[Jzr171]

I have seen conflicting statements around the internet. In your experience, is it best to record at a lower volume and then boost or record loud and then turn down.  I have tried both. When recording lower I found it was hard to get things, like vocals, to ever stand out. But then going loud leads to clipping. I have heard the right compressor is the answer to either method. But I am not good with compressors.  Any tips are appreciated.

[tone]

First of all, a compressor will not fix clipping. Once you've recorded something so loud that it clips, you can't process the clipping away again.

The trick is to spend some time setting up your signal so that it;s hot (loud) enough to work with while not loud enough to clip. Personally I prefer a track to be as hot as is reasonably possible- that way you don't introduce extra noise from adding gain to a quiet signal.

Compressors only balance out the loud/ quiet parts of your take, making the whole thing a more consistent level. Some people use them for tracking, but they're no substitute for a good signal chain.

[Boydie]

I don't want to confuse but I disagree slightly with @tone on this one, which is why you are probably finding conflicting advice about the best approach - neither approach is right or wrong, just getting to the same place in different ways and both approaches can be used in different situations

The one thing that everyone agrees on is that clipping whilst recording is a VERY bad thing as once it happens (and is recorded) it is there forever. Whilst nowadays there is "audio restoration" software that "may" help fix it later it is definitely the best approach to avoid clipping at all costs

In the days of recording to tape the mantra was very much what TONE is saying - record as loud as possible without clipping

The reason for this is that tape usually had a "hiss" and noise so the goal was to get a really good signal (i.e. Your recording) to noise ratio - essentially a loud recording to drown out the hiss

Some people use compressors whilst recording to help control the volume of the audio being recorded but I would strongly advise against this if you are not familiar/comfortable with compressors - and even if you are I would still advise against it as you could easily get a lifeless recording if you squish all of the dynamic range out if whilst recording, which is then committed to the recording and can't be easily undone

With the advent of digital recording the tape hiss issue has disappeared. Also, the improvements to resolution and bit depth of recorded audio is also improving the quality of digital recording

In my experience the new challenge is to manage the noise that can be introduced by the preamps on your audio interface, which can be very noticeable when they are working hard

I therefore prefer to record a bit quieter to reduce this noise

As a general rule of thumb I would recommend that you aim to get your meters to peak at around -12db - remembering that you can only change your recording volume via the audio interface and not within your daw

I think this will give you plenty of "headroom" (i.e. If you get carried away and sing a bit louder you have some room before it will clip)

When it comes to mixing - if you can't hear your vocal clearly then simply turn everything else down and turn your monitoring up

During mixing your should not be concerned too much with increasing volume (I typically mix so that even with everything playing my master meters are peaking at around -6db)

When your mix is done you can either get the track mastered to increase the volume to a level that would sound normal against commercial mixes or simply put a limiter on your master bus to increase the volume without clipping - I wouldn't worry too much about this stage yet, just concentrate on getting comfortable setting your recording and mixing levels

[cowparsleyman]

@Boydie - Spot on.

I agree, anything over the 0 on your faders (0dbfs) is a big NO, as that cannot be undone with any plugin etc. you might be able to hide it with other stuff over the top, but best avoided.

A mic will pick up anything around it, if you are recording in studio it'll pick up scratching your eyebrow, touching your guitar, tapping your foot, there are REALLY sensitive, if you are recording at home it'll pick up dogs barking, planes, cars, heating pumps, shouts upstairs saying dinners ready...so if you sing quietly your singing won't drown out the dogs, so to do that you need to have the level of you vocal much higher than the background noise. Audio interfaces also generate their own noise, which sounds like a permanent hiss or a hum, to get rid of that can be very hard but usually just turn all unused inputs off and have your input up as loud as you can get without clipping (going in the red)


First get the level from whatever you are recording from, say a Vocal mic at a level that is not going into the red at all on the Audio interface at the loudest part of the song, it might be a good idea to set your mic up so you can see both from where you are singing, then you need to look at the level on the DAW track...

A mistake that is often made is that everyone tries to get as close to 0 as possible, but as Boydie says, there is no need, digital recording has a lot of headroom which sort of means that you can record lower than you think and it'll be OK to make it louder later, but no above 0.

As he says between -18 and -12 is a good level on tracks and I use exactly the same guide level -6dB for master buss levels.

Also as he says dynamic range is SO important, compressor and maximisers can squeeze all life out of a song, so I try and avoid them and keep the song really airy and open, I heard a podcast by Daniel Barenboim last year about what music is (Reith Lectures BBC podcasts), and he said that it's all about space, music is nothing if there is no space, and I agree.

Ian Shephard is a bloke who is on a mission to end the war on crushing dynamic range, there is a link to his stuff on the recording sub forum, it's well worth a listen, as is understanding what Loudness Units are, this really helped me improve my mastering.

So low vs loud? I'd rather get it right.


