home recording cd question

[shawlie]

Thanks for the more advice and ideas.

RustyAxe, I do use normalize, but never tried it with the Zoom itself. I notice when I look at the waves on computer that turning the Zoom on and off makes peaks higher than the guitar playing itself. I was afraid that it would normalize it based on the highest peaks - which are from me turning the Zoom on and off. But I will try it and see the difference.

Originally Posted by Pokiehat

What is headroom? Headroom is the difference between 0dB (the clipping point) and the peak signal level. You can use up this 12dB of headroom in a number of ways, one of which is to normalise the track to 0dB (thus gaining +12dB of volume).

Using a compressor to squeeze loudness out of a multitrack recording does not take into account that it has been well mixed. Therefore it is entirely possible that your mix is quiet because it simply been mixed badly with poor headroom management and poor gainstaging. One of the byproducts of a good mix is that its nice and loud before you have to resort to compression.

Thanks very much for the good information, very interesting. I don't quite understand the first thing in the quote, though - would you normalize after you set a limit? Does that make a difference?

And I agree - I'm sure the things I am making are not mixed well. It always sounds very "booming", even thought the volume is never really that good.

Like another suggestion, I have tried raising things to peak at around -3 to -1. It does help the volume.

Which makes me now wonder something else - say I am making a song with two guitars and a voice. I will cut off the peaks (turning on and off the recorder), then normalize it (default setting).

I will then turn down the volume on my computer to about a quarter to listen. I turn down the low end of the guitar a little, and a little on the voice, too. Then I do not know how to approach the volume as a whole - the track volumes versus the master volume on the mixer. I have no method, is there a method to it? There are three volume levels to think about (the computer, the tracks and the master volume) - is there a standard way to approach it?

I don't expect great quality (I'd have to play better...) but just something that's a bit ok to listen to. As an example of something I tried the other day - the reverb is awful heavy (hate how the piano sounds) but is the volume sounding good enough to continue this way, or should I start all over?

very short sample without the vocals:
sample clip

Again, thanks for all the ideas and advice.

[Pokiehat]

Originally Posted by shawlie

I don't quite understand the first thing in the quote, though - would you normalize after you set a limit? Does that make a difference?

I don't understand what you mean.

Let me try another example.

If you have Audition or Soundforge, use the FM generator to create a mono 60hz sine wave with an amplitude of -15dB. Now I want you to create an identical 60hz sine wave with the same amplitude.

Sum both of these waveforms together. What happens is that you get the same 60hz sine wave but with double the amplitude.

Now, reverse the polarity of the second sine wave (turning it upside down basically). Sum them again. You should get a flat line. Hopefully seeing this should make sense because if you have 2x sound waves and you mathematically add them together you are continuously summing the amplitude of both waves. What happens when the frequency of one or both waves is different? They will partially cancel each other out and alternately reinforce each other at cyclical intervals.

You need to do this test so you can see it. Right now I'm guessing it sounds very abstract but when you see it then it makes perfect sense.

So what happens if you have a sub bassline in key of B1 (60hz fundamental) and you also have an 808 kick drum (decay phase keyed to 60hz B1) striking whilst the bass is playing? What happens is that you hear the bass normally and the kick normally except when they overlap and sound at the same time. When this happens you get a sudden spike in amplitude which eats up alot of headroom and makes the kick drum sound way too loud.

This is bad mixing technique. There are ways to stop this kind of interference:

1) lower the volume of the bassline at the intervals where it sounds at the same time as the kick drum. This is typically done using a compressor and the sidechain input to automatically apply gain reduction to the bassline when it detects an amplitude spike on the kick. A similar but less accurate effect is to ride the volume fader of the bassline and drop it low whenever the bass hits.

2) add a phase offset to either the bassline or the kick drum. This will make them cycle slightly out of sync and will reduce the amplitude spike when both strike together. Be warned that this causes periodic destructive phasing. Read up about it.

3) Simply tune the drum and the bass apart so that they cycle at different periods that do not reinforce each other's 'fundamental'. A kick drum of course is not a harmonic instrument and thus does not have a pitch reference but an 808 kick can have a long decay phase which can which is why I mention it in this example.

A person who is clever at mixing can use techniques like this to avoid clashes and minimise the amount of 'doubling up' that wastes headroom. Poor mixes are usually very quiet because of things like this.

