normalizing and e.q.'s (dumb questions)

[B.SABBATH]

hi

I have a question.. Is it really necessary to normalize if you dont want the dynamics and life sucked out of a recording?

AND the second dumb question is I have a graphic e.q. with 7 bands of e.q. frequency lights for each side.. Should I look at the lights and mix so they light even? What I mean by that is maybe mixing more on sight than sound.. I see how my favorite sounding c.d.'s look as far as the lights on the e.q. goes, but when I tried to match that with my own recordings it sounded waaaay too crisp and harsh to me.. I saw the 16k lights barely registering so I boosted the hell out of the 16k and it sounded terrible... but some of my favortie sounding cd's register quite high in that range..

Any thoughts on these rediculous questions? I saw there is a free normalizing program offered on the front page by 'Audio Grabber' but I havent tried it yet.. I dont record with my computer..


I'm asking this in the newbie section to avoid being flamed.. but feel free to because I get a good laugh out of it sometimes..


thanks
sam


www.nowhereradio.com/samij/singles

 

[mixmkr]

sabb...here's my take on your question.

when you boost 16k, you may be boosting frequencies that you haven't really recorded, thus effecting your sounds in a bad way. Had you, [for example] recorded real cymbals, that frequency would most likely be present. But instead, you're boosting the upper end of some of your synth patches, maybe..... kapeesh?

also...I'd [just about always] mix with your ears, and leave the meters alone for the most part. Your listeners probably aren't staring at meters either. I've had mixes that seemed really lob-sided on the VU's, and yeah...it bugs me, but if it sounds allright, I usually let it go. Use them as a reference, like you appear to be doing, but I wouldn't worry about them in the end result [too] much.

Now...some 'real' pro is gonna get on here and flame me about clipped levels, crappy EQ balances you're gonna get, lousy mix levels, O dynamics.......

 

[B.SABBATH]

I've had mixes that seemed really lob-sided on the VU's, and yeah...it bugs me, but if it sounds allright, I usually let it go.

I'm glad it's not just me.. It's getting to the point where I'm getting so rediculous that as I'm listening to the mix, even if it's sounding good, I'm looking at the low registering 16k lights and thinking it's not as crisp as it should be.. I have this one CD that lights the 16k just as high as 36bass freq and it sounds great.. I recieved disasterous results when I tried to copy that e.q. look.. I hear what your saying that I might be boosting the 16k of a synth or something when I do that, so I guess I need to figure out where the friendly 16 is because I seem to be boosting the 16k noise.. hhhhhhhmmmmm..... thanks for the breakdown mix.. How do I find where the friendly 16k is? Most CD's sound too crisp when I play them through my monitors, (20/20/'s passive) but sound great when played on my regular listening speakers.. My stuff seems to sound best on the events.. I dont know..

what about normalizing, do you normalize?

thanks mix

 

[Guernica]

........not to distract from your original question sami..... but exactly what the hell is normalizing anyway? I thought that normalizing was just a fancy word for dithering down to 16 bit from 24....... is this right? Im thinkin im way off........

 

[lyonhard]

also...I'd [just about always] mix with your ears, and leave the meters alone for the most part. Your listeners probably aren't staring at meters either.

Exellent advice mixmkr, VUs should be basically used for a reference for clipping and the like (Overdrive, etc.). Nothing beats a good set of ears when going for quality sound.

Also B.Sabb, what kind of monitors are you using? Different monitors handle different frequencies differently.

Lyon

 

[knightfly]

If you don't record with your computer, then since you're concerned with normalizing can I assume you're using a workstation?

Regardless, normalizing is almost never a good idea - you should boost levels during record so normalizing isn't necessary. Here is a link that will keep you reading for a while -

http://www.digido.com/

scroll down the right sidebar about 60% to "Level Practices in Digital Audio", for starters. Pay particular attention to the paragraph in the first article, titled The Myth of "Normalization" -

On the hi freq wierdness, mixmkr had some good thoughts - He's right, people don't LOOK at music, do they? Meters are handy for figuring out WHY you don't like a sound, and maybe WHAT you can do about it - they almost NEVER will tell you IF it will sound good, other than maxing record levels in the first place so you aren't adding any more noise to the mix than necessary.

Personally, if you can I would recommend ditching the graphic EQ entirely. They are famous for phase shifting, and not in a good way - this is especially true for boosts. One of the biggest "Newbie-NoNo's" is forgetting that EQ's can also be used to cut. Generally you are better off cutting freq's that aren't needed, rather than boosting freq's which are needed.

