Recording level in digital

obee

New member
I recently went "digital" and I'll never go back. I've found that a level somewhere in the -5 to -8 gets the best sound for guitar and vocal. If I even try to inch up to 0 (not going into the red) it doesn't sound as good to me. But my analog mind keeps saying to push towards the red(even though I know you can't oversaturate in digital or it sounds crappy) Am I on the right track (so to speak)?
 
Last edited:
I record as close to 0dBFS (dB Full Scale) as possible. But as you know, don't go over or it'll sound like a bad dream.:D
 
I get it up :) around -3, and with some mild compression on the track, I get almost no clipping.
 
The principle of digital recording is the same as that for analogue - the best results and lowest track noisefloor will be obtained by recording a signal as close to 0dB as possible. The only difference is, where in the analogue domain you can overload a track and cause a sometimes desireable effect, tape saturation, you have no such effect, or leeway in digital. One clip and you're F'd.
Record as close to 0dB as you can without clipping. If it doesn't sound good to you if you go nearer than -5 or -8, that can be for one of two reasons:
1. Clipping anxiety :)
2. You need to calibrate your meters.
 
DOBRO!!!! What do you mean I get ALMOST no clipping????
You can't clip! No clippin' allowed!:D
 
sjoko2 said:
The principle of digital recording is the same as that for analogue - the best results and lowest track noisefloor will be obtained by recording a signal as close to 0dB as possible. The only difference is, where in the analogue domain you can overload a track and cause a sometimes desireable effect, tape saturation, you have no such effect, or leeway in digital. One clip and you're F'd.
Record as close to 0dB as you can without clipping. If it doesn't sound good to you if you go nearer than -5 or -8, that can be for one of two reasons:
1. Clipping anxiety :)
2. You need to calibrate your meters.

I used to think the very same, however I dont anymore. A lot of engineers will tell you that you SHOULD NOT record as close to OdB for some very good reasons. Without writing a several page explanation read "setting Recording Levels" by Jeremy "Jezar" Wakefield. As a very respected engineer he makes some excellent points. For example, why should you record a constant hi-hat pattern at just below 0 dB? For the mixdown you will have to really trim it down. I realize you get the best signal to noise ratio, but after reading this article and talking to some other engineers I think this is bad advise.

Heres the link...http://home.onet.co.uk/~jzracc/levels.htm
 
Last edited:
I have just read the article concerned Greggy, with ever increasing amazement. It contains to many .... well .... I don't even know what to call it. Lets be polite - a complete lack of knowledge of digital sound? I also wonder how long ago this was written? It certainly must have been quite some time ago.
Anyone who can write "why should you record 16 bits when you will only use 12" really needs to do a bit of studying. This is in adition to remarks like the fader being less sensitive in the lower regions - while a digital fader works in equal increments throughout the entire path.

Don't regard this as critisizing collegues. But it is a fact that many of them are entirey accustomed to analogue ways, and ways of working suited to analogue. They often continue to use these ways in the digital domain, until they find that some working practises should be disgarded for a different media. I had to make the change, I found some practises hard to discard. I have spend a lot of time training engineers on mainframe digital consoles for one of the main manufacturers. For some people its like taking the wheel of their car and telling them to drive through the rush hour with a joystick instead, its different! Others take to it like ducks to water.

Just one thing about the High Hat. Of cause, its a HH. Use sence and reason. You're not going to spend a load of time setting your levels so your HH is just below clipping out, you don't need to. But keep the level high, and you keep your noisefloor low. Just plain fact. Record it without to much shelving and / or EQ, and when you apply your shelving - your volume has dropped considerable already, and you have plenty of room for processing.

Another point worth making. Its very important to ensure your meters are calibrated correctly, especially when you're working with digital. Another thing is that cheaper DAW's do NOT have the best of meters. Often they are not fast enough to respond to peaks. In that case it really is a matter of using your ears!
 
Yes if you dont hit 0db your leaving off usable bit length, but does that make the track a technicaly worse recording?

Has anyone taken 46 tracks of vocals and instrumentation with most everything maxed out at 0db? This can be a horrible situation to mix. What about the fact that you have left no headroom on your output bus? So now I'll pull the whole mix back a bit to get it under control. And then I have to do a lot more compression to get the tracks to fit properly. The result? A low energy low audible volume mix that sounds like some life is gone.

And lets assume we are usuing quality A/D converters! The early more inexpensive A/D converters required you to try to reach 0db to obtain bandwidth.
 