The first thing I recommend you do is to go and download Voxengo SPAN. Its a free VST spectrum analyser which you can use to get a visual representation of your guitar and your voice. You will need to run this in a VST compatible program like Soundforge or Cubase. By looking at the spectrum you can see where alot of the acoustic energy is (concentrated around the fundamental and the most significant harmonics). Now you can use an EQ to shift the emphasis of one away from the other so they overlap easier. You can link the two with compressor sidechain and let a comp handle automatic gain reduction. You can spread one in stereo by adding a time delay to one of the channels. You can do an often used tactic where you record one pass on your guitar, then record another pass with the previous take in headphones. Pan one hard left and the other hard right and the slight differences in timing give a constantly shifting phase relationship which makes your vocals sit easier in the middle.

I recommend trying out a combination of all the above to see what works best. There are no rules but you can inform yourself about recording and mixing and you can think of it structurally and methodically.

Like another suggestion, I have tried raising things to peak at around -3 to -1. It does help the volume.

Leave yourself some headroom. This is important. One of the big problems alot of people have with digital recording is that metering in most software sucks balls. What happens is that you get people recording their guitar and vocals as hot as it will go (right up to the clipping point) and then they dump the .wavs into a pro tools LE session with the entire mix riding the red line.

If you do this it will close certain doors with respect to what you can do to this mix post recording. For a start, any sort of additive equalisation or any additional sounds are going to cause the mix to clip. Which means you will turn down the master volume.

Always leave yourself some headroom in a mixing session. More headroom means you can mess around with the mix and add things and the beauty is that you can always raise the volume later if you have headroom left at the end of the session.

I will then turn down the volume on my computer to about a quarter to listen. I turn down the low end of the guitar a little, and a little on the voice, too. Then I do not know how to approach the volume as a whole - the track volumes versus the master volume on the mixer. I have no method, is there a method to it? There are three volume levels to think about (the computer, the tracks and the master volume) - is there a standard way to approach it?

remember what I said about summing. The meters on the master bus represents the sum peak amplitude of all the channels. This is hard to visualise but mute all 3 channels first. Enable the first one and press play. The master bus should be peaking at the same volume as channel 1. Now unmute channel 2. The master bus volume should increase because now it represents the sum amplitude of channel 1 and 2. If this level jumps dramatically and it sounds awful with terrible disparity between loud and quiet then something is clashing. Try using some of the techniques I mentioned above to make them sit easier together. You will note that if you do this successfully, you will gain headroom on the master bus. i.e. the master peak volume will go down.

I don't expect great quality (I'd have to play better...) but just something that's a bit ok to listen to. As an example of something I tried the other day - the reverb is awful heavy (hate how the piano sounds) but is the volume sounding good enough to continue this way, or should I start all over?

very short sample without the vocals:
sample clip

Again, thanks for all the ideas and advice.

I can barely hear the piano. It sounds very centred too. If you want you can upload 13 seconds of the piano and the guitar and I'll mix them for you. Its only 2 channels so it'll be a piece of cake. Then I'll post screenies and tell you how I did it.

[shawlie]

Wow, thanks for the very interesting reply with a lot of detail.

I use a program from Magix and am not sure how to make a sine wave... but thinking about what you say, I think I can visualize it - also after reading (and trying) what you say about muting the tracks and turning them on one at a time (and your bass and drum example, too).

That two or more things at the same time with similair hz will add to the overall volume, making it so you will have to turn down the master volume to avoid peaks. So the important thing (I think I get from reading) is to either slightly alter the timing, the eq or the volume (with compressor) to avoid these peaks when everything overlaps? And that using something like Voxengo SPAN as you suggest will help see at which hz these things overlap?

(And I think I will also have to read what the actual difference is between volume and amplitude, now I think of it).

I will try to download it, and hope it will work with my program. When you say "reverse the polarity" (turn it upside down) - is that the same as "inverting phase"?

I have uploaded a 13 second clip - one of the guitar, one of the piano/keyboard and one with both. I would very much appreciate a look at how you would do it - and find it very amazing and very nice for you to take so much time and trouble with helping. Very much appreciate it and will try to figure more things out on my own.

guitar
piano
both

[Pokiehat]

I will try to download it, and hope it will work with my program. When you say "reverse the polarity" (turn it upside down) - is that the same as "inverting phase"?

Polarity is not the same as phase but you are on the right track.

