Call Us: (800) 344-4361 between 9:30AM and 5PM

Mastering Audio by Bob Katz This Third Edition of the "audio bible"…

Twitter Feed!

bobkatz48bit: 2nd prog rock album from White Birch of Norway. Norwegians are prolific. I think they like to spend all winter inside making music! 10 hours ago from Twitter for Mac
Digital Domain's Audio FAQ is the most comprehensive set of audio answers online on the web. Spread the word! Your audio questions answered here. If there's a topic you need answered, please contact us.

Frequently Asked Questions - Search Results

FAQs - Search Results
Search FAQs
From: Barton Michael Chiate
Subject: Re: Question
My comments are:
Bob, Thanks for all your articles. Great stuff!! Required reading for my students. I have a question. Maybe you can help. I have several recordings of classical music that were recorded to DAT through an Apogee AD 1000 with UV22. I dumped them into my computer  to edit at 24 bit.(Figured higher is better)
 
Hello, Barton. Thanks for writing and for your fine comments.
The only reason to "bump them" to 24 bit is if you're processing the material or doing any editing or gain changing. Otherwise the 24 bit didn't buy you anything, as I think you already know
 
I have not done aything to them but edit. They are now ready to master but I don't know the best way to get them back to 16 bit. I have an AD8000 with the D/A card and a Finalizer Plus. So far, I have been using the dither in the TC to a DAT to burn CD's but I would like to avoid writing the file to disk again so I can burn the CD for the duplicators.
 
I see. Well, Murphy's law says to get a bitscope (always have one around) and make sure you didn't accidentily put in any extra bits. If the program did put in extra bits, then ask why and find out why. If the bits are due to legitimate processing, then you really have a 24-bit file. Then use the UV-22 (or something better!) to go to DAT and then load that back into the computer if you must. Or just loop out and then back in....our Sonic system can play back 24 and record 16 simultaneously, can yours?
 
Or, if it's really 16 sitting in a 24-bit slot (8 bits zeros), and you're stuck, then just copy it out to DAT or S/PDIF-AES CDR and then back into the computer. That's probably the fastest way.
 
Don't go through analog. Why do that? If you have the AD8000 you must have a DAT machine around somewhere; doesn't the AD8000 go D-D?
 
Bob
From: Mark Gemino
 
Bob,
Finished reading more of the information on your web (Boy so much there to comprehend!).
And it raised a few more questions. I will primarily mix to my 1/2" Analog deck after I find some GP9 tape and rebias as you suggested.
 
I think that's an excellent idea.
 
But i would like to back that up with a digital copy as well. I have access to a Protools V Mix system with a 24 bit (new) Adat bridge . It has a AES in/out and stereo monitor jacks as well. Would this be better than using the 96Khz finalizer @48Khz and stereo dithering to 16 bits on the TASCAM DA30-DAT, that way i could keep the audio files in @48Khz /24 bit on a hard drive stereo AIFF file type? (in theory).
 
Definitely store at 24 bits for archiving and retaining the resolution.
 
I think that in your case I would like to receive 48 kHz/24 bit Pro Tools (preferably stereo interleaved AIFF or WAV) files as well as the analog tape. The Pro Tools files represent a pristine exact digital output of the board, and the analog tape represents a colored (possibly more "beautiful") output. One or the other will sound better through the mastering process, and we will make the best decision which is best.
 
I do not have a bitscope or oscilloscope to measure what Protools does and to compare each i have to power down reconnect/swap and then listen?
 
If you connect as I described, you should be in pretty good shape. Keep all faders in pro Tools at 0 dB, turn off the dither in Pro Tools (and if you have Pro Tools HD, use the Surround dithered mixer). Make sure if you do use the Finalizer for any interconnections, that it be in total bypass.
 
A) ADAT MDM's out into Optical Digital in of TASCAM TMD4000 console (set to 24 bit word / no Dither) AES out to Digidesign 24 bit ADAT BRIDGE AES input and out the monitor outputs to the 1/2" analog deck.
 
I think the D/A converters in the Finalizer are superior to those in the Digidesign monitor outputs. If possible, for the analog tape, go out the 24 bit AES mix output of the console into the Finalizer (set to bypass) and the Finalizer's analog output to the 1/2". The Finalizer also has tone generators. I don't know if it does more than 1 kHz, I've forgotten. But at least it can do 1kHz at -14 dBFS to get you started. For aligning the 1/2". You still need a technician with a phase oscilloscope and ideally a distortion analyzer for aligning the 1/2".
 
And feed the 24 bit AES output of the Finalizer into Pro Tools. Not too bad, eh?
(only issue here is i don't know much about setting word length or Dither but from
what i read you might suggest word length of 24 or more if available and no Dither.
 
Correct!
 
D) Rent a super D/A converter which is well regarded and forget about the Adat Bridge /Finalizer?
 
If possible... Something like a Prism, Weiss, or even in the consumer domain, a Mark Levinson, or Muse.
 
I understand your 83 SPL reference for monitoring but electrically i am still confused... you said -14dBFS (console 1Khz) equals 0 Vu on ATR 60 1/2" deck
 
This is for the level recording onto the 1/2" and is secondary to the monitoring consideration.
 
