Need Help Understanding Recording Levels

J

JLLFAN

New member
That's for all that great advice, Slackmaster!

I'm using Acid ph1. I think it's an old precursor to all the Acid products that are out there now because I can't even find it on Sonic Foundry's web page.
The editor I'm using is Sound Forge XP 4.5. That's where the 'dynamics' effect is.

Acid ph1 is handy for its multitracking capabilities and the fact that it's got about 200 different drum loops to play around with, but I'm beginning to feel the pinch when it comes to mixing (and having to move back and forth between acid and sf xp. Do you have any experience with N-track? Is it a respectable mixing tool? Do you have to pay for all the good effects (I know you have to pay for the compressor plug-in)?

You mentioned the normalization and noise gates. How do these work? Do I apply the gate after normalizing, before, while recording? Help!!!! BTW, I'm using a Sound Blaster Live Value card and am currently recording Line-in, though I have tried micing it and loved the tone. (Not that this makes any difference, I just thought I'd mention it )

Thank you much!!
 
I'm currently recording guitar through the line in on my sound card. I'm just at a loss as to how loud I should be recording. The on-screen monitor bounces around the yellow and green, sometimes into the red.

I'm supposed to stay out of the red, right?

The monitor also seems to be keeping track of the highest peak hit during the session.

Is this the number I should watch? What should I keep it below? And what about when I run the "Dynamics Effect" (it seems a lot like compression)? Should I be bringing the volume down after doing the compression to avoid getting distortion/clipping (though I can't really hear any)?

Any help anyone can provide with answering any of these questions would be GREATLY appreciated! I'm just getting started in recording and am having a great time. I just want to get better at it. Thanks!!
 
Keeping the level as close to the red without going over is the general rule. The peak meter is useful to watch the overall level of what your recording, throughout the song.

Try it at different levels, if it peaks the odd time and you don't hear any distortion, leave it. I'm pretty suspicious of the level calibration between my soundcard (aark), Cubase, and Soundforge.
 
Thanks for the speedy reply.

But a couple of follow-up questions:

Does this general rule of "close to the red" apply before I start using EQ and compression? And should I ever EQ and compress the whole song, even after I've done it to the individual tracks?
IF the volume goes way up after compression, (bringing the graphical image up from -12 to -2) should I bring the volume down, normalize,etc (with the software, not with a slider)?

Thanks a lot for your help!!.
 
As a general rule of thumb, you want to keep your INPUT levels high to improve the signal to noise ratio (more signal, less noise).

I leave my Line In settings all the way up and control the actual recording level with my mixer (e.g. at the source). On the mixer I also like to try to get the signals as hot as possible and control the actual output to the computer. Of course this process will differ depending on equipment and what not. Just try to get as much signal onto tape (your computer) as possible with as little noise.

As far as staying out of the red. The red zone starts at 0db and you typically want your signal to get as close to this point without crossing the line too much. (One handy function you should look for in your software is "normalization"...it will normalize your track to 0db. However, this will boost your noise too unless you're using a gate.) Going over the red line is referred to as clipping. I've found that you can get away with clipping in very short little bursts without having any distortion. I usually crank on my guitar strings as loud as I can when I'm getting ready to record. Then I adjust the level until it only goes into the red in the extreme case. This typically gives me an average signal in the upper yellows which is good. The more the better. Basically you want to ride that red line.

There are several things can cause problems for you here. First of all your microphone might not be responsive enough. If you do have a good mic or aren't using a mic, then it might be your soundcard that isn't responsive enough. When this happens you'll get a signal that's way down in the green until you really lay into the strings, at which point the signal will jump up into the red. No amount of tweeking will help in this case. I tried using a cheap ES soundcard for a while and man, I was either in the red or barely getting anything. Sucked. You also might not be playing evenly enough...recording can really make faults in your playing stick out.

A compressor can be used to "crunch" your signal a bit. It can quiet loud sections and give a boost to quieter sections ....thereby 'compressing' the signal. So no, you shouldn't have to bring your volume down after applying compression, because that's the compressor's job! Learning to properly use a compressor is not an easy task...which is why I tend to stick with my compressor's presets :)

You said "dynamics effect"...sounds like you might be using Diamond Cut software? I have that...not too great but yes the "dynamics effect" is a compressor and noise gate. I have found Diamond Cut to be interesting, but real noisy for some reason.