Check these out to get a better idea of the concept:

http://www.communitypro.com/files/li...PHASE_TECH.pdf
http://www.kevinkemp.com/homerecordi...ial/micing.htm

These are difficult terms to explain because they are abstract. However they are simple concepts and are very easy to visualize so illustrations are important. Oddly enough, it is easier to understand whats happening by 'seeing' it rather than 'hearing' it although if you can do both at the same time thats even better since you can begin to correlate the change in what you see with what you hear.

Some technical terms in audio are ironically enough quite difficult to illustrate in auditory terms. For instance, dither is much much easier to explain with images than it is with audio dither tests.

I don't know what Magix music maker is like these days but the last time I used it (5 or 6 years ago?) it was necessarily limited. I don't think you can do things like run a spectrum analysis in it. However you can get all the tools you need to test the concepts for free. Check out the FL Studio 9 demo (save disabled but fully functional). Voxengo SPAN is freeware. Voxengo Audiodelay is a time based delay plugin thats also free. mgPhaseshifter is free. The only problem is that if you go down this road you open up alot of doors and things can get complicated fast.

When I get home I'll mix it for ya. By the way, do you have the piano without reverb?

[shawlie]

Thanks a lot for the links - the pictures do help! I can see why the reverse polarity would cancel each other out (reminds me of a show I saw about people trying to make quieter machines, but that may be something different).

Just not really sure why two of the same signal makes it so much louder (but I suppose "knowing" is almost as good as "understanding", at least at the moment).

And the illustrations about phase were also very interesting - along with the sound samples. Pretty incredible how much of a difference it makes when you hear it. And now I see more the difference of polarity and phase (the one has an opposite wave form, the other comes in sooner or later). But seeing how phase can look like reverse polarity at a certain point was interesting, too.

I admit - I will have to read that pdf file about ten more times, it is a little confusing. But again, pictures do help explain it.

Then of course - knowing how to actually use this kind of information is a different thing. Also because it says that it reacts differently at different hz levels. I will try and download the Voxengo SPAN and look at the other demos/freeware you point out - it is much more complex than turning the master volume up and down... it is confusing, but it is interesting to read about (and I think it would be fun to try some of that phase changing stuff and see how it changes the sound).

The reverb on the piano is from the keybaord I think (or else I edited the original wave file - I think I might have). So here is a sample without reverb I just played and recorded:
no reverb piano
guitar piano (no reverb)

Again, thanks for the information and links. I guess it will be a little hard to learn some of this stuff, but if it can help to make things sound a little better, it will be worth spending some time on.

[Pokiehat]

Sorry for the lateness. Yesterday I took a looksie at the 2x files you uploaded and in a futile sort of way had a go at fixing problems to do with recording in the mixdown. Now, I know this doesn't work very often but I thought I might get lucky.

Anyway, once I was done with that I finally got around to mixing it this evening. Heres the result A/B with yours, normalised to the same RMS level (-21.9dB), scan window size = 1995 milliseconds:

http://www.mediafire.com/?innmxm1vuhw

I'll go over the signal chain in a moment and then give you suggestions on how you can fix some things on the recording side and generally just be aware of certain things that will help you immensely in the mixdown. So with a little work on your end you should be getting much much better results.

[Pokiehat]

Right the way it works is as follows:

Channel 1 = guitar panned left
Channel 2 = piano
Channel 3 = guitar panned right + phase offset with channel 1 = 10ms

Sends = Reverb (100% wet)

Channel 1 and 3 routed to Send 1 with a little bit of SIR Reverb on it using a Concert Hall impulse from a Lexicon PCM91. Both channels are not routed back to the master bus.

Channel 1 and 3 routed to Channel 4 so this is the combined guitar track. Channel 2 is connected to the side chain input of a compressor just before channel 4. The way this works is that when the piano makes the channel 2 meter jump up, it triggers the compressor to automatically reduce the gain of channel 4. If you concentrate you may notice that the guitar gets quieter when you hit the piano harder. This is just to give the piano a little bit more breathing room whilst having the guitar recess around it. The compressor is set to 1ms (fast) attack, about 470ms release, 9:1 ratio (yes its quite high). There is no makeup gain.

Channel 2 is routed to to Send 2 (SIR Reverb + TC System 6000a impulse response - 'f1-0-4 warm cathedral')

Channel 2 has phase offset of 5ms between left and right.

