Normalizing without the hiss and fuzziness

Discussion in 'Mastering' started by hxckid88, Nov 21, 2005.

  1. hxckid88

    hxckid88 Active Member

    When I record, I record the guitar where it is loud enough, but not too loud where it is fuzzy and bassy (and of course, I make sure it doesnt clip at all).

    Generally, the recordings are low, and I realize, I should try and normalize them (being a newb to recording and all). In CUBASE LE, there is a normalize option for the selectected tracks, so I selected the sections for track one, did normalize to maximum etc...

    And I did that with all tracks... It seemed fine except a little bit of clipping occured, is there like a general rule when normalizing a mix? Notice how I said I did each individual track up to the same level (set to max). Should I normalize, and compress? And normalize again? I always hear of that technique to get the "togetherness" and "tightness" sound, but I don't know how to go about doing it correctly and, well, good.

    And at such high levels (levels you would find on a CD), it seems that there is a lot of hiss. I assume most of it is coming from my amp/speaker, seeing that I am using a practice amp, no hum/hiss eliminator, and I am recording in my home. But is there a way to get rid of it, or at least have minimal hissing? And it's not like uber loud, it just sounds as if my speakers are turned up all the way and there is that natural hiss from so much gain or whatever. (excuse my lack of vocab and correct terminology).

    Also, as I normalize the tracks, and add more gain, it seems to get a little more grainy, even though there is no clipping. Does this happen or am I just tripping?

    Any tips and suggestions are definately appreciated :) I love you guys :wink:
     
  2. hueseph

    hueseph Well-Known Member

    When you normalize a track, your faders should be at unity gain and your max db for normalization should be 0. There shouldn't be any clipping. If there is you should compress or limit first depending on how bad your peaks are.

    From what I understand, the way normalizing works is, the average peak of your track is calculated and raised to the level that you set. So , if you have an average peak of -10db but you have few peaks at -6, those few peaks are likely to clip. I'm sure someone will come along and correct me soon.

    The other thing to consider is, because normalizing raises the overall volume of the track, your noisefloor also is raised. Hence the hissing.

    Things you can do about the hissing are, edit out the spaces between your actual playing or gate the track if you have to. You can try equing it out tbut that will effect the tone of your guitar.

    The bottom line is, try to avoid any kind of processing if you can. Do your best to get the loudest signal to disk beforehand. What soundcard are you using?
     
  3. hxckid88

    hxckid88 Active Member

    Thanks for the reply. I've taken a look at the noisegate and limiter, but not sure how to use it correctly. I know as for the noise gate you basically set a floor/ceiling, but I dont know what is best...

    I'm using a Firebox along with Cubase LE.
     
  4. IIRs

    IIRs Well-Known Member

    Don't normalise, ever. At best its a waste of time, at worst (if you record 16-bit files) you are adding truncation distortion each time.. that's probably why it sounds grainy!

    If its too quiet, turn it up! ;)

    As far as noise goes, try to back off the distortion a bit, use good quality cables, check the guitar is properly shelded & keep it well away from CRT monitors etc.. when you have reduced the noise at source as much as possible you can then try a software noise gate to clean up the rest.
     
  5. Reggie

    Reggie Well-Known Member

    Allow me :wink:

    The normalizating that is typical in recording software finds your highest single peak (say, -5.5 db) and raises everything to bring that peak up to zero (+5.5 db of gain) or whatever level you want to normalize to.

    If you have several multitracked files that you normalized, the combination of all of them going through your main bus will cause clipping.
    So yeah, normalizing is weak.
     
  6. hxckid88

    hxckid88 Active Member

    So then the alternative for raising volume is...?

    I mean, I'm no professional (we'll see 5 years from now once I get my degree :wink: ), but lets take for example an average CD, all tracks on the CD is the same level, same type of sound, at least from what an average listener can hear.

    Should I just naturally record loud? If that happens, the microphone might clip (im using an SM57). Plus, if I record guitar too loud, the amp and cab might start getting too grainy, too hissy, it'll sound better at lower volume (at least that what it seems like in general).
     
  7. IIRs

    IIRs Well-Known Member

    The channel fader..? :roll:

    You need to understand the difference between peak levels and average levels: peak levels tell us how much headroom we have in a digital system, but do not reliably indicate apparent loudness, which is dertmined by the average level.

    The standard normalise function places the loudest peak at exactly 0dBFS.. but this is pointless as it is almost certainly not the correct volume if you intend to mix it with other tracks. An RMS normalise function is more likely to acheive constant apparent volume, but you risk clipping unless you normalise to a very low level.. either way you eventually have no choice but to use your ears and the channel faders, so why do you insist on first applying a destructive normalise which, as you appear to be recording 16-bit files, is damaging your audio irreversibly every time you do it..? :?
     
  8. hxckid88

    hxckid88 Active Member

    Thanks for that, whenever I record I just record as is, which usually means its pretty low, which usually means I have turn up my speakers all the time, then people listen to it and theyre like what the hell its really low...

    Thats why I wanted to learn how to bring it up to a listenable level without distortion...

