Optimum levels for recording meters in Sonar

I have a number of questions here so I hope to draw in some users and cakewalk tech support people to answer them. 
 
First off I have the latest Sonar X3 Producer edition (upgrade) and the Roland quad capture 2x2 audio interface.
 
This is how I record. Please tell me if I am doing something wrong.
 
I first insert drum and usually some synth chord progression. Then I push the "auto sens" button on my Roland audio interface and sing or play as loud as I think my voice or guitar will need during the recording. It sets the gain with the auto sensor. This part I have no problem with. The problems starts after I record. I will notice that the sensor will set itself but often my record levels are low once they are wave files. Sometime too small to even see without expanding the track view. I prefer not using a compressor in the record process. I usually record at 44khz 24bit.
 
First question
What should the peak meters read in Cakewalk while recording and how do I gain them so I end up with healthy waves?
 
Second question
In the compressor section I will set the meters so they only flash green so no orange or red is shown in the various in/out meters in the pro channel? Is this right?
 
Once my signal comes out of the pro channel it is usually left with little track headroom to boost gains. 
 
Third question
This is the worst one: Once I get my mix done I usually find my mix levels in the master bus are so low I end up cranking the amp in my mastering software. Should I normalize my tracks or crank it up at this stage?
 
I usually don't normalize my tracks because I prefer to re-sample my tracks as little as possible. They are already resampled once with V-vocal (I don't care for melodyne that much sorry.)
 
I end up using the cut and add audio gain to louder and softer passages so in the end i am doing a bit of re sampling.
 
Isn't using a preamp to bring my tracks out of the dirt adding a bit of noise? What is better using a preamp on the master or normalizing the waves? 
 
Fourth question. This is almost embarrassing to ask because I have been using cakewalk since cakewalk for dos.
In my mix, at what DB should the main vocal roughly be peaking at.
 
I have heard it said the bass kick drum needs to peak at around minus three DB but usually it is just too loud for my mixes. So that does not seem to work. 
 
Most of my songs have a bass drum, a bass and (lots of) vocals, acoustic guitar parts, pads and solo fill instruments etc.
Decibel wise what instrument should I set to a certain DB level and then fill the rest of the instruments around? This is so when i master and compress the mix I have healthy gains/headroom etc. I am sorry if my questions are simplistic I just usually do what sounds right and get by that way which has led to some really disastrous mixes.
 
Fifth question.
Since a have a ton of recordings with really low db level audio tracks (all over the map) would you suggest me normalizing all of the tracks first? If so, what DB level should I normalize them to? If not should I just use some type of gain? There is also the post and pre issues that are totally confusing. There does not seem to be a stand alone plugin for just gain and often my track gain levels get maxed.
 
Related to this gain thing I have noticed a lot of technicians use the equalizer before the compressor in prochannel. Because most equalizers do not have a makeup gain this seems pointless because the signal does not come into the compressor at unity gain then...
 
It seems logical to take the unwanted frequencies (or add) to a signal before compressing but it seems to never work right. How do you work around this problem using the prochannel EQ?
 
Because I end up with needing a preamp between the EQ and compressor I don't usually place the EQ prior to the compressor.
 
Check out rexredmusicartist on revebnation, sorry for the plug :)
 
 
Thanks in advance for any help here, much appreciated. 

jkleban
Just an old rule of thumb from my analog days is that you want to record as hot as you can but not distort.... a little above the limit on tape and you got some warm compression, a little above the limit in digital and you get crap.
 
Jim

Totally wrong.  John is correct.  With 24 bit recording you can afford to choose a reference level like -20 dB FS and set your rms levels way down there.  There is NOTHING to be gained by recording hot other than the possibility of clipping.
 
What you should do though is adopt the K system and install some proper VU meters.  It makes setting levels very easy and you cannot go wrong and you won't bring a clip light on anywhere either.  Setting peaks to a certain level is not great.  You should be setting your rms levels to be constant (on tracks and buses) not your peak levels.  The built in headroom of the K system will take care of peaks.
 
Putting Sonar meters into rms mode wont help either.  Sonar reports rms levels as being 3 dB down on correct rms levels and rms levels show up way too low on the display.  A VU meter will show 0 dB on the meter when the rms levels are at the calibrated reference. (eg down at -14 or -20 etc)
 
But that is whole other story.  One which I have told many times but many don't seem to get it or want to get it.
 
If you are really keen go to this link. There is a decent explaination of it there:
 
http://forumsarchive.presonus.com/posts/list/42222.page
 
Sorry it is the Presonus forum but I am sure you will cope. I have explained this all too on this forum but it has been a while and you may have to search for it but it is well explained here too. Presonus has a new forum but this is the old forum.

Checking out your link now Jeff thanks for the info. I am not exactly sure what you mean Jeff about the VU meters I will study your post thoroughly thanks.
 
I am not really having trouble obtaining the a nice recording it is just that the waves are all different sizes depending on how a vocal is sung. If i have a loud section in the vocal then I lower my recording gains than the small quiet parts/tracks have tiny waves. 
I can't seem to find a DB standard or middle point for mixing these various levels of recordings. 
 