Channel 4 gets complicated because theres a whole lot of tricking on this channel to spread it out trying to let the piano peep through. Theres a multiband stereo imager on it. I'll have to open up the project to get all the numbers

Here are the EQs with spectrum underlays for each channel:

Channel 1 & 3:



Channel 2:



The piano looks smaller at first but you can see the harmonics (the large series of peaks). If you obscure or EQ out any of these then it sounds unnatural. Pianos work real well in a sparse mix where you get the full effect, note decays, damper noise etc. Alot of that can disappear if you smother it in a cluttered mix. Your guitar was well recorded but the piano has a very noticeable amount of hiss. It may also be slightly out of tune thought that might just be my EQing. Thats the other tricky thing about pianos. You can make them sound of out tune if you EQ away too much of the fundamental - enough that the first harmonic becomes the dominant pitch reference.

Personally I would rerecord the piano if possible. You can work with what you have but certain doors will be closed to you. For instance, I initially wanted to make the piano really airy and keep the guitars nice and bassy/midrange but when I put a shelf EQ at 10khz on the piano it just amplified the hiss really badly so I ended up turning that high shelf into a low pass filter (see spectrum of channel 2). This de emphasised the hiss but I couldn't get the high end up and over the guitar so I went the opposite way - made the guitar more twangy by scooping out some of the lows and around 500hz so the piano fundamentals poke through.

The thing about properly recording your instruments is that it gives you alot of freedom to trick it in the mixdown. Poor recordings can be used and I've seen some producers work with astonishingly poor recordings with hiss and buzz and everything and they can still get results but try it yourself at some point, You tend to find that you have to shift the emphasis away from the flaws in the recording and so you will end up mixing it one way instead of having the freedom to mix it any number of ways.

Mixing is one of my weakest areas so a really good mixing engineer/producer can probably trick it alot better than I can. You don't need high end gear really. I used a bunch of free plugins (Audiodelay, Fruity Compressor/Peak Controller mainly). GlissEQ and FL Studio I had to pay for but its worth it. Multiple spectrum overlays on GlissEQ make it the most practical EQ I've ever used. I prefer it to hardware EQs costing many times the price.

The things you probably want to focus on are mic placement, room acoustics, eliminating electrical noise from getting into the recording and things like that. The other thing is focus on playing really tight. Your rhythm is a little slack in places. If you play really tight, even bad mixes don't matter to a point. Listen to any of Hendrix's recordings because they are all astonishingly badly recorded and improvised in some of the most hamfisted ways but they sound amazing anyway because that band was tight as hell.

Lastly, if you learn just enough about mixing to understand what you need out of your performance and recording, then you'll be good. For example, if you are tracking live and you have the piano and your guitar fed into your headphones for direct monitoring, and you notice that the piano and guitar are really getting in the way of each other, consider one of you playing in a different register. I tried pitch shifting the piano down an octave and it sounded nice. It took the piano out of a problem range I was having on the EQs. The only problem with pitch shifting to that extent is that it wasn't recognisable as a piano any longer but you get the idea. With good monitoring you should get a rough idea of how its going to work in a mix, where it needs to go. You can roughly identify where there will be problems and possible solutions.

PS: Its really weird getting .wav files with 7.2dB+ of headroom. Normally when you get a PT session every channel is kissing the red line. That said, you can record much hotter. I would even suggest you run it up to the red line and dial it back so its not clipping during the performance because thats how you are going to get the most dynamic range. Once you record to 32 bit float, then you can gainstage and alter levels non destructively so you don't need to use all that dynamic range if you don't need it. But if you record too quiet then you need to add alot of pain post recording and that amplifies all the low level stuff too, background noise, hiss, amp buzz, someone coughing far away etc.

PPS: I forgot to check for mono compatibility. Try collapsing the stereo image to see if anything disappears. If it keeps most of its integrity then thats good for people who listen on mono devices (i.e. clock radios). If you are destined for radio play you will at some point need to consider the issue of mono compatibility.

[Fran Guidry]

Originally Posted by Pokiehat

...
PS: Its really weird getting .wav files with 7.2dB+ of headroom. Normally when you get a PT session every channel is kissing the red line. That said, you can record much hotter. I would even suggest you run it up to the red line and dial it back so its not clipping during the performance because thats how you are going to get the most dynamic range. ...

Please do not follow this advice, sorry, Pokie.

