Use all them there bits?

[rick-slo]

Lately, even in our 24 bit recording age, there has been what IMO is ill advised advice to use all possible bits (record very hot going into the DA converter). This however will result in poorer sounding recordings.

To quote John Scrip:

"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."

This is the full article:
http://www.massivemastering.com/blog...ing_Levels.php


A couple of threads on the subject well worth the read are:

http://www.gearslutz.com/board/high-...12-dbfs-2.html

http://recforums.prosoundweb.com/ind...els#msg_328563

[ljguitar]

Originally Posted by rick-slo

...You're in the age of 24-bit digital recording - Relax and enjoy the headroom.

Hi rick…
Yes, with a quiet room and quiet gear this is great advice.

[rick-slo]

Originally Posted by ljguitar

Hi rick…
Yes, with a quiet room and quiet gear this is great advice.

Yes Larry. Probably always good advise. Room noise will be the same fraction of the total signal whether amp is cranked up or not. Same with noisy mikes. Noisy preamps just get more noisy cranked up.

[sdelsolray]

I agree with not recording "hot". However, John Scrip's statement is more geared toward those mixing dozens of tracks. For two track wonders like us, recording at -20 to -12 dbfs is fine, but there's nothing wrong with recording a bit closer to full scale.

[rick-slo]

Originally Posted by sdelsolray

I agree with not recording "hot". However, John Scrip's statement is more geared toward those mixing dozens of tracks. For two track wonders like us, recording at -20 to -12 dbfs is fine, but there's nothing wrong with recording a bit closer to full scale.

I have settled on using -14. John mentions the level that gear is designed to run which is a different issue then summing tracks.

[RRuskin]

This argument has been going on since analog noise reduction was introduced, if not earlier. Case in point: Some people used to complain that dbx I while reducing tape noise, would sound crappy because of pumping & breathing on transients when printing hot. Explaining that the very idea of noise reduction was to allow for the lowering of record levels which made the dbx process invisible, increased headroom & dynamics while reducing distortion from just about every link in audio chain fell on deaf ears. I suspect people will be slamming things to 0dbfs no matter what they are told. Hey, more is better - right?

[DupleMeter]

I do professional audiophile recordings of classical & tradition jazz music. In my world most engineers agree that -12dBfs is the top peak during tracking. Then leave no less than 6db of headroom (-6dBfs peak) for the mastering engineer (though many of us stick to a -9dBfs peak, which a good mastering engineer appreciates).

Think of it this way - what equipment do you own that operates best at 100%. Does your car operate best when it's redlining? I doubt it. What about your stove. Do you cook everything at 500º? Now how about your home stereo (for those of us who still have them). Do you crank your amplifier to 100% to listen to music (loudness wars aside). I'd wager not.

For those with decent monitoring and a decent sounding room, you will notice that as you approach digital full scale (0dBfs) the soundstage closes in, the mids become oppressive and the "air" around instruments disappears. If your monitoring isn't good enough to hear it, do you really want someone else hearing it that way?

Keep in mind that digital equipment is referenced so that either -20dBfs (AES standard - USA) or -18dBfs (EBU standard - Europe) is equal to 0dBu (0 on an analog component). Do you really want to print your tracks that hot? Even when I was recording to tape I never ran things that hot...and tape is where you had to print hot to keep the SNR manageable.

The bithead mentality came from the days of 16-bit recording. And even then it was misguided. It's one of the main reasons digital had such a bad reputation in the early days - it did sound like crap because it was recorded like crap.

Just some things to think about.

[ljguitar]

Originally Posted by DupleMeter

...Even when I was recording to tape I never ran things that hot...and tape is where you had to print hot to keep the SNR manageable.

Hi DM…
Yes, but it is so hard for some folks to realize the recording process has changed and not just slightly modified...particularly those who have come to digital from the tape world.

This is why (as Rick mentioned) ''I suspect people will be slamming things to 0dbfs no matter what they are told…"-RRuskin.

It is hard to shake old ideas when we are marrying analog equipment with the digital process. It still often ''feels'' the same as before on the recording end of things. I’m still running analog mics into tube preamps, and the tube preamps into an interface box. Only then is it interfaced with a digital converter (invisibly inside a box somewhere).

In many recordists minds the digital interface/hard drive/SD card etc have merely taken the place of tape (without dropout and the need to degauss and realign the heads frequently).

The meters look the same, the pad controls, pan controls & everything else are digitally made to look/respond the same.

I believe this is part of the reason it is hard to modify our proceedures.

[KevWind]

yep hard to break old tape habits

I note that on my Rosetta 200 on the meters instead of - 0- is says "OUCH"

[Bob Womack]

A lot of the business of recording high levels comes from two sources - 1) habits from analog days, as Larry mentioned, and 2) the more official one, fear of the quantizing errors and zero-crossing noise that happen at low levels.

No, you don't want to record too low. However, a good dithering plug-in in your output bus will reduce the problem with zero crossing. My facility, which does both audio post for video and music, has the -18 = 0VU standard. Even then, there are rich voices I record who can drive the system into digital distortion at an average level of -10 VU (-28 absolute) without being excessively peaky - they just have a certain "rip" to the voice that reaches out and touches the top end.

So, like any other physics-based process in the world, it's a compromise. Don't get caught at either extreme.

Bob

[Fran Guidry]

Originally Posted by ljguitar

Hi DM…
Yes, but it is so hard for some folks to realize the recording process has changed and not just slightly modified...particularly those who have come to digital from the tape world.