What software are you using to record anyway?

Slackmaster 2000
 
In addition to the solid advice given to you by S2K, consider this advice given on another recording level topic forum on this BBS: there are great differences especially between level meters on software multitracks.

I think it was sonusman who posted a reply to that topic by saying that if you can't hear distortion, there <is> no distortion. This simplicity of this statement blew me away becuase I too was overly concerned with my meters getting into the red. With the software I use (Vegas), I have transient peaks consistently hitting up to 7 and 8 dB on the meter but guess what? No distortion.

I have to push my levels this high just to get a decent volume from my mixes. This way I can reduce normalization, thereby reducing unwanted noise. I seriously doubt there is any standard for calibration for the meters on software.

Anyways, take my advice with a grain of salt and use your ears to judge.
 
I appreciate that dmcsilva.

I'm concerned, however, that the distortion will appear as I slide the volume of the track up to the point it should be at for the mix. But I guess that I just have to play around with it to make sure that I only have to turn volumes DOWN--not Up! DOes that sound about right?

WAIT! I just realized that the tracks are getting recorded kind of quietly. The meters peak into the red zone, but the playback is kind of quiet.

If I boost the master volume for the song up so that it is constantly peaking in the red zone around 4.0, as long as I don't hear any distortion, I shouldn't worry about it? Is that right?!

Thanks again!

[This message has been edited by JLLFAN (edited 11-30-1999).]

[This message has been edited by JLLFAN (edited 11-30-1999).]
 
Yup, if you don't hear it, it isn't there is the rule I live by now. As for your concern about individual track volumes, I always adjust them first and then bring the master up to the desired volume.

dmc
 
JLLFAN:

I use n-Track. It's pretty good... though probably not the best. When you consider the price, however, it's unbelievable. I don't know exactly what you want to do, but yes, n-Track as a "built-in" mixer interface that gives you total control over the mix. From what I know, n-Track will do 99% of what the real expensive multitrackers will do...it just doesn't do them quite as well. :) Note that I am 100% satisfied and I'd recommend downloading the trial version.

n-Track comes with free reverb and a couple other lesser effects. However, it supports DX plugins so you can download other effects. I'd recommend Blueline effects. Fully useable shareware. Everything you'll need from compressor to chorus...

About normalization and gates. You don't have to normalize anything but your final mixdown, but I like to normalize each track because it's easier to control a mix by lowering levels than raising levels (IMHO). It doesn't matter when you apply the gate, although your threshold will rise after you normalize up (I'd assume). I don't really like noise gates because they kill some of the dynamics of the sound...and don't really eliminate much noise if your setup is ok. The trick here is to play around. I might of just had a bad experience :) First and foremost, try to get the sound as good as possible before you have to screw with anything. Don't "fix it in the mix" as they say.

And like dmcsilva said, do a search for posts around here by Sonusman. He gives some great advice and I've listened to some of his mixes...he knows his stuff. Plus he writes really long detailed messages when he's around.

Slackmaster 2000
 
Aw shucks Slackman, now you made me reply with nice talk like that about me..... :)

I have some tracks that I have recorded on my ADAT's that pretty much stay in the red the whole time. Now, please keep in mind that this was intentional and the results were intended to be something that worked very well with a lead vocal and lead guitar part for two different songs, so, I would never recommend this for general use. Anyway, as I was saying, the tracks are pretty much in the red the whole time (which on ADAT's means that you are either at 0db, or above, no way to tell because the meter only goes to 0db on most digital tape decks) and you can only hear a tiny bit of digital distortion on the tracks. Like I said, the results worked well for a special application but the point is that you really can get away with "recording in the red" if the track sounds good. Isn't that what recording is really about? Getting a sound? Who the hell cares what the manuals say!!! Or the lying cheating big time engineer/producer in Mix-o-sell-a-bunch-o-print-ads magazine who claims that all of his tracks get recorded clean and quiet. Dirt and grime is what pop music is made of!!!

Anyway, I will say it so that you can hear it from the horses mouth, if you can't hear the distortion, it really isn't there!!! If you think that by turning up the track in the mix is going to add distortion then just "solo" the track and turn it up loud just to make sure. You also have to weigh up whether the tiny bit of noise you might be hearing is really going to be heard in a mix with several other tracks. Often, background noises that are at -40 or lower just aren't going to contribute anything to the finale mix, so you need not worry about that kind of stuff.