Record with peaks around -12 dbFS and raise the level later when you mix and master. Even if you're recording 16 bit, if you're not in a soundproofed space your background noise limits your dynamic range well below the range available in your recording gear. Even if you're using an H2.

http://www.massivemastering.com/blog...ing_Levels.php

Fran

[rick-slo]

Originally Posted by Fran Guidry

Please do not follow this advice, sorry, Pokie.

Record with peaks around -12 dbFS and raise the level later when you mix and master. Even if you're recording 16 bit, if you're not in a soundproofed space your background noise limits your dynamic range well below the range available in your recording gear. Even if you're using an H2.

http://www.massivemastering.com/blog...ing_Levels.php

Fran

+1
In 24 bit recording don't even think about approaching the red line. Recording at -12db or even less will sound better and leave plenty of room for post recording software changes.

[Pokiehat]

I disagree and the blog you posted does nothing to dispute anything I said. You should be recording hot if you want to get decent dynamic range anyway. Some people take it a bit too literally and hand you a session where every track is kissing the redline in soundforge but hey, you can work around that. Anyway, when you are recording, the signal goes to an AD converter, gets sampled and then you import this digital reconstruction into your DAW session.

First you need to make sure the peak signal going into the (digital) mixer is referenced to 0 VU or +4dBu. If you ever use a really nice desk theres tonnes of headroom above 0 VU and you can keep on pushing up the gain for like +18dB before the signal turns to crap.

If you find yourself working in and out of the box you will need to use a line trim plugin to ensure your meters have the same reference level (in this example you may want to calibrate your digital meters to 0dBFS - 18dB) which will give you roughly the same amount of headroom as your desk. When I'm recording hot as hell I use Sonalksis Free G for line trim as and when needed which as the name implies is free and does exactly what it says on the tin.

If for whatever reason you want to dump a session from Pro Tools into an SSL the same problem can arise in reverse. If you get tracks slamming the redline in PT referenced to 0dB(FS) and then put it through an SSL, the meters will be completely pegged from the start and any outboard additive signal processing on top of that is going to turn the whole thing to crap. So instead, trim the output from PT so you have a good 18ishdB headroom so that 0 VU = -18dB and you can go back and forth to your racks and a desk and your meters wont be completely out of whack.

Anyway, I like being able to record hot as hell because I can always use a trim plugin and there are times when I do want to record hot, or I do want to drive an analogue filter or a tube amp into saturation because the distortion is what I want. But maybe I'm in the minority in thinking that this is a creative decision and that the important thing is to know when its happening, that you are doing it intentionally and that you do it out of creative necessity rather than because you don't know any better.

[shawlie]

Wow, thanks very much for the mix and the explanation!

Yours is so much cleaner, crisper, less "busy" soundning - the difference is really inceredible (can't hardly stand to listen to mine now, actually!) Also much wider sounding, without sounding artificial (like mine tend to using reverb).

I think I hear how the guitar softens a little with the harder piano notes - it does calm it down, makes it much nicer.

I think I understand at least some of what you explain, but do not know what "send" and "side chain" mean. I will have to look that up.

But what I at least will try to experiment with:
I love the guitar panned left and right - the one on the right with the phase offset: I assume that is so they aren't playing exactly at the same time, and then you don't get a build up of sound level (like your explanation of revesre polarity and phase offset)? I think I can find that on my program and am excited to try it out.

For the piano - you used a phase offset between the left and right stereo waves? So only one track, but slightly "out of synch" between the left and right? Is this also to keep things clearer and less muddy?

I think I understand a little of what you say about channel 4, but again I have to look up what side chain means. You have it so that when channel two reaches a certain level, the compressor on channel four lowers the combined guitar track? (just not sure how to go about doing that, but again - will look up about "side chain").

I did not know you could use a compressor like that - setting a level on one track that will effect the volume on another. So I will really look into trying to figure that out - seems very useful.

Maybe a stupid question, but did you just combine channel 1 and 3 to make channel 4, and then do you keep channels 1 and 3 or just the combination (channel 4)? Or does the compressor on channel 2 only effect channel 4, leaving the levels of 1 and 3 un-touched? I guess my question is- are there "two" guitars, or "four"?