Anyway, as always, hope this helps

cpm

[Jzr171]

@tone  @Boydie @cowparsleyman

Thanks for the feedback. I believe I recorded most of my last album aiming for 0 db. Some songs worked out better than others this way. I found the songs I recorded last came out a little better than when I started.  The whole process took a year of on and off work.  It sure beat the first record I did with an old band from high school. I think I recorded everything just under peak. Which wasn't entirely by choice. The mic I had back then was USB powered and I really had no control over the gain. Now I have real mics with real mixers and real DI pre-amp boxes, so I have ways of control. And using Logic X I feel I have as much available to work with as professionals have on the DAW side of things. Now if only I could figure out how work compressors without just using presets...

[cowparsleyman]

Glad to assist, with compressors my humble advice is really understand what compressors are for, really understand the terminology, that'll help, rather than paying lip service, and maybe take a term such as ratio, leave the rest of the settings alone, so you can hear the affect of just that, take a lone kick drum, and a compressor and just experiment, it would help if you know what kind of sound you are after, and how you want it to sound in the mix, it's helpful to be ultra picky.

If you do stumble upon a sound you like, save it as a preset and name it so you know when you come to use it again, what it's good for, I usually put a flag infront of my presets so I can recognise them from system presets, maybe prefix them with a z, or your initials...

Hope this helps

cpm

[Boydie]

@Jzr171

When I finally get some time I do plan to do,some tuition videos or write a book with a view to putting a lot of production, song writing and music theory stuff in to very simple terms that are accessible to beginners - but are not "dumbed down" so they are still relevant to intermediate and advanced people

My (extremely shortened) take on compressors is....

Compressors can do lots of creative things and be used to add "colour" to a sound, which should be considered separately to understanding the actual role of a compressor

All a compressor does is help you manage the variation between loud parts and quiet parts in a track

The best way to think of a compressor is to imagine that you have an "imaginary friend" in your studio and you want them to turn the volume of a certain track up and down to make the volume more consistent throughout the song

Think about the things you would need to tell them...

They would need to know...

At what volume level should they bring the fader down (this is the THRESHOLD control on a compressor)

How quickly should they pull the fader down when the volume exceeds the level you told them - e.g. You may want to leave a little delay to get the initial crack of a snare and then bring it down (this is the ATTACK control on a compressor)

How much should they pull the fader down when the volume exceeds the level you told them (this is the RATIO control on a compressor)

How long should you reduce the level for before you let it go back to normal (this is the RELEASE control on a compressor)

Finally, with all of this volume reducing the track may end up quieter than it was originally so there is a control on compressors called MAKE UP GAIN to bring the overall level up

These are the most common controls and different compressors usually have a variation on these controls. In some cases (especially emulations of hardware units) some of the controls are "fixed" and can't be changed

So, as @cowparsleyman has said - armed with this knowledge you can start thinking about the sound you want to get from using a compressor and then start aiming for it

I think it is useful to think about what you would ask your "imaginary friend" to do with a fader and then use this to set your compressor

E.g.

You may have a kick drum where you don't want too much "punch" but you want the "weight" of the kick drum to come through

You would ask your friend to turn the volume of the kick drum down only when it makes a really loud sound (i.e. the initial beat of the kick) so you want a high threshold. You then want to bring this initial beat down in volume very quickly so you set a fast attack time. You also want to bring the volume down quite a lot when it gets loud so you set a big ratio. You then want the "weight" of the kick to come through so you set a fast release time to let the volume come back up quickly. Now that you have got rid of the initial big thud you can use the make up gain to raise the level of the kick drum without it clipping as the loud bit has been reduced in volume (compressed) leaving the "weight" of the kick drum - mission accomplished 

Now you have a snare where you want to keep the initial crack and then bring the volume down quickly so that the snare does not get in the way of other instruments

So it would be a slower attack time as you want to leave a bit of time before the compressor "bites" and pulls the volume down. You may set a lower threshold so that the snare sound after the initial hit is pulled down a lot, which would be especially effective when combined with a big ratio. You want a slow release time to ensure the compressor stays active - but you need to make sure the release is short enough so that the compressor "lets go" before the next snare crack or you will lose the initial crack of the next snare. Finally you may decide you do not need any make up gain at all as you have preserved the "crack" of the snare (i.e. It is not being compressed because of the slow attack setting) and you wanted to bring everything else down - so mission accomplished

I have never seen a "plain English" explanation of compressors so I hope these explanations help you get to grips with compressors.

[Ramshackles]

From an electronics point of view, to get the absolute maximum possible signal to noise ratio from your preamp you need to record as hot as possible without clipping.
I am agreeing with @tone here. @Boydie is reaching more or less the same conclusion, but confusing the reasoning some what.

It is important to understand that we are dealing with 2 different kinds of 'decibels'. In the analogue domain (i.e. what the meter on your preamp or audio interface is telling you) we have 'DBu'. In the digital domain (inside your DAW), we have 'DBfs'.
 0 DBFs is the absolute maximum digitial signal - anything above and you are clipping. The DBu scale is different - 0DBu is about 0.775 volts and is this way for historical reasons (it is based on a 600 ohm load dissipating 1mW of power). Preamps generally clip around +24Dbu so if you are recording at 0 DBu you will have plenty of headroom.