Anyway, just turn it up and listen!!! That is all the average listener does..... :)

Ed Rei
Echo Star Studio www.echostarstudio.com
 
Well, since you asked.....

Mr. Sonusman has been doing a new day job (yes, even us "professional engineers" in Portland Or. have day jobs sometimes). I work for a company called MTI. We build switching equipment and displays for car, home stereo, and home theater rooms in stores.

Maybe you have visited a Curcuit City, or Good Guys, or The Wiz (for you east coast members), or an Ultimate store, or a Walmart. If you go to the car audio room, you will have a display where either you can push a button at a component, or, a central controller to pick which deck, speaker, and/or amp you would like to hear. Well, that is the stuff that I build now.

The job entails a bit of wiring, quality control, contruction, and some very basic electronic engineering. Kind of fun. You can visit the companies website at www.mti-interactive.com to see some of the stuff we do.

Anyway, that has kept me busy since this is the busy season for the company. With Christmas coming up, everyone wants there new display to sell product for the season. So, that means usually 6 10 hour days a week for me. I recently visited Long Island to install a system at the new The Wiz store in Bay Shore NY. 14 days straight with the average day being 13.5 hours!!!!

Sorry that I have been a bit sporatic with replies lately guys. I have also recently moved about 60 miles from where I was living, and in between moving, and trying to get the studio going again, things have just been hectic. But, life is starting to settle in again. And that means sonusman should be back to haunting this place on a somewhat regular basis again.

Rumor has is also that I will have a banner link to my website from this website ( www.homerecording.com ) with a feature article written by me of a topic of Dragon's choosing. I am really looking forward to this as one of my secret passions in life is to write and teach. Look for the link probably by the first of the year.

Anyway, keep on recording guys. My experience with this stuff has been that if you pay attention to what you are doing, read as much as you can about it, and think things through logically, you can really turn out recordings that start to rival the "pro's".

Ed Rei
Echo Star Studio www.echostarstudio.com

[This message has been edited by sonusman (edited 12-01-1999).]
 
Ed:

At least the day job is somewhat related to music. Plus it never hurts to gain a little electrical experience.

And I think that the article idea is excellent!

JLLFAN:

You wanted to know about normalization? Welp, I found a post from the writer of n-Track on this subject. It's a bit intimidating the first time you read through it (plus his english isn't great, I think he's in Italy), but it should help you out in the long run.

---BEGIN MESSAGE---
Every processing done on a digital signal adds noise, but the noise that the program adds when amplifying the volume of a single track is definitively not hearable. Also, the tracks are first converted to 32 bit, amplified and mixed together, and then converted back to 16 bits. The noise of converting back to 16 bit (C) is much more than the noise that results from amplifying a 32 bit signal (C32).
On the other hand, if you normalize each track, you amplify a 16 bit signal, leaving it to 16 bit format, thus you add a C noise for each track. Given N tracks, and assuming that you leave the volume sliders to exactly 0 dB, you end up with a noise amount equal to (N+1)*C.
If the sliders are not at 0 dB, it's (N+1)*C+N*C32
Avoiding normalizing you end up with an amount of N*C32+C.
Suppose you're working with 10 tracks, and that C32=1 and C=5: the noises become 55 for the normalization with 0 dB sliders, 65 with sliders volumes other than 0 dB and 15 for the case without normalization.
Assuming I didn't make any mistake, this is anyway a VERY rough calculation, but I hope it helps to explain the concept: the mixing process amplifies the tracks anyway, so it's better to not amplify (normalize) the tracks when not necessary.
BTW, this is basically an academic discussion, since unless you don't amplify a track tens of times, the noise that is being added is probably much less than the "built-in" noise generated by the soundcard when the track was recorded.

Flavio Antonioli.
----END MESSAGE----

Slackmaster 2000
 
Thanks Slackmaster! (And everybody else who's been writing) I appreciate all your help. And the benefit of your experience (which definitely saves me a lot of trial and error time).

Thanks again!!

[This message has been edited by JLLFAN (edited 12-03-1999).]
 
You must log in or register to reply here.
Back
Top