Yes, my keyboard may be slightly out of tune - it sounds that way to me sometimes. I'm not sure how to record it right. Now I play it and record it straight onto my keyboard. After that, I put a line into my Zoom2 and record it onto that. I import it into my computer, but keep getting a mono recording. So I use an older program I have, and select "stereo to two mono objects" - which gives me two tracks, one in stereo and one that is empty. I throw away the empty one and export the stereo track as a wave. I don't know why I get so much noise (the keyboard is just noisy, I assume?) or why it's always mono.

I tried to quiet the room as best I could and used some blankets a little behind the Zoom to see if that would help. But yes, it's not going to be great, I know. And the mic was fairly far away (an arm's length and I have fairly long arms). Timing does give me problems, and have trouble with it. Even more when I'm trying to do more than one instrument - timing is off on the guitar, then the timing is off on the piano, then I have to put the tracks together by ear until it sounds more or less in synch. Yes, something I'm always trying to work on, that is true.

And later on in the song, I do go down an octave on the piano, just for variation. I like the sound of the high notes, but wil try out your advice - if the mix would sound less busy and yet more pleasant with lower piano, I think I'll give that a go and see how it works out.

I'll have to keep looking back at your advice and will keep looking things up, too. I'll try my hand at something this weekend and maybe I will improve at least a little. Things are still confusing, but using phase offset and panning are things I could maybe start trying now.

It really is amazing what you did with that clip - and thanks again for all your help and time!

[Pokiehat]

the side chain is an input on a compressor. A compressor has an input (the signal you want to compress goes in here), then the signal goes through a peak detector which tells the compressor how loud the signal is in realtime. Then theres the actual compression bit which is determined by the threshold (at what level the compressor should begin working) and the ratio (how much the compressor should work when it detects that the signal is above the threshold). Then theres an attack/release envelope which determines when the compressor should reach full gain reduction after detecting a peak over threshold (attack) and when it should cease gain reduction after the level dips below threshold again. Then it goes through another gain stage (makeup or output gain) and then to the output stage.

The side chain is sort of like a direct line into the peak detector bit of a compressor and skips the input. Basically, instead of using the input signal level for peak detection, it detects the peaks of a completely separate channel. In this case, the gain reduction is happening on the guitar based on the loudness of the piano, not the guitar.

A send is a type of bus on a mixer. You have a row of channels with inputs and outputs. I redirected the piano to channel 2 for instance. Now I connected the output of channel 2 to a send bus which has a a reverb on it set to 100% wet. This is better than just putting the reverb on channel 2. By putting the reverb on a send bus I can put however many channels through the send bus with the 1x reverb unit. Early on, I tried putting the guitars through the same bus but decided there was too much verb going on so I dialled it back.

I love the guitar panned left and right - the one on the right with the phase offset: I assume that is so they aren't playing exactly at the same time, and then you don't get a build up of sound level (like your explanation of revesre polarity and phase offset)? I think I can find that on my program and am excited to try it out.

Its an age old trick. Ideally you are supposed to double track your guitar. So you do one mono take and then you immediately do another mono take. Your timing needs to be good. If your rhythm is too loose it will sound very bad. No matter how good you are though, your timing wont be machine perfect so there will be subtle drifts and delays in when you pick notes. Not big enough to be perceived as an echo but what happens is that you have a constantly changing phase relationship between the two takes. I just improvised with what I had which was 1 take on the guitar. The difference is that the phase relationship between channels 1 and 3 are a fixed 10 milliseconds. I encourage you to try double tracking your guitar and see how it works out for you.

There are times when you won't want or need to do this and its preferable instead to capture your guitar from lots of different mics spaced around a room. But double tracking is helpful for making room in a mix and making guitars sound wide and huge. Pretty much every heavy rock track with thrashing distorted guitars is double tracked. Its a really common trick.

For the piano - you used a phase offset between the left and right stereo waves? So only one track, but slightly "out of synch" between the left and right? Is this also to keep things clearer and less muddy?

Yes. Channel 2 is dual mono. The right channel has a very short time delay so it starts a few milliseconds behind the left channel but the delay is so short that you do not perceive the right channel as an 'echo' of the first. I used mgPhaseshifter for this and it wraps around and whilst I use milliseconds, it is more appropriate to talk of a phase shift in terms of degrees because the left channel is the reference from which the right channel is shifted.

However, if you want to hear for yourself the kind of effect 'stereo widening' does to a sound I recommend playing with mgPhaseshifter and Voxengo Sound Delay (both free). It can produce some odd effects, exagerate the width of an instrument and appear to make it sound as if its coming from a different direction instead of dead centre. This is not the same as panning.

