzer to look at peaks in different frequencies in real time and to graph those peaks.

  Using the Cakewalk analyzer I determined that the channel meters in Sonar X-1 are not accurate. In some cases I was able to drive a channel up over 5db in the red and still only show just under 0db on the analyzer. The two don't agree with each other. My ears detected no noticable digital distortion, and in fact, the volume was at a level I had wanted all along but was afraid to go into the red on. Most of my material in the upper freq spectrum was coming in at -6/-10 db according to the analyzer. A smaller amount of material was touching the 0db mark but showing a redline on the Sonar meters.

 For purposes of loudness comparison ONLY I have posted a before and and after test.

 Before- http://soundcloud.com/starise/just-the-same

 After- http://soundcloud.com/starise/just-the-same-remix

 I am aware that the guitars are out of sync in some places. In fact the timing in general needs some work as does the drums,so don't be to tough on me lol! The lead guitar part is slightly distorted and this is because it went in too hot and is not the result of the levels settings.In short the mix isn't finished but this will at least show you how the volume and some clarity came back with less compression and moving levels up slightly.

  The whole RMS and peak theory thing reminds me of the old days when I was in electronics and we measure things in RMS. I guess maybe I need to return to some of that theory again. Like I said before, the more I learn, the less I know.

 In the case of the guitars, the software programs GR4 and Amplitube 3 both have built in compression and I was already using that, so adding  more compression was overkill. 

 If I were to add compression on an individual track,from now on it will be done very judiciously and it won't be the first tool I reach for to get more volume. I think I would work the EQ first to open up the mix and give room to all of the parts. Next I would adjust levels using something like the analyzer instead of relying on the channel meters entirely. Only lastly would I think about compression and be very careful its not sucking the life from my mix.The thing I really like about the Cakewalk analyzer is that you can solo tracks and look at the results for individual tracks too. Great stuff.

 

 
 

 

Batsbrew- I can't afford the high end stuff and I don't entirely know what I'm doing. I am having a lot of fun trying to learn though.I think its a dangerous place to be in when an individual thinks they know everything about anything .  A man who admits he doesn't know what he is doing is far batter than a man who tries to cover up his lack of skill and BS everybody into thinking he does.A man that thinks he knows what he is doing either better be right or have a friend nice enough to tell him the truth. Agree? 

i totally agree with this.


that's why, typically, when folks ask this kind of question, and it's obvious they're just getting started, i usually suggest taking their project to a Pro Mastering facility.
that's what i did, with my first two projects.
then, i learned from it.




i've been learning all about mastering since about 2006, so by now, i've got a lot of miles under my belt.
it's not easy.
and a lot of times, whether you like it or not, it really matters with the quality of gear available to use.....
and that's my personal battle, learning my mix room well enough to overcome acoustics, gear, etc.




there are tons of 'plugs' out there, to try to make things easier..
but the bottom line is, it boils down to decision making.


it's easy to simply 'boost' levels..
and that's what i'm hearing mostly....
a lot of arbitrary boosting of levels, and the thing is, everything interacts with everything else.


every time you brickwall limit something, it takes the signal of one thing DOWN, and brings the signal of another thing UP, and suddenly, it's not your mix anymore.


the art is in finding the balance.


well, that, and having a stellar mix going in.

the more I learn, the less I know.

The definition of an expert is someone who knows enough to begin to grasp the depths of his own ignorance.

Unfortunately, the Dunning-Kruger Effect guarantees that no one can never truly comprehend just how deep that ignorance goes. So whatever we think, it's probably best to assume we're wrong, either in part or in whole.

Some instinctively recoil from this concept, terrified at just how enormous the blackness of the unknown is. But it's important to remember that D-K does not negate the value of acquired knowledge, just reminds us that it's incomplete and that we can never say "oh, I get it now" and stop there.

OK, back to topic. Here are some amplitude statistics for your two samples, which I have labeled "before" and "after":

Chorus:
                 BEFORE        AFTER
Max sample       32767         32767
Peak amplitude   0db           0db
Clipped samples  21            24470
Ave RMS Power    -8db          -3.6db (L), -6.5db (R)

Verse:

                 BEFORE        AFTER
Max sample       26028         32767
Peak amplitude   -1.4db        0db 
Clipped samples  0             19044
Ave RMS Power    -11db         -3.7db (L) -7.9db (R)

Entire song (less intro and fadeout):

                 BEFORE        AFTER
Max sample       32767         32767
Peak amplitude   0db           0db
Clipped samples  92            125787 (L) 49690 (R)
Ave RMS Power    -10db         -4db (L) -6db (R)


It would appear that you've done the song no favors in the remix, Starise. Subjectively, it sounds much worse to my ear than the original. Objectively, the statistic bear that out.