It seems if i were to normalize sections of recorded materials to a middle of the road volume level it would save me a lot of headaches later on.

Thanks for clarifying that for me that is a drawback and i noticed that the track levels looked like they worked but did not a while ago I did not really put it into a complete thought like you have, thanks.

Jeff I think that is one of the most amazing music related things I have ever seen in your introduction on the K metering system. It is completely over my head but I now plan to learn it and use it if it is within my budget. 
 
I don 't have any of these meters I don't think. I record in mono fro most all things so i would think one meter would suffice (not sure). I suppose some pink noise etc. playing from my PC  would suffice for calibrating the meters. It seems complicate but I imagine if it was all set up and calibrated it would be fairly cut and dry and easy to use. I think I vaguely understand the concept. I think in essence you are saying that transients are a more constant way of measuring sound than relative DB. (not sure if i worded that right)
 
I have always dreaded learning about VU meters and DB but it is time to take the plunge. ESPECIALLY if i can achieve near perfect input tracks. I am very confused between RMS and VU. Need to study the other links too. I hope it is not outside of my (poor struggling artist) budget. I want to learn this. 
 
I clicked on the paypal link and it said 8 dollars. What is the total I will need to spend on this awesome project?

The MeldaProduction MFreeEffectsBundle includes a handy noise generator

Jeff,
 
I appreciate being told I am totally wrong but I think we are saying the same thing.. I was just making a point that you don't want to be too hot coming in with DIGITAL for the nasty clipping (CRAP).... and you don't want to be too low either (distortion with low dynamics)... I think your point of the exact gain is a much more elegant way of saying what I meant to say.
 
Sorry about that but I was trying to keep it simple.
 
Jim
 

Jim I agree with what you are saying re 16 bit recording. But with 24 bit recording the theoretical level for noise and distortion and problems is way down at -144 dB FS (-138 dB real world) So even right down at -20 dB FS we have still got 118 dB of dynamic range to play with. That is why 24 bit recording is so good.
 
I just want people to let go of the myth of recording everything so hot. It is the reason many get into so much trouble later on.

i Just found a definitive answer to one of my clipping issues. 
 
The Sonar help file says
 
Clipping LED. Shows if the input signal to the module is clipping. If there is any distortion in the ProChannel signal chain, the clipping LEDs let you identify where the clipping occurs.
 
Comment:
 
I think I have figured out why my master output is so quiet
 
LED lights in the pro channel have four colors they cycle through
 
1 green
2 yellow
3 orange
4 red
 
Red is the only one that is clipping
The other colors just mean they are nearing clipping
 
This is why my masters are so quiet.
I have been keeping them strictly in the green 
 
Every time I would look for the help file I would be taken to this forum with a jumble of words. (No offence peeps)
 
In finding this about the LED and the previous discussions here it gave me an idea to look at the output levels on the track channel meter.
 
I know this sounds stupid that i have not done this before but... I am learning, beside writing, recording and publishing and having nearly 100 thousand followers online that I invited one by one I have to learn the nitty gritty of this stuff.  Somethings my mind does not like to learn and meters and thinking in DB is one of them.
 
Now I know why I could not get the console emulator to respond much.
 
I thought yellow was clipping and my track volume was useless at that point.
 
I set out to figure out why this level thing was not working today and gosh darned i did.
 
I might add that the Sonar help file could go into levels and metering a bit more. What seems obvious to the person writing the manual is not obvious to someone who has been using cakewalk since cakewalk for dos.
 
Thanks for letting me think out loud and bounce my ideas till I figured out what was wrong. :)
 
You have been a good help. My dilemma was probably too simple to figure out or articulate. Once i figured it out it seems quite obvious in hindsight. Green is not the DB peak ceiling, orange is.
 
I have clocked many many thousands of hours in  Sonar and finally I will be getting some hot mixes.
 
Please detail the help files in sonar concerning the LED light threshold for those as dumb as me and cannot figure it out. 
 
and Jeff I will be buying the meters soon so please stay tuned to this thread and I will be back to comment and make inquires if you don't mind.
 
THANK YOU ALL VERY MUCH!
RexRed
 

Rexred....
 
I was trying to suggest that you turn up your input signal, louder than you think right up until you get RED led on your interface and then turn it down so it NEVER turns red during the entire track.  The QUAD capture does this with AUTO SENS but you have to perform and record a demo to the setting so it will do this for you... perform it all the way through (at least until to get to the loudest part of the performance) so that auto sense works for you.
 
I used to have an OCTA CAPTURE but sold it (it was good) when the UAD APOLLO became available.  They are both good interfaces but the UAD one allows you to insert UAD plugins with ZERO latency and either record or monitor with them in real time.  I love UAD plugins and this was just the bomb for me as my audio interface.
 
Good luck and it sounds like you are on your way to making music instead of reading about how to use your tools.
 
Jim

While I am at it I might like to suggest a couple features I would like Cakewalk programmers to implement.
 