If you want an excellent book on the whole range of mixing questions, I would recommend Bobby Owsinski's book Mixing Engineer's Handbook - Amazon.com carries it if you can't find it locally. Hope we helped some... Steve

 

[B.SABBATH]

hey G,

normalizing is what a lot of people do to get the highest levels.. My understanding of it is it's almost like compression where it squashes the levels so they dont bounce around as much and you get get a louder overall mix on a cd.. I have heard things that were too normalized and the mix lost all of it's dynamics.. the soft parts were as loud as the heavy parts.. but I have also heard it used well.. When that song came on, it was loud and tight.. I'm not sure if I should be messing with it because my tunes have a lot of mood swings.. but maybe it works better on stuff that has the same dynamics all the way through.. maybe you should check out that link that knightfly dropped as well.. I havent checked it out yet so I might be wrong about everything I just said.



lyonhard - mainly I use event 20/20's passive but I check it out on three other sets of regular speakers as well.



knightfly - thanks for the info.. I record with a Tascam 788 portastudio (24 bit 44.1)...The only thing I use the graphic e.q. for is its lights. I never monitor the sound that goes through it.. I have cut frequencies before but I hear what your saying..I'm gonna check out that link.. thanks again.

 

[M.Brane]

I try not to normalize unless it's absolutely necessary.

When you normalize you are bringing the level of your highest peak up to 0db. This also raises your noise floor along with the program material so the only time I do it is if I just can't get the track loud enough in the mix without compressing the piss out of it.

I always try to track stuff as close as possible to what the actual level will be in the mix. Even though I'm working with 16bit converters I find this sounds better than pushing everything up just to get those upper bits then pushing it back down with compression or the faders in the mix.

YMMV

 

[Seanmorse79]

Normalizing is different than compression. Normalizing just brings the track's overall level up (or down) based on either peak or average levels - it doesn't do any squashing. It's similar to just bringing the fader up, but you're actually amplifying the recorded signal (and all of it's noise).

The way I look at normalizing is ...if you need to use it, then your levels weren't where they should have been when you recorded it in the first place. It's okay to use if it's only adjusting levels a slight bit and your signal was recorded nice & hot, but I used it as a crutch when I first started, and it made for some really crappy sounding recordings (because of how I tracked).

I don't use it anymore, and I don't miss it a bit

 

[Seanmorse79]

M.Brane, you beat me to it

 

[BrettB]

I only use it on my final mix, after I've used compression and all the AQ and stuff I needed. After all that is done, and maybe it should be a little bit harder, I normalize.

 

[Shailat]

Have any of you checked to see if a mix pre normalized and after sound the same?

Do yourselfs a favor and try once to normalize a mix and compare it to a non normalized mix. Set the levels about the same and check audio quality.

 

[Seanmorse79]

Do yourselfs a favor and try once to normalize a mix and compare it to a non normalized mix. Set the levels about the same and check audio quality.

I have ....what am I missing here?

 

[mixmkr]

yeah...Sabb, understand the difference between compression and normalizing.

Normalizing does have its uses...like bringing up non-critical .wav files for use on the web, etc. But, with good recording techniques, you'll find that you'll eventually do away with it.

In most cases, good use of compression should actually be inaudible. All it is, is a fancy volume control...as you don't wanna hear someones hand riding that fadar up and down really fast (like its pumping..eh? ) Compression can be used anytime in the recording chain, imo.

as far as finding that "sweet" 16kHz....you gotta record it first...right? Then it will be there to tweak. That's possibly the difference betweem the CD's your listening to as a reference and your stuff. Are you using a mic and signal chain that will capture 16kHz on a quality level? Possibly not, also. It may start rolling off big time before that, especially if your using stage mics, or other dynamic kinda mics, like the 57.
I agree on [NOT] running the final mixes thru normal, consumer, HiFi style EQ's...as usually doing more harm than good.

About your CD's sounding too crisp on your 20 20's but NOT on your stereo, leads me to believe your hifi speakers are lacking on the high end output, and/or possibly your 20 20's are in an environment where the high end is actually getting accentuated. But that appears to be the opposite of the norm, actually. I find your mixes to be pretty balanced, however...EQ wise. But on the other hand, I'd [kinda] want to hear you say that they are matching the commercial cd's on your recording system too. You're too strange dude

 

[B.SABBATH]

I'd [kinda] want to hear you say that they are matching the commercial cd's on your recording system too. You're too strange dude

I understand the strange dude part, but I dont understand the rest of that sentence.. Also, my recordings dont match the levels of the CD's I listen to.. I'm hearing you on the normalization.. I shouldnt have used the words 'squashed' I just ment less level movement.. I'm gonna have to mess around a bit.. I'm sure theres got to be some sweeter 16k on the electric drums, cymbals etc. Yeah the only thing I do with the graphic e.q. is stare at the pretty lights.. It's good for that and it looks cool.. that's the main reason why I leave it hooked up.. I figured maybe I could use it as a spectrum analyzer and trust it, but that's not the case.. thanks for the input mix.. I'm going to check out some things.. Always interested in your thoughts about the sound man...


sam

 

[TexRoadkill]

Yo, Blackie. I think you're gettting Normalization confused with Limiting. Normalization has no effect on the dynamics of a track. It raises all the levels by the same amount. Limiting is a more severe form of compression and will only raise the lower levels and if used with a heavy hand can really squash the sound.

