Home recorded Loudness vs CD Loudness

[nabskater]

Hey guys i was hoping sombody might be able to help me out with a question thats been sitting in the back of my mind for some time now. So when i record stuff (digitally) i try to get it as close to clipping without going over. The finished song usually has one or two places where it clips but isnt really audible, so i dont worry about it. I usually run the finished song thru a cmopressor and that helps even it out a lot, but what i was wondering is how are the songs that are professionally recorded on CD's and stuff so much louder than home recordings for the most part? Is this usually due to the difference in recording or what? Thanks

[pc]

Mastering baby. It's all mastering. Try this. After you mix your song, doing what you are already doing to avoid clipping while still getting the highest level possible, export it as a .wav file (or however your software does the mixdown process).

Now, once that's done, reimport that .wav file to a stereo track. You'll probably see that all of a sudden the level is fairly low by comparison--no longer up near 0 dB. You can then add some mastering plugins (or master on your own with additional eq, compression, etc.) and get the levels back up to 0 dB or slightly under and re-export. You should then hear something closer to the level of pro CDs.

Now... pro mastering engineers are doing this with true mastering gear, such as multiband compressors. By compressing certain frequency ranges and leaving others alone, the perceived "loudness" to the listener is greatly augmented.

[boobtube21]

I know they've been discussed here, but this reminds me again of the BBE sound maximizers. Are those still in use? I understand they do something similar to if not exactly what you're describing, pc.

I knew a guy in high school that bought one and it immensely improved the sound of his band's demo. And this was with a 'top of the line' cassette 4-track.

[pc]

Well, I'm in the camp with Fezz and Gravity Jim with those.

A BBE sonic maximizer isn't a multiband compressor as much as a specific-band delay box. In the 70s, that technology was called the "Aphex Aural Exciter," and it is exactly the same thing. It was the flavor of the month for way longer than a month. They (or their software counterparts) are still used today, but anyone with a decent ear can hear the difference between the Aphex/BBE stuff and true mastering with multiband compressors.

My opinion FWIW: Mixes done with Aphex do indeed make the mix pop up and grab you--no doubt about it. However, listen to three CDs in a row mixed like that, and you'll just feel tired and angry. Properly mastered CDs can be listened to all day. Again, that's just me.

[boobtube21]

That makes sense. Actually, I do remember reading in a magazine years ago something about "delaying the time it takes for certain frequencies to reach the ear", but I couldn't get my head around that at the time. I kinda-sorta can now if I really stretch my imagination, but it was probably a shaky description to begin with.

I know exactly what you mean about the ear fatigue/headache factor of some recordings, I've experienced it. I listen to a lot of independent-artist recordings which I'm sure don't always have the best mastering/production.

These multi-band compressors you speak of intrigue me. Do they come preset to filter out known-annoying frequencies, or is it something where a skilled engineer has to dial in specific freq's to be compressed for each particular recording? Is there a moderately-priced version (that still works well enough) that can be used with, uh, a reel-to-reel tape deck?

[Gravity Jim]

Fezz is exactly correct. The Waves L1 and L2 and TC FInalizer processes have created the Loudness Wars. I use an L1 on every master I deliver, but I never crank it to Rock Radio Ready levels. It's easy to forget that the person who is listening to your CD has a volume control on their stereo, usually parked at something like 1 or 2... they've got WAY more amplifier to use if they want you louder.

Multiband compressors are a great way to smooth, pop, "EQ," or otherwise polish a master. The plug ins do come with pre-sets. They are nothing more than starting points, and it takes years to get good at using one. I've been using the Waves C4 plug in for years, and I'm still not good at it. :)

[clayville]

Not that my stuff is worthy, but I've been digging the results (usually) from playing around with the PSP VintageWarmer plugin on my home recordings
http://www.pspaudioware.com/indexen.html

I have the early version, not the just-released VM2.

Its essentially a multiband compressor/limiter that seems to beef up and round out individual guitar tracks quite well, but also (using different presets and settings) can create the sort of "big sound" you're talking about when applied to a master track/late stage mix. I've never tried the (I guess standard) technique pc mentions of just applying an overall wash on an already bounced mix. I tend to do it all at once as above.

When you play as slow as I do, dynamics are your friend... but used in moderation I like the overall sound better when I kick this thing in. 'Course some presets can flatten it all to a uniform roar too.

Keeping in mind this is a 128K bitrate mp3, and that I didn't do any tweaking of the mix to create this example, here's a JamZone track from last summer with and without. I just plunked down a "light multiband compresser limiter" preset on the master track. One thing it did was exaggerate the reverb on the lead a bit too much.