Be careful when changing phase as if you go back to summing tests you will find that it creates periodic destructive phasing (comb filtering). And if the phase offset is 180 degrees then summing to mono will result in completely cancellation. If you are ever destined to get airplay on mono devices (like radios) then you need to be careful with messing with phase and you need to keep checking for mono compatibility by turning the master bus to mono and see if the song keeps most of its structural integrity. It is possible to make some instruments periodically disappear so it might sound great in stereo but turn it to mono and parts of your song are gone.

I tried to quiet the room as best I could and used some blankets a little behind the Zoom to see if that would help.

Blankets don't really do anything. Furthermore, I don't think its always a good idea to try to deaden a room by turning it into a makeshift vocal booth. I've built a completely mobile recording setup that fits in a backpack so the great advantage with that is that you can setup and record in places that have fantastic acoustics. traffic noise is a problem if you live on a busy street but the hiss on the piano sounds electrical in nature. Its not a ground loop. It may be that you recorded it way too quiet and then tried to add loads of post gain which also amplified the hiss. That or your preamp couldn't supply enough gain to bring the mic up to line level or something. I don't know the details of how you recorded the piano.

Maybe a stupid question, but did you just combine channel 1 and 3 to make channel 4, and then do you keep channels 1 and 3 or just the combination (channel 4)? Or does the compressor on channel 2 only effect channel 4, leaving the levels of 1 and 3 un-touched? I guess my question is- are there "two" guitars, or "four"?

I kept channels 1 and 3. You can always go back to them and make changes. I sent both of these to channel 4 for convenience. Basically when I got the balance between 1 and 3 right (volume, EQ etc), I disconnected them from the master bus so they make no sound on their own. I connected both to channel 4 instead, the output of which is connected to the master bus. Then everything I do on channel 4 affects both channels 1 and 3. I only hear channel 4 from this point on unless I disconnect it and plug 1 and 3 back into the master bus.

If I want to lower the volume of these 2 track guitar, it means that I don't have to lower the volume of channels 1 and 3. I just leave 1 and 3 alone and lower the volume of both on channel 4. This is not the same as 'grouping' because FL Studio doesn't have groups. A group is where you choose like 2 or more channels and link all the faders and buttons so what you do to one of those channels happens to all of the channels in the same group.

There is only 1 guitar in this track. However, I used the recording of your guitar twice. Also, the end result sounds like only 1 guitar which is important. If you listen to an album like Nirvana's Nevermind you may wonder why you can't get such an awesome distorted guitar tone with your axe and pedal board on its own. This is also a very good reason for why you should not listen to a professional recording as a technical demonstration of the worth of a guitar. Theres a good chance its been tricked to hell and you won't be able to tell if its the tone of the actual guitar or the tricking that you are more impressed with. Usually its a combination of both which leads to inevitable disappointment when you finally get your American Jag and SM57 for Christmas and you still can't sound like Kurt Cobain. I heard that the chorus riff in Smells Like Teen Spirit was quadruple tracked and layered.

I tried to quiet the room as best I could and used some blankets a little behind the Zoom to see if that would help. But yes, it's not going to be great, I know. And the mic was fairly far away (an arm's length and I have fairly long arms). Timing does give me problems, and have trouble with it. Even more when I'm trying to do more than one instrument - timing is off on the guitar, then the timing is off on the piano, then I have to put the tracks together by ear until it sounds more or less in synch. Yes, something I'm always trying to work on, that is true.

Ok, I recommend trying out and reading up about different micing techniques because you can do crazy things if you have more than one mic. For instance, you can use a cardiod condensor angled towards the nut to catch alot of fret noise and top end. You can also be plugged into a pickup and recording at the same time into a second channel which always gives good, even bass. And you can also set up an omni mic a good few metres away to get the sound of the room.

Thats 3 channels you can mix down to a stereo pair and give more or less emphasis to the fretting sounds, natural reverberation etc. Mix in more of the omni mic source and the mix feels more distance. Mix in more of the close mic source and its drier and sounds closer. You can use mic positioning and direction to create an artificial sense of space in the mix and provided you are recording in a great sounding room (concert hall!), you shouldn't have to resort to tricking it in the mix. Thats the ideal situation though and not many people have access to a nice mic locker and access to a great sounding room whenever they want. So we make do.