This is why (as Rick mentioned) ''I suspect people will be slamming things to 0dbfs no matter what they are told…"-RRuskin.

It is hard to shake old ideas when we are marrying analog equipment with the digital process. It still often ''feels'' the same as before on the recording end of things. I’m still running analog mics into tube preamps, and the tube preamps into an interface box. Only then is it interfaced with a digital converter (invisibly inside a box somewhere).

In many recordists minds the digital interface/hard drive/SD card etc have merely taken the place of tape (without dropout and the need to degauss and realign the heads frequently).

The meters look the same, the pad controls, pan controls & everything else are digitally made to look/respond the same.

I believe this is part of the reason it is hard to modify our proceedures.

Actually the meters don't respond the same. That's a big part of the issue.

Here's a Sound on Sound article from 2000:

http://www.soundonsound.com/sos/may0...es/digital.htm

Fran

[ljguitar]

Originally Posted by Fran Guidry

Actually the meters don't respond the same. That's a big part of the issue.

Hi Fran…
I was referring to the digital meters in my last tape gear (with the overload light which flashed to let you know when you were too hot) as opposed to the good-old-needles we had formerly been used to.

You are likely right that some of the digital ones built into software are generally less specific and less accurate and suffer from the lag of time that it takes to process signals. I’d love for them to have more 'segments' in the display...

<edit>I meant to say thanks for the article you linked - so Thanks for the article you linked, Fran!<edit>

[Pokiehat]

I still think it depends entirely on what you are doing. Heres an example of where I completely go against the 'rule' so to speak.

I do alot of my own sound design and although I play and record guitar tracks, I mainly produce my own electronic music and create alot of the sounds myself. I use the D16 Drumazon demo for some bass drum type sounds and this is a 909 emulation. The 909 is a very famous Roland drum machine from the 80s.

Now, the drum is fairly weak on its own but there are ways to make it truly devastating if you want it too. The most common thing to do is to use a compressor to literally 'warp' the drum and bring more of it up to the red line Another is to distort it or use a combination of the two. I've got a 909 kick in Soundforge right now and I'll admit, I wanted to distort it and make it hard and heavy and its pegging my VU meter like you wouldn't believe. Soundforge's VU meter can give a readout way over +3 VU and this kick is pushing up to 11.5 VU. However once the 'shape' of it is done, you bounce it down to .wav and reimport it along with your other sounds into a mix and thats all gainstaged so I can mix with a decent amount of headroom. The same drum that was pegging the VU meter earlier is sitting around -2 VU thanks to the fact that I trimmed it using Sonalksis Free G before the signal hit the mixer. One of the problems with hitting a mixer with an +11.5 VU signal is that you will peg the VU meter all the way and have zero visual feedback. You will probably have to bring the channel fader way down which means the effective range of the fader is a small fraction of what it should be and makes fine adjustments annoying.

I've seen the opposite too though. If the signal is too quiet before it hits the mixer, you can run out of fader before you reach the desired level and +post gain is going to raise the noisefloor relative to peak. I think that everyone mixing ITB should have a gain trim plugin in their repetoire. Free G is free so theres no excuse.

I do agree that if you are mixing big multitracks, you need headroom. If you are mixing part ITB and OTB then it helps to have the same reference levels otherwise your metering goes totally out of whack and its not hard to see why studios have a standard 0 VU = -x dB reference if their clients are constantly demanding to use the SSLs or Pro Tools. It makes workflow consistant for everyone. I do agree that if you are used to pushing another +20dB out of an SSL above 0 VU before crapping out the signal then you'll probably want to trim 20dB off the top of your DAW meters so functionally it gives you the same amount of breathing room as an SSL. Thats assuming you even want to mix ITB if you have access to an SSL.

Me? I think its generally always good idea to leave a sizeable amount of headroom when mixing so you can do additive signal processing (EQ, +gain on your compressor, distortion etc) and to accomodate multiple tracks that when summed will consume that headroom in short order. But there are times when I'm not mixing but doing something like sound design and thats all about making timbre and creating form and shape and I tend to use tools like compressors in very different ways for very different purposes. I don't always use compressors for gain reduction or to reduce dynamic range but use varying attacks and releases to shape the transient or decay of a drum sound in a way that you can't really do with an amp envelope. I also end up pegging meters in a way that some would find scary if they saw the readouts but didn't understand the method behind the madness.

[Bob Womack]

FYI, we moved out of a SSL4048G into our current Nuendo workstation.




One of the things we kept were our Dorrough meters, and we are probably going to add software Dorrough plugs. What are Dorrough meters? They are LED ladder meters that display VU or persistence levels and peak levels at the same time:



If you notice, there are contiguous chicklets lighted from -25 to -3. That's the VU reading. Then there is one chicklet lighted up around +10. That's the peak reading. I'm rather addicted to them. I spent years on VU meters and learned to estimate the peaks, so I find a peak meter, well, mmmm... less than helpful for program material like mixed music. However, you can glean a whole lot of info from the Dorrough - the VU level, the peak level, and the ratio between the two. That sort of analysis is very helpful in trying to get a handle on why a particular mix might be louder than mine. For instance, the Eagles latest album, Long Road Out of Eden, sounds very dynamic but isn't. There aren't a bunch of peaks riding way out above the VU level. The engineers and mastering guys got a real punchy kick drum that isn't very peaky. Marvelous work. Between analyzing the waveforms and watching the Dorroughs, you can learn a lot.

Bob

[ljguitar]

Originally Posted by Bob Womack

...Between analyzing the waveforms and watching the Dorroughs, you can learn a lot...

Wow Bob…

Nice!