    Well, I learn something new every day, don't use the normalize function? haha
     
  9. IIRs

    IIRs Well-Known Member

    if you have recorded 16-bit audio at too low a level and the channel fader can't apply enough gain you would be better off using the make-up gain of a compressor plug than applying a normalise.. unless you first convert the files to a higher bit-depth of course. Better still to record 24-bit audio in the first place, as then you can record at low levels without any quality loss.. and if you must normalise it won't sound so grainy afterwards ;)
     
  10. hxckid88

    hxckid88 Active Member

    The Firebox DOES record in 24-bit, I just figured it would take more out of my CPU, not to mention I only have 512MB PC3200 RAM at the moment. So far I haven't experienced any slowness. I have also been recording at a 44.100hz, it seems to not let me choose 96k even though the firebox supports it.

    What I have done before is what you mentioned, IIRs. I used the compression plug and did a gain make-up, and that seemed to help, rather than the distorting normalize feature.
     
  11. IIRs

    IIRs Well-Known Member

    Recording higher bit depths won't affect your cpu much, but it will make your HD work harder. I would advise you to record at 24-bit all the time if your system can keep up.. higher sample-rates are less important IMO, especially if your recording is destined for CD.
     
  12. hxckid88

    hxckid88 Active Member

    I just figured it all out, recording in 24-bit/96K doesn't make a difference, my computer can keep up just fine.

    But I seriously need a bigger harddrive, exporting these .wavs are like 90MB (with lots of effects). I already have like 1.5GB across maybe like 8 songs or somethings.
     
  13. IIRs

    IIRs Well-Known Member

    24 / 44.1 is good enough for rock n roll if you ask me :lol:
     
  14. hxckid88

    hxckid88 Active Member

    Sweet, sounds like a plan. I couldn't tell the difference between 44.1 and 96 honestly. It sounds cleaner but its hard to put that on a scale... It just sounds good to me haha.

    Anywho, I've been writing songs to experiment with effects and cleaning up hhiss. thanks for the advice.
     
  15. GregP

    GregP Guest

    I just wanted to chime in and agree with IIRs. At best, normalizing is identical to using the gain fader. At worst, you can get distortion/graininess.

    Since your best-case scenario is identical to using track gain fader, you might as well just use the gain. :D

    An important thing to consider is that when mixing a track you will almost inevitably have the average levels of individual tracks far below 0dB anyhow.

    As an example-- and this is NOT the only way to do it, it's just something I picked up as a tip along the way and it works for me so I use it-- when I start mix-down I begin with the drums soloed, and set them so that average levels are at around -9dB. Now, this is danged QUIET compared to a full mix. :D To the point where lots of people would say, "Gee, the quality doesn't seem there. There's no impact to the drum track!" But of course, it's an illusion. Volume is a deceptive but powerful illusion, and one that's hard to shake.

    BUT, persevere. Tell yourself it's just an illusion that the drums have no punch, and carry on. OR, turn up your monitors if you must. ;) Anyhow, then you mix in your instruments... usually bass first, so that it 'sounds' right with the drums. I don't have a real recommendantion for levels... as soon as I've done my drums at minus nine-ish trick, the rest is subjective for me.

    By the time you've mixed in all your tracks, you'll probably already be clipping your master out and will be adjusting individual tracks downward. NOW who's worrying about them sounding too quiet? It's all you can do to keep things manageably quiet now! Once it's all mixed, then you can start farting around with the mastering of it all. This is a bit of a grey area for me, but suffice it to say that by the time you've done some master limiting, your perceived loudness will start making its appearance.

    And all you have to do is be patient and wait for it. <grin>
     
  16. Thomas W. Bethel

    Thomas W. Bethel Well-Known Member

    The reason a CD sounds the way it does is because AFTER the recording session and mixdown session the CD was mastered. That is where the loudness is matched song to song and everything made to sound the best it can. Imagin, if you will, you just bought a new car and it looks sweet. But you take that new car to a detail shop and they make it look FANTASTIC. That is what mastering is all about. (it also helps to go to someone with the equipment, the monitoring setup and the experience to do a good job on mastering) Lots of places ADVERTISE mastering many do not really do what they say they are doing. Buyer beware.
     
  17. Cucco

    Cucco Distinguished Member

    That ain't much. For recording projects, many of my clients get their own hard drive. I usually run up to 5-10 gigs of information per song. That's each track, never deleting anything. (Scratch tracks, etc get moved to their own "dead tracks" and labeled so that if I need them, I can get to them later with ease.)

    In general, I'm going to echo many others' statements here. Normalizing is a process which should never be done.

    Also, I'm going to correct some misuses of terminology.

    Average level refers most commonly to RMS level. This is what determines the "loudness" factor of your track. If the drums are averaging -9dBfs, your drums are damn loud. Really damn loud. The mix would collapse at this volume when you add anything else to it. If the peaks are reaching -9dBfs, then you're probably a little closer to reality.