Steve Albini is a producer that does very little tricking in the mixdown. Hes all about mic placement, distance, direction, mic choices and combinations and using as much of his environment as possible without resorting to artificially constructing it with things like hall reverb simulations.

[Fran Guidry]

Originally Posted by Pokiehat

I disagree and the blog you posted does nothing to dispute anything I said. You should be recording hot if you want to get decent dynamic range anyway. Some people take it a bit too literally and hand you a session where every track is kissing the redline in soundforge but hey, you can work around that. Anyway, when you are recording, the signal goes to an AD converter, gets sampled and then you import this digital reconstruction into your DAW session.

First you need to make sure the peak signal going into the (digital) mixer is referenced to 0 VU or +4dBu. If you ever use a really nice desk theres tonnes of headroom above 0 VU and you can keep on pushing up the gain for like +18dB before the signal turns to crap.

If you find yourself working in and out of the box you will need to use a line trim plugin to ensure your meters have the same reference level (in this example you may want to calibrate your digital meters to 0dBFS - 18dB) which will give you roughly the same amount of headroom as your desk. When I'm recording hot as hell I use Sonalksis Free G for line trim as and when needed which as the name implies is free and does exactly what it says on the tin.

If for whatever reason you want to dump a session from Pro Tools into an SSL the same problem can arise in reverse. If you get tracks slamming the redline in PT referenced to 0dB(FS) and then put it through an SSL, the meters will be completely pegged from the start and any outboard additive signal processing on top of that is going to turn the whole thing to crap. So instead, trim the output from PT so you have a good 18ishdB headroom so that 0 VU = -18dB and you can go back and forth to your racks and a desk and your meters wont be completely out of whack.

Anyway, I like being able to record hot as hell because I can always use a trim plugin and there are times when I do want to record hot, or I do want to drive an analogue filter or a tube amp into saturation because the distortion is what I want. But maybe I'm in the minority in thinking that this is a creative decision and that the important thing is to know when its happening, that you are doing it intentionally and that you do it out of creative necessity rather than because you don't know any better.

Here's the "dumbed down" (his words) summary of mastering engineer John Scrip's blog post on recording levels, at the link above:

THE "DUMBED DOWN" VERSION: Stop recording so hot. Instead of trying to get your tracks to peak at -2dBFS, have them peak between -20 and -12dBFS and your recordings will almost undoubtedly sound better. Mixing will be easier. EQ will be more effective. Compression will be smoother, more manageable and predictable. You're in the age of 24-bit digital recording - Relax and enjoy the headroom. Even if your only concern is the volume of the finished product (which would be a shame, but it happens), recordings made with a good amount of headrom are almost undoubtedly better suited to handle the "abuse" of excessive dynamics control. QUIETER recordings have more potential to be LOUD later. It's because they're usually better sounding recordings in the first place.

Since you're suggesting recording close to 0 dbFS, if I can understand your recommendations, I'd say John's blog post very directly speaks to, and contradicts, your advice.

Sorry to keep bringing this back, but the bad advice to try to hit 0 dbFS is all over the internet. There are excellent technical reasons to aim at -18 to -12 dbFS and no reason to aim higher except lack of knowledge. Whenever I've persuaded anyone to record at more appropriate levels they have always reported an immediate improvement in the sound of their recordings.

All the best,
Fran

[Pokiehat]

You do realise that you can trim it down to -18dB before it hits the mixer using a trim plugin right? And there are plenty of reasons to drive things into saturation and distortion. John's blog post is about proper gainstaging by giving yourself a similar amount of headroom as you will get from an SSL console with a signal peaking at 0 VU. The point I'm making is that you can have a PT session with all the channels slamming 0dBFS and its really not that big a deal. It happens all the time. If you want a similar amount of headroom as you are used to on an SSL just use the trim plugin.

There are times when I think its wholley appropriate to slam a mix. I remember watching an interview with Noel Gallagher where he said that back before Oasis was big, he used to just turn every dial on his amp up to 10. Oasis was loud, brash and brickwalled in performance so its fitting that Definitely Maybe was loud, brash and brickwalled on recording. You may not like it but I sure as hell did and I can't imagine Oasis being recorded any other way.

[rick-slo]

Originally Posted by Pokiehat

You do realise that you can trim it down to -18dB before it hits the mixer using a trim plugin right? And that there are plenty of reasons to drive things into saturation and distortion right?

If that is the sound you want coming out of your pres I guess.