Let me speak about my bud Jezar.The man didn't fall off a turnip truck recently.He spent 5 years at Pinewood studios in the sound effects dept.He has produced and engineered commercials for radio and TV and pop tunes for movies and radio.
His technical engineering knowledge in reference to digital gear is apparent from his stint on the Psion design team as well as the free direct X and vst plugin Freeverb which he has given away the source code to (open source,yes!).
Jezar lives in the UK but I have corresponded with him via email for almost two years.The man is a stone cold pro who is an enormous asset to our community.
Follow this link to his web page and look around.
http://home.onet.co.uk/~jzracc/index.htm

Tom
 
sjoko - what I *meant* was, I almost never clip using that approach. :D BTW, have you heard the track that clips on Morissette's Jagged Little Pill? It's right there - you can hear it - a little bit of ugly in a good, good tune. I wonder if they named the album after it.
 
Like I said Tom - I wonder how long ago it was written, I think it was long ago. I hate spouting critique at collegues, never mean anything personal, but stick firmly to what I said :).
Dobro, if you listen carefully you can hear a lot of clipping on many tracks. Digitally recorded tracks I get in for mastering have, more often than not clips in them. Sometimes I can edit them out, other times they are to large to edit, in which case its the customers choice, and often the overall performance wins out, the clipping is left alone.
 
skojo2
Nothing personal bud!I like a lively exchange of ideas and don't mind how often folks disagree with me.I did want to set the record straight about Jezar's well respected place in our community.I recall the date of that article's posting as being spring or summer 2000.Not this year,but not ancient history either.There was a conversation on the n-track forum (Jezar's home gear isn't PC based but he diddles around with the pc for fun) at fasoft.com about headroom old style analog vs modern digital.I remember a point he made about recording minor percussion like claves.
Backing up,the preface to the arguement involves the difference in the old signal to noise ratio which forced you to slam levels and the much wider relative space we have now to place the level.If you are going to have the final level of the claves at ,say -10 dB,and you aren't forced by technical limitations to push 0 dB,there is no practical reason that would REQUIRE you to flirt with oversville.
At least that is my understanding of his point.
Nice to butt heads with you,sjoko2.
respectfully

Tom
 
Logic, sense and reason are all wonderful things :) Especially when applied to recording.
Of cause, don't risk clipping, why should you. Lots of things come into play, like the noisefloor of your system and the noisefloor of your signal chain. ehh - all shit everyone knows already anyway, or should know.

From my experience it seems you can place people in groups - Group one, with firm analogue roots, is the group whom would apply EQ and all sorts of processing to a signal at the tracking stage, recording a signal in the level and sound they perceive to be its place in the final mix. Group two, use microphone placement and acoustics to track a full spectrum signal for each track, as direct as possible, leaving processing to a later stage.

There is always room for differences!
 
Good point about stylistic differences based on the changing nature of the beast.The new way will certainly use more "raw" recordings and have you do all the processing in the digital domain.
I'm an old tape head with a lot of old habits.I admit the superiority of the new gear,but I think the comprehensive understanding of the "old breed" engineer was head and shoulders above some of today's technicians (using mic position as EQ,etc...).
The ancient Chinese had a curse "may you live in interesting times".Hoo boy!That's us now.

Tom
 
LOL now you're talking. Mic position. To me this is the one most important element there is, together with acoustic environment and acoustic positioning. Its the most exciting part of recording, the most difficult part to get right, and the most satisfying if you do get it right. Also the reason I am building my new studio where I'm building it. Killer acoustics in the tracking room with 29" ceilings and huge variable bass trapping, and a vocal booth with high ceilings, no paralel surfaces and variable trapping as well. If I have to EQ anything going to tape - I feel I've done something wrong.
 
I admire that getting it right straight to disk approach with the right mic, right position. I read an interview with Al Schmitt recently, and he talked about a session he did where he used no EQ while tracking, and no EQ in the mix. Now, *that's* recording. From my limited perspective, I'm also impressed by people who can get the level right off as well. My understanding is that 24-bit not only gives you more headroom, but more 'tailroom' as well - you don't have to worry about the noise floor as much. So if you don't *have* to track every signal as hot as possible, why do it every time? :)
 
sjoko2 said:
Killer acoustics in the tracking room with 29" ceilings and huge variable bass trapping, and a vocal booth with high ceilings, no paralel surfaces and variable trapping as well.
29"????? Well... I guess at 29 inches, standing waves don't have a chance to build up a whole lot... but how do the musicians play lying down on their sides?!?!?! :D :D

Mind you, Sonexx and other acoustic deflectors are probably cheaper by the inch! Excellent way to save on building materials!!! :D :D :D

yeah... I know... I'll shaddup now! ;)

Bruce
 
Back
Top