WITHOUT PSP Vintage Warmer:
http://www.soundclick.com/bands/song...songID=4173840

WITH PSP Vintage Warmer:
http://www.soundclick.com/bands/song...songID=4913546

Probably proves the Pro's points, actually...

[refin]

Not that my stuff is worthy, but I've been digging the results (usually) from playing around with the PSP VintageWarmer plugin on my home recordings
http://www.pspaudioware.com/indexen.html

I have the early version, not the just-released VM2.

Its essentially a multiband compressor/limiter that seems to beef up and round out individual guitar tracks quite well, but also (using different presets and settings) can create the sort of "big sound" you're talking about when applied to a master track/late stage mix. I've never tried the (I guess standard) technique pc mentions of just applying an overall wash on an already bounced mix. I tend to do it all at once as above.

When you play as slow as I do, dynamics are your friend... but used in moderation I like the overall sound better when I kick this thing in. 'Course some presets can flatten it all to a uniform roar too.

Keeping in mind this is a 128K bitrate mp3, and that I didn't do any tweaking of the mix to create this example, here's a JamZone track from last summer with and without. I just plunked down a "light multiband compresser limiter" preset on the master track. One thing it did was exaggerate the reverb on the lead a bit too much.

WITHOUT PSP Vintage Warmer:
http://www.soundclick.com/bands/song...songID=4173840

WITH PSP Vintage Warmer:
http://www.soundclick.com/bands/song...songID=4913546

Probably proves the Pro's points, actually...

Sounds good Clay,I hear a difference.

[pc]

It all has to do with how we perceive sound. Limiting peaks overall and then raising the level as high as it will go usually results in a louder but mushy mix. Compressing the peaks/lows on certain midrange frequency ranges while leaving the peaks/lows on the high and low frequency ranges can allow you to get a mix to sound even louder but still sound good. (There are also reasons to do it the opposite way).

Jim's right though, it is an art unto itself and even with a really high-end plugin like the Waves stuff, you really have to know what you're doing to see good results.

Here's the other thing though, just like I personally believe AAE mixes had their hayday, multiband compression on mixes is becoming tougher to do well also. Why? Because in the 70s & 80s, we were listening to rock albums on vinyl or cassette, on our home stereos or on the radio, and as Jim suggests, it was not at all uncommon for singles sent to radio to be mixed differently just for that medium.

As the 80s went on, boom boxes came into high fashion, soon to followed by walkmen. Now iTunes-type downloads are hugely prevalent, with people listening to them on iPod headphones or through computer speakers. The Mastering Engineer, trying to mix for the most common medium of play ("where will this be listened to?") has a lot more really disparate possibilities.

The key to good engineering is to get the best sound out of your lousiest monitors so that all other speakers that play the music will sound equally good or better. Of course you cross reference to far-field monitors so you know you're not setting your low-frequency instruments too high, but if your mix sounds good on Auratones, it's gonna kill on good speakers.

By contrast, the key to mastering has always been to take that mix and make it KILLER for the radio, or on someone's home hi-fi, etc. There is definitely a certain way to master for those two media, and I think the increased number of common media has changed that process significantly.

My point being, big-name mastering houses are still the common thing for commercial CDs from well known artists, but don't take anything I've said to mean you can't make a good mix/master with really well-though EQ and the right amount of "normal" compression and/or limiting.

[nabskater]

Yeah, i dig what you guys are saying. But when i reimport the file, its barely touching the clip line on some of the peaks, jsut where i had it. The process i usually go thru is to mix thie thing, not really worry about clipping, then once the whole song if mixed and sounds good i export if then compress the whole thing and save it seperatley, so that way its pretty much the same loudness throughout the whole thing. Theres still dynamics, but its just more constant.

[Sage]

nabskater, I might be able to shed a little light...

Most professionally recorded CDs have a much more thorough compression job done on them. It sounds like you're running the entire finished song through a compressor. A mastering engineer will usually get much more detailed with the song, increasing and decreasing the compression levels, thresholds, etc. much the same way you might set volume on an individual instrument's track. If you're familiar with automation curves, this goes a long way towards explaining it.

A mastering engineer will also apply equalization to the final mix, and will do so in such a way as to complement the compression. Again, he will change the equalization at different points in the song, either to accentuate certain parts or to even things out.