One is in V-vocal. I plan to use V-vocal over Melodyne. I assume Cakewalk switched to Melodyne because it has a few nice features. I would use Melodyne only if i wanted to use those added features. Like converting waves to midi and taking out some unwanted blips in a wave. Also the harmony feature is nice. I sing my own harmonies though.
 
Granted V-vocal, is buggy and is the source of lots of cakewalk crashes. I will be editing in V-vocal and I will click the undo button and sometime lose whole large swaths of editing. This is remedied by saving and lot and by restarting Cakewalk and V-vocal's undo is restored.
 
Sometimes Cakewalk will freeze up and it is directly related to V-vocal.
 
But I find Melodyne is even more buggy and more often causes Cakewalk to freeze up.
 
Besides the editing features for waves in V-vocal are MUCH easier to use and for this reason I will be using V-vocal as long I as I can.
 
Even before Melodyne was integrated into Sonar it was terribly buggy I tried it and did not like it at all then. I was so glad to return to V-vocal. 
 
I think V-vocal is using Melodyne in the background that is okay as long as the editing functionality of V-vocal still works in Sonar.
 
I hope Cakewalk still develops V-Vocal for Sonar and gives users a choice.
 
What would be nice in V-vocal is the smart tool interface.
 
Also the if there were added demarcations for 8th notes and not just quarter notes or maybe even a smart grid in V-vocal. :)
 
Also I have asked about this before. I will mention it again in case not one saw it.
 
It would be nice to be able to add a second ruler up top so I could have two places for markers.
 
I like to put vocals a up top in the various markers so i know where i am in the song.
 
Then I also like to add chords C, Dm, F, etc when I get to recording the guitar/keyboard part.
 
Currently I have to delete my lyrics markers to add chords because they coincide with lyric changes and I cannot add two markers on top of one another and read them very well.
 
So a second ruler just for markers would be nice. (please)
 
One might say why not use lyric view? Well it is way to buggy and takes me HOURS to add lyrics. Often I struggle and struggle with lyric view and finally give up after tons of wasted time. I can add lyrics up top in about 10 minutes.
 
Don't forget a single hyphen in Lyric view or it is all over.
 
Please fix lyric view...
 
Other than those minor issues I REALLY love Cakewalk.

The signal to noise ratio of a final recorded track does not have anything to do with the recording level or even your preamps and converters on the way in  (well hope so!)  The amount of noise present in any recording is usualy due the amount of noise present in the source. (I have found a well set hardware downward expander on the way in can clean up noisy guitar sources big time, no work required later to remove any noise)
 
So while the approach of recording just below clipping seems OK,  it would still be too high for a K system refernece I would imagine.  The problem with this approach is it is easy to just get a little loud and hence you will clip because you are simply too close to that point.  Ruin a good take.  Who wants that. (not all clipping is audible too but that is a case of luck more than anything)
 
By setting your average rms levels down at a K reference you have got all the headroom of the chosen K ref up your sleeve.  Even if you were working at K-14 then the chances of you suddenly getting 14 dB louder is actually quite remote and hard.  And at K -20 even more so, near impossible to suddenly get 20 dB louder.  Us drummers can do it though with a very loud snare hit for eaxmple.  But as I have said before with drums it is more about peak metering and just making sure that the loudest things I might do on the kit will not clip.  (this is more like the record hot as close as you can to clip approach and that seems to work for drums. I still make sure that the loudest drum things I might do only get to -6dB on the peak meters)
 
The VU's wont help you a lot with very transient sounds but once you send a group of transient sounds to a buss and put a VU meter on the buss, there will often be enough rms energy to move a VU meter more normally. The more open and ringy the drums are the more the rms VU metering works.

Jeff,
 
I agree with you 100% but the loudest below clipping technique is something that you learn how to do over time.  It is almost like how hard to hit the guitar strings..... I don't push it too hard but I do like it hotter than not.  It works for me and I am sure that your science is more accurate.  I just want to record, keep it simple.
 
Now, if I were a PRO in a PRO studio, I would be the musician and not the engineer.  It is amazing what you can do in the comfort of your own home nowadays, no?
 
Be well,
Jim

And the Yamaha mixer is the reason I can push the volume to a higher level than that of my interface.  
I actually record different ways using both my interface analog inputs and my 01V into SPDIF. 
The 01V has good metering, And it has compression.  
The Scarlett 6i6 has a ring around the level knob which tells you next to nothing. So I agree with Jeff about the K system, in a way I have always done something similar I want a REAL VU meter. As I said, I have never had an over when using my Yamaha as I've used that mixer more than anything else in my studio. 
If I do use the Scarlett I find it best to record on the low side as I cannot trust what I'm seeing. As a result those tracks are defiantly quieter. 
And yes when it then comes down to mixing and hitting the master too hard, certain tracks I want at the top like vocals, So if I was to record my vocals with the scarlett they would not cut through the mix without either pushing the gain or normalizing in wave lab. ( sorry I would never use Sonars normalize tool, long story) 
In the end I use a different approach for each instrument or vocals. But it all starts with a proper set up of the pre amp. It always did and hopefully always will. I think pre amps loose something when not driven properly. After all, it is analog up to a point.