If you have some real cymbals or something you can record to help give you more high end content that may make your music a little more crisp and professional sounding. It's hard to get a lot of high end sizzle out of guitar, keys and vox.

 

[Shailat]

I have ....what am I missing here?

In critical listening conditions together with a person who builds vintage pre's for a living, I tested files in Pro Tools using three different files.
1 - mix to analog board
2- mix in PT
3 - Normalized file

The normalized file sounded the worst by far. It was so clear and cut that even a deaf person could have heard it.

I posted the results at the time on a thread by Ethan Winer back a few months ago and I might still find them hanging around somewere although I doubt it.

The other engineer with me was willing to vouch for this as well and gave me his email just in case anybody wanted to ask him specificly (although his english is a bit weak).

 

[Seanmorse79]

Shailat, I wouldn't question your ears or your skills by any means, but I have done quite a bit of testing myself and found the results quite different. When I normalized a track that was already peaked at 0db, normalizing did absolutely nothing to the track.

Believe me, I'm up for learning how to correct something I'm doing wrong ...actually, I don't really normalize much anymore, but still, I'd like to know. How did the file sound "worse by far"? Was the file mixed out at a solid volume before it was normalized?

 

[Shailat]

I don't remember exactly the chain (it was half a year ago.....)

If I remember correctly all the files were compressed with a Drawmer and all the files that werent mixed to the HD were put on the HD through a digital connection. So basicly all the files when they got to the HD were at the same basic volume after compression. Then I Normalized one of them and it and it degraded the sound turning it into something grainy. The bass got thin and the freq content of the whole file had changed compared two the other two files.

I was actully working more on the internal summing of a digital system then the effect of normalizing.
Maybe in the near future I'll try some more tests with a deeper look at Normalizing, but I really don't use it ever.

 

[knightfly]

"When I normalized a track that was already peaked at 0db, normalizing did absolutely nothing to the track."

That's exactly what should have happened. Normalizing shifts the level of the entire track upwards until the highest peak is at the level you choose (if you have that option in your software) Sooo, if the file you normalized already had a peak (or more) at "0", then nothing would have been done to the file. If, however, you did that to a file whose highest peak was at -10, then everything, including noise level, would have been raised by 10 dB.

If we follow established practices for maximising EACH link of the recording process as analog DEMANDS, then we should first of all record at the hottest levels we can, consistent with not clipping. Doing this at the first stage will ensure a better sound from each track, with minimum tweaking afterward.

Every time you change ANYTHING in a WAV file, you are doing math changes with the computer/DAW. Things that will make these changes LESS destructive:

Record as hot as you can without "overs" - when you lower the level of a track in the mix, you're lowering the noise level also.

Use the highest bit rate/depth available to you, with rate preferable double the final product. (88.2kHz for CD's)

Use the highest internal bit depth available (32 bit floating point is common in better software)

Use as little EQ as you can and still get the sound you want. This is also a math function, and changes the files. If you have to EQ, normally cut is better than boost. Best methods are "cut narrow, boost wide" , also, if you have to boost and have multi-band overlapping EQ available, boost a little at two close freq's instead of a lot at one freq. This lessens the phasing damage of EQ boosts, probably moreso in analog than digital.

Work at the elevated bit rate/depth until the mix is done, then mix down and dither ONLY ONCE, to the rate/depth you need (16 bit, 44.1 kHz for CD, etc.)

All the above are aimed at doing the LEAST DAMAGE to the recorded tracks during mixing.

If you can't afford converters that cost several thousand dollars, high sample rates will usually sound better than 44.1, because of the expense of making quality "brick wall" filters used at 44k. This is still widely debated -

Getting back; besides raising the noise floor when normalizing low level tracks, the other main reason not to normalize is when putting together tracks to finalize an album. Perceived loudness has NO BEARING WHATEVER to peak loudness - This can only be determined with EARS, which is one of the reasons we pay mastering engineers to make our album sound "right" - Sooo, when you normalize each final mix and then put them together to make an album, the perceived loudness of each track will NOT be the same, guaranteed. If you've normalized all the songs, then the mastering engineer has to find the loudest SOUNDING track and adjust the others DOWN til they SOUND the same level, or in the case of a ballad, SOUND QUIETER than the rest of the (rockier) tracks. Leaving the mastering engineer some room to work without this hassle will get you a better sounding record.

Not outa info, but outa time for now - hope this shed some light on the normalization thang... Steve