I know I know, there he goes again flooding the forum with ignorant questions again. SO WHAT! There are a lot of smart experience people who can give me pointer on this: My show songs are completed. Now to burn them all to CD's and get them ready for opening night. Oh brother let's hope so.
Question - How do I normalize each musical number so that their overall levels are near the same? Something I've repeatedly have had trouble with. One song tend to be quieter than the other and visaversa. Or is that, visaposta? Or, visacopa? Never mind...
Any tips? Sirloin tips maybe? How about a hamburger on whole wheat bread and hold the Mayo in the palm of your hand? eeeK!
What software do you use? Some sequencer programs have a normalizing feature in them. On a Mac DP does. You might want to check the one you are using to see if it has this. Also I use Toast and Jam for burning my CDs Jam does a capable job of normalizing. IÕm sure there are plugs, probably some free, for the PC that do it.
By the way, got your pm. Mail comes about 3:00. Will let you know.
Let me tell you my experience from theatre-man to theatre-man . Normalize in level the loudest cue, that is, make the top level reach the 0 dB mark (you may also want to leave 2 or 3 dB of headroom). And then, you have no option but to work at hand the relative levels of the other cues if you want a good job. Almost every program you can use to burn your CD should have level control for each cue (even Nero, I guess).
Just jump form one cue to another at random, and check if volume is consistent. Keep doing this untill you are happy with all the cues. One evening spent in doing this boring thing yields to a really polished master!
All I'll say is that I'd stay away from normalizing software. They produce a harsh sound overall and will bring everything up and cap the top end and you'll end up having a song that sounds static. Use the aforementioned method of bringing the volume up. Just use volume control on the master output channel and try to keep the peak of the song around 0 db. Do this for all of the songs and you should have a pretty relative volume level for all the tracks. Don't be alarmed if when doing this you have small peaks up into the 2 and 3 db range, that can't be helped.
Assuming your software can support "peak" and "RMS" normalisation, and can calculate statistics of audio, here's a good method. I use Sound Forge 7 to do this.
Starting with what should be the loudest track, normalise it using peak mode (so the loudest peak is 0db or close to it). Then use the statistics to find out the average RMS power of the track.
For any tracks that should be equivalently loud, normalise them in RMS mode (not peak mode) using the average RMS power of the loudest track as the peak RMS level. Any tracks that should be quieter, use a lower RMS level in the normalisation - you can experiment "by ear" to find the appropriate levels.
One piece of advice - once you've normalised a track, save it to a new soundfile. Keep the original un-normalised, so that you can go back and make changes if you want to without having to re-render the audio. (hard-won experience talking!)
Hi Styxx, I totally agree with Joseph about staying away from normalizing features. Here is another idea, I am assuming that you use some type of DAW with fader (volume) automation. Start a new project and import or drag, all of your songs, one after the other in the order that you want into one stereo audio track. This way you can quickly access any portion of any song. If you find that a particular song starts way too quiet, correct it with the fader volume and record that automation, keeping an eye on the master levels so that they never go too high. Once you have the levels of all the songs sounding the way you like, you can then devide the stereo track to separate the individual songs, retitle them, and then make your final CD from those soundbites or tracks.
Just an idea,
When you (Joseph, Gwydi, DPDAN) refer to "normalizing" software, to what are you referring? Forgive me for not know much about this, but are you referring to software like Peak? Or plug-ins within, say Logic or DP, like a limiter? Could you please offer more info on why?
Whether a plugin or a separate program, the normalizing feature tries it's best to keep the volume consistant to a specified limit. Usually this normalizing feature takes away the dynamics in places where it is not desirable. Even though the system is able to look at peaks and dips in the volume, it is not able to control the volume smoothly and musically like we can, especially if we are familiar with the music content. That's all
"I think you're confusing normalisation with compression, which can behave as you describe."
You're right, I was mistaken, in Digital Performer, normalizing takes the file and increases the loudest peak to reach "0" db, my mistake. I have never used it because I was under the impression that it did not affect the entire track. My boo boo sorry.