In the remix, you have severely overcompressed the song, removing all the dynamic range and reducing clarity and punch. There is no longer any dynamic difference between the verses and choruses, so the song loses excitement and becomes tedious to listen to.

You have also introduced actual defects in the form of digital overs - lots of them - that give the song a grainy sound. Commercial recordings generally do not have overs, even when severely compressed (Death Magnetic being a notable exception, which is why it's notable). They just have better gear than us, so it's best if we NOT try to get our peaks right up to 0db but leave a little headroom, at least 1db.

Spectrally, both versions are a bit ragged, with a slightly annoying peak at 3-4KHz in both. This may be the result of your speakers being a little weak in that range. In the remix version, there is a new, slightly less-annoying peak around 300Hz. The overcompression has actually reduced the low bass slightly, probably not what you intended.

Mix-wise, the biggest problem I see is a rather large disparity between left and right channels. Note the 4db difference during the verse. I'd suggest listening closely with headphones and work on the overall left-right balance.

So where does your loudness actually fall in the commercial-record continuum? For top-40 pop, your original version is a little quiet but your remix is over the top. If you're mastering for the teen market, just back off enough to get rid of the overs and you're there. But if you're after more of a classic rock sound, then the original mix is already hot enough and could even stand to be less-loud if you're after a higher-quality product that other musicians might appreciate.

Some advice: measure your room and your speakers and identify their strengths and weaknesses. The general rule is whatever frequencies your monitors are good at, those are where your mixes will be weak, and wherever your monitors are weak, that's where your mixes will be overcompensated. We all have less than ideal monitoring environments, which is why objective measurement is so important. Pro engineers will often tell you otherwise, and say "trust your ears". But they work with better speakers and better rooms than we do. Sadly, you cannot trust your ears.

Also, make sure you're monitoring at consistent levels. You should have two repeatable volume levels, one for mixing and one for mastering. Mix quiet, master (moderately) loud. The exact levels aren't as important as being consistent. Use an SPL meter to calibrate your speakers so you'll know they're at the same setting each time.

Chappel- I'm going to give the voxengo plug a shot. I like free :-)

 
Yes! Voxengo SPAN is a MUST!  I use it all the time.  It's great for locating problem frequencies, locating areas where there's an overload in a particular area of the frequency range and a look at the overall curve for the mix.  After a while, you kinda get to know what looks right and what doesn't.  Be sure to utilize the the CTRL button when scanning with your mouse to isolate and hear certain frequencies.

 I can be quite bullheaded sometimes and I have heard that in terms of compression less is more so I had to go and remove the compression in the second example and have Bitflipper explain to me why that was a no no and prove to me in chart form why this was the case..so I apologize for not getting it at first. Sometimes I need to read and re-read something for it to stick . This was the case with parallel compression. Middleman mentioned it early on but it didn't quite lodge where it should. Mattplaysguitar further emphasized the benefits of it and FINALLY it clicked. 
   I think now I can make some headway with this. Thanks again for your patience.

Dave King

Chappel- I'm going to give the voxengo plug a shot. I like free :-)

 
Yes! Voxengo SPAN is a MUST!  I use it all the time.  It's great for locating problem frequencies, locating areas where there's an overload in a particular area of the frequency range and a look at the overall curve for the mix.  After a while, you kinda get to know what looks right and what doesn't.  Be sure to utilize the the CTRL button when scanning with your mouse to isolate and hear certain frequencies.

Spot on guys!

I was listening to a piece that is partially mixed and 2 bass notes in particular kept on jumping out of the monitors. Sweeping an EQ around was of little benefit in finding the exact frequency, but SPAN homed in instantly. A quick notch at the fundamental and first couple of harmonics was enough to smooth out the entire bass track

Depends on the speakers, and how directional the tweeters are. It's a bit less critical with Genelecs and Mackies than it is with ADAMs, for example. I have the same speakers as you, and with them, the ribbon tweeter is a laser beam.

bitflipper


Depends on the speakers, and how directional the tweeters are. It's a bit less critical with Genelecs and Mackies than it is with ADAMs, for example. I have the same speakers as you, and with them, the ribbon tweeter is a laser beam.

Yes, spot on about the ADAMs Bit - move your head an inch or two and you're out of the sweet spot.