(the TMD4000 can be set internally to +24dBu /+20dBu default/ +15dBu for Analog CR/Stereo outs ---would this effect AES levels in any way?)
 
Not at all. And since you're going to avoid the analog outputs of the console they become irrelevant, too.
 
a) Select 24 bit Word length & No Dither on TMD4000 @48Khz sample rate Digital console output.
b) Turn on tones/ Internal to TMD4000 console select (1Khz).
c) Set the level on the output buss level meters to read -14dB with main fader at the 0 (zero) position. (AES output).
 
An accurate meter is very critical. If you want, you can make me a CD-ROM of this test tone as set by the console... feed it through your chain and into Pro Tools, then send me a CD-ROM of this test tone from the console and I'll let you know if it made it.
 
d) On Finalizer select 48Khz /16 bit and stereo Dither check the meter reads -14dB (AES input) & L/R Analog out to 1/2" deck.
 
DO NOT DITHER THE OUTPUT OF THE FINALIZER! SET THE FINALIZER TO 24 BITS (AND BYPASS IT, THAT IS, SUSPENDERS AND A BELT) AND FEED ITS ANALOG OUTPUT TO THE 1/2" DECK.
 
e) Set ATR60 L/R input controls to read 0 VU on 1/2" deck ( use Quantegy GP9 tape at 400 nw/M equals 0 VU).
 
Correct. You're getting it! Let your technician give me a call if he has any questions about what I mean by 400 nW/M = 0 VU. Aligning an analog deck is what separates the men from the boys :-)
 
f) Check DA30 DAT reads -14dB (AES input). AES connects out to CD input
 
I see why you want to dither the outputs of the Finalizer... to feed the DAT. No, don't do that. Make two mix passes, one with the Finalizer set to 16 bit dither, to feed the DAT, and the second pass with the Finalizer in Bypass for its analog outputs to feed the analog tape and its digital outputs to feed Pro Tools. I don't even think you need to have the DAT backup if you have Pro Tools.
 
g) Check CD RW5000 reads -14 dB (AES input).
 
Digital is "absolute." If you are feeding -14 dBFS out of the Console, it will automatically be -14 on all the digital recorders and Pro Tools. Provided the console has an accurate test tone generator.
 
Clock concerns: i normally use Adat sync to digital console @48Khz sample rate , should i be using word clock and tie this to finalizer/adat bridge if used?
 
Good question. You really are thinking! In this case, if the Finalizer locks well to the console under the following circumstances, then you will have the LOWEST jitter if the Finalizer D/A ***IS*** the master clock FOR THE ENTIRE LASHUP:
 
The Finalizer has the ability to be the master clock (internal). If the Finalizer has wordclock outputs, then feed its wordclock to the console and let the console slave to it. If the Finalizer has a spare SPDIF output and the Console can slave to it, then slave the console to the Finalizer's SPDIF, so you can feed the Finalizer's AES output to Pro Tools and slave PT to the Finalizer.
 
This is a complex patch and you have to make sure there are no glitches or clicks, but it is the absolute best patch from the point of view of getting the best sounding analog output from your D/A converter (in the Finalizer).
 
next connect cd player to playback pink noise and set console CR /PB for 83dB spl w/ one speaker at a time (not sure if pink noise is narrow band or not?) using Radio shack level meter no weighting if this is a comfortable level to mix in... begin to mix
 
I said that? Depending on how much compression you are using in the mix, this will be an uncomfortable level for mixing. Likely you will have to turn down the monitor about 6 dB. If your monitor is turned down more than about 6 dB from that reference, then you are likely over compressing.
 
listen back at different PB levels
w/ speakers:
A) HS824
B) NS10's
C) Urei 809's
D) AKG headphones
 
Good. The reference speakers, the merrier, but know which one is your standard and determine if it translates to the rest. In your list above you don't have anything I would consider a "standard", so try to include some wider range, accurate speakers in your listening.
 
play Cd on:
A) Sony player in kitchen
B) Sony Disc man w/ Sony MDR headphones
C) Apple disc player on computer
D) CD player connected to Console
 
Even more important, a real good, big hi fi system in someone else's house, someone you know and trust.
 
Hope this helps,
 
Bob
From:mblock
 
MBlock: Bob,
Would you please help me to settle a debate I've been having with a friend of mine? I maintain that technically it is wrong to digitize audio in the -10 to -20 (peak) range. You're just throwing away bits."
 
Bob Katz: Hi, Mark. In general, from the point of view of the A/D converter, more is better, as long as you are not overloading any intermediate analog stage in line in order to do it. And as long as you have an accurate overload indicator. (See my article on levels: Part I)
 
OtherGuy: "Your interest in making a track as loud as possible would seem to sacrifice dynamic range."
 
BK: Making an initial recording peak near 0 dBFS is not the same as making the final mixed production as loud as possible. Making the track peak near 0 actually maximizes the dynamic range of the recording with respect to the signal to (dither) ratio of the recorded multitrack medium. Afterward, in post production, you can