Others have pointed out that too much compression kills the dynamics and causes listener fatigue, and that too often modern recordings sacrifice sound quality for loudness. I agree, but this is irrelevant to your question. The answer is that you just need a more thorough compression job. You may also want to look at applying a limiter in order to get rid of a few standout peaks.

This is a more complex task than applying compression to a single instrument, because your final mix is a more complex sound file. It takes patience and focus, but the return on the investment is worth it.

[clayville]

Another thing, I never let the software "export" my mixes. I send them from the computer to my board then back into the computer to record on two tracks. That way what I hear is what I get. When I let the software mix to a stereo wave file, the new file is rarely the same as what I hear.

I might be being a little thick here, but I'm having a conceptual breakdown on this (the perpetual danger of the self-taught). Can you elaborate a little on the signal chain here? What's doing the recording of the signal back from the board? Inserted confusions below...

I send themWHAT? THE MIX YOU'RE LISTENING TO WITHIN THE SOFTWARE OF YOUR CHOICE? from the SOFTWARE ON MYcomputer ??to my board SENT AS JUST A L-R AUDIO OUT?then back into the computer to record VIA DIFFERENT SOFTWARE?on two tracks.

[Gravity Jim]

You can play two tracks from your software and record those same two tracks back into the same software at the same time, clay. I do the same thing fezz describes for multi-track mixdown and for stereo mastering (unless I'm shooting roughs to a client, in which case I will "bounce" the stereo mix, with its processing, to the hard drive as an MP3 or AIFF just to save time).

[clayville]

Oy.... as you know, there's a wealth and world of stuff the software can do that I've never tried, terms and techniques that make my pea brain melt. But just so I'm sure I understand in my l'il ProToolsLE world...

Are you talking about a submix sort of technique?
Setting up Sends from all (or some) of the tracks to, say Bus 7-8 and then opening a new track whose input is that bus? And then hitting Record? And using that track as your master?



I'm shuffling through my PT books trying to figure it out...

[Gravity Jim]

Well, fezz is talking about something like this:

1. Send the output of a stereo track in your software to, let's say, "Out 1-2" or whatever you call it, so that it shows up on channels 1 and 2 or your mixer (that is, actual audio playback).

2. Set up your plug-ins and do whatever processing you're gonna do, listening to channels 1-2.

3. Then, when you're ready to mix down, instead of doing it in software as a bounce or routing the audio internally in the computer, you would send the output from mixer channels 1 and two, to, let's say input buss 3 and 4 on your computer.

4. Make a new stereo track, using 3 and 4 as inputs, and arm those tracks to record.

Now, when you hit record you will be playing audio out channels 1 and 2 and re-recording them on a new stereo track on input buss 3 and 4.

In this way, you are making a a new recording of exactly what you hear in your mixer, and not letting the computer create a new mix that you actually have not heard. Your final mixes will sound more like what you mixed when you do this.

[clayville]

....woah!....




thanks for the clarification!

[Gravity Jim]

So, I was right?

The whole time I'm typing, I'm thinking, "maybe he meant something else!"

[clayville]

Awww man! I just tried it on an old file, routing and re-routing various buses, sends and inputs in ProTools (shaky ground for me anyway, so good practice), bouncing out to wav (and re-bouncing the original version out to wav instead of mp3). My ears are listening back and forth through the monitors.... then the cans... then comparing back to the original non-bounced file... and I was sitting here developing even higher esteem for the spectacular and discerning ears you guys must have to be able to hear any difference!



But thanks, truly, for all the lessons you guys impart -- even when MMMV!

[pc]

The Pro-Tools bounce plays the file and records it, so if you're using Pro-Tools, you're already doing this. If you hear the file being play during the bounce, it's real time.

I was going to say this, just to say, you know, that ProTools is pretty kewl. But you beat me to it. Clay--if you use BTD on ProTools, you don't need to do what they just described--it's already happening.

[Rickenjangle]

A friend of mine says he does submixes of, say, the bass and drums, and then runs the new stereo track through a compression algorithm. He says this helps him keep the levels and S/N ratios in check, and enables him to treat the tracks appropriately for the program material.

This sounds similar to the concept of multiband compression except that he's doing it step by step.

I've had good luck with Sound Forge's Wave Hammer algorithm. I always run a preview, and make adjustments before processing. It's so easy, a dummy like me can make it work. I get pretty close to Pro volume levels that way.

By the way, I had to edit out a lot of technical jargon - for some reason, this forum doesn't like certain keystroke combinations. If anyone is interested in exactly how I use SF, just email me.