    Basically, you do still want to record hot in a 24 bit environment (hot = getting your levels to reach close to the top of the meters - notice, I said close - NOT over. Digital clipping is bad.) With the additional headroom that you get with 24 bit, it's not essential to keep your meters close to the peg, but don't let them stay at -20 dBfs. (This is really personal preference. With 24 bit, you do have at least that much more headroom, but coaxing that much more gain out of a standard DAW can be a royal pain in the ass.)

    The simple rule is, if it clips or you get a digital over on any of your individual tracks, turn it down until you don't.

    Then, when everything is summed, you may still get clipping on the master bus because you've got a lot of hot signals cramming into one 2-channel bus. This is where mixing comes in. Lower the levels until you no longer have clipping. My personal tactic in this case is to leave at least 3dB of room. That way, a Mastering Engineer is not restricted when they get my project - they can EQ, compress, limit and so on and have a little room to play.

    The volume should not be adjusted in the Master bus except for by a mastering engineer. Period. Get the sound you want while mixing. Don't get it too soft or your ME will hate you. Don't get it too hot, or your ME will hate you. Simply put, get your guitars, vox, drums and everything right during the mixing phase. Each channel should be at an appropriate level that you give yourself that 3 dB cusion that I spoke of a moment ago, but you still have a good mix of usable amplitude (or "volume").

    Yup. Listen to Tom. This is 100% correct. One of the main jobs of the mastering engineer is to correct the synergy of the album. They fix the levels of the songs, spacing between them, fades, etc. It ain't just pushing everything through a limiter and getting it loud.

    AMEN Brother Tom!! Preach on!!

    A couple rules here -

    1. If they don't tell you what they're doing to your mix - don't take it to them.

    2. If they won't let you sit in on the Mastering session - don't take it to them.

    The reputable guys out there know damn well that their careers aren't in jeopardy if a client watches them work or asks what is done. It takes GREAT ears, excellent knowledge and butt-loads of equipment to do a good mastering job.

    There's a company around here who gets the majority of the mastering jobs in this area and it just makes me sick. Their mastering set up is a noisy PC in a warehouse running out of a Sound Blaster card into a Sony reciever powering a pair of Alesis M1s (I know, they've got their loyal following...)

    As for mastering equipment, they have exactly this:

    SoundForge 7
    Waves L2 software limiter
    Waves Q10
    Roxio CD burner

    Their mastering process involves loading in the mix, limiting it until it's almost a square wave, cutting out the low and high frequencies and then burning it to a disc.

    They get clients b/c they advertise like hell and their rates are cheap.

    I wouldn't send my worst band to them for mastering. But then again, I won't master them myself either.

    Just some thoughts (long-winded ones at that...)

    J.
     
  18. hxckid88

    hxckid88 Active Member

    Thanks for that, Cucco, that was helpful.

    From what I understand, I've been doing this all along, thanks to a little a common sense.

    When I record, I just try and make sure my Firebox doesnt show clipping (referring to the LED light), by simply adjusting the knob on it. When the riffs are down and recorded I readjust using my ears, making sure the meter doesn't clip (a little yellow will tell me that I'm right there). Basically you're telling me I should just adjust everything BUT the master volume.

    But see, using my equipment for personal use, I *cough*"Master" my tracks, noob style haha. But I think I'm understanding the process better :D
     
  19. Cucco

    Cucco Distinguished Member

    Hey - No problem!

    There's nothing wrong with a noob doing as much as they can to a track to make it commercially viable. Mastering your own is really only a cardinal sin when you are operating a commercial facility or you are really trying to sell quality for your clients. You'll actually find that, many clients are willing to get their tracks mastered when they find out:

    1. How affordable it can be.
    2. How much of a difference it can make.

    Where are you in the country/world? You might be able to watch an ME do some magic and you'll better understand what you need to do as a mixing engineer to get ready for mastering.

    J.
     
  20. hxckid88

    hxckid88 Active Member

    I'm actually a senior in high school, I'm moving to San Jose,CA to go to Foothill Community college (GREAT music technology programs) and SJSU. Which is some 30-45 minutes from San Fransisco. I plan to do lots of internships ;) I'm really excited to get going on my career, and hey, I might find out I like something better, related to music, or maybe not related to music... But I do plan on getting in on some action and watching some engineer's do their job.

    I recorded my bands old crappy demo (http://www.myspace.com/astoriaca), keep in mind they're our old songs if you listen, ew, our new stuff is much better. Anywho, people thought THAT was good. And I guess it's because smaller locals bands really just play shows to have fun and that stuff, to have a demo out even it was recorded like crap. What I've been doing is asking local bands if they'd like to record with me, and I give them the truth that I'm a "noob" at recording, I can make it sound decent, I cant make it sound professional, and that I'm still learning so as long as they work with me I work with them. Most bands offered to pay me $200 for a 4-5 song demo. And thats like WOW to me. Studios charge $200 an HOUR lol.

    But anywho, I guess it often depends on your expectation. If you have a old beat up 1990 Honda Civic, are you really going to pimp it out and get a detailed and spend the money? Some people only expect something listenable.
     

Share This Page