So what is 0DBu in DBFs? It depends on the calibration of your A/D converter. Since DBFs is digital, we can basically denote 0 DBFS to be equal to anything in terms of the actual volume coming out of your speakers. There are standards which vary slightly (e.g. between USA and Europe), but the audio interfaces you and I deal with will generally be calibrated to a set figure with no option to change it.
For example, in my SPL Crimson, 0 DBu = -15DBfs. So if I record at 0DBu on my preamp then I will have a whole 15 DB of digital headroom to play with.

So here we come to the same conclusion as @Boydie - if your DAW is saying -12 to -18 DBFs, you are doing OK. The mistake many people make is to set their preamp to -12 or so. Now we are recording way to quiet so your signal to nosie ratio will be terrible.


If you look at the specifications of preamps, they will state their input noise and/or signal to noise ratio at a specific gain. This gain is usually quite high - e.g. the Focusrite ISA one says: '-126dB measured at 60dB of gain' - 60DB is the top gain the ISA One will do (remember this is the amount of amplification they are adding to your signal - it does not mean that the signal coming out is at 60DBu). The tell you the stat at this gain because (usually) it is where the noise is lowest.

The grace design m101 tells you at 3 different gains: http://www.gracedesign.com/products/m101/m101.htm.
Best performance is at 40DB of gain out of a possible 60DB. Also worth noting that the top gain (60DB) still has significantly lower noise than at 20DB.


I generally agree with Boydie but in this case I am going to have disagree a little bit - as a guy with an electronics background - making your preamp 'work harder' does *not* reduce the quality of the signal, quite the opposite in fact. Its also important to talk in terms of ratios. Yes, turning up the gain will also turn up the noise. BUT the signal is turned up by a large amount - thus you have a higher signal to noise ratio.
The thing you are recording will also have a lot to do with this. Obviously a quite vocal you will need to add a lot of gain to to get it to the required level, whereas a louder instrument will need less gain.

So we come to the actual conclusion, which is 'turn your preamp up as much as possible while leaving enough headroom to work with'. This means that you can actually reduce noise by recording as hot as possible (basically, don't clip) and then turning down the volume of the track in the DAW to what you actually want.
(Of course you can destroy these gains by turning it all back up again at the end, so don't do that. Golden rule of gain staging is don't attenuate then amplify and vice versa.)

Here is what I do.
- Looking at the meter on the preamp (i.e. the analogue meter), turn up the signal until it is just clipping in the very loudest parts of the performance, then I back it off 2-4Db.
- If your A/D converter has a meter, look at that and make sure its not overloading. Many audio interfaces have the preamp and converter in 1 unit with a meter. It will be an analogue meter so just do step 1. E.g. on my SPL Crimson I have 3 LEDs indicating 'Signal', '-6Dbu' and 'OVL'. Even the manual for the Crimson tells you to 'turn up the gain until the OVL indicator flashes in the loudest section of the performance then back off the gain just a little bit.
- Finally check the input signal on your DAW. If you are worried about it being too high, back off your preamp gain a little more.


Bottom line - if the meter on your interface is showing around 0DBu, you are getting the best out of your preamp and your digital level should be absolutely fine

[Boydie]

@Ramshackles

@Boydie is reaching more or less the same conclusion, but confusing the reasoning some what.

With the greatest respect I think going deep in to the theory may cause more confusion than what I posted

I clearly stated there were 2 approaches - it is fine for you and @tone to subscribe to the "record as hot as possible" approach - and I never said this was "wrong"

I was just saying that in my real world experience audio interface preamps can be noisier when cranked - and why risk clipping when you can quite easily record at a lower volume without the worry of "tape hiss"

I guess the bottom line is experiment with the 2 general approaches and find what works best for you - and the approach may vary depending on the source being recorded

[Ramshackles]

I was just saying that in my real world experience audio interface preamps can be noisier when cranked - and why risk clipping when you can quite easily record at a lower volume without the worry of "tape hiss"

This is the confusion. Preamps are noisiest at low gain- 0-20 or 30db. As long as you aim to have your preamp meter averaging around 0 DB/VU (if it occasionally goes over its fine), then your digitial level will be fine.

[Boydie]

Thanks @Ramshackles

I think we are essentially saying the same thing

I was always referring to the meters within the DAW as most home audio interfaces do not have VU meters - so to answer the original question posed by @Jzr171 the best way to go is to adjust your pre-amp gain (using the knob on the audio interface) so that the meters within your daw are bouncing around -15 to -12DBFs

Yoj then just need to make sure there is no clipping occurring in your preamp (unlikely at this level) but you should always get in the habit of checking your audio interface for clipping lights (usually simply a red LED)

By taking this approach you are starting with good "gain staging" that will give you a healthy signal

You should also find that at this level your preamps are not being "cranked" as I still stand by my experience that cranked preamps may cause unnecessary noise if they are driven

I hope you can see from the discussion that although there are different "views" (which may explain your confusion when looking around t'internet) there is general agreement on the optimal recording level

The other thing we agree on is to not record as hot as possible - this is a "hang over" from tape days where driving in to tape drowned out the tape his and sometimes led to desirable tape saturation

Record at the volumes that have been recommended above and at the very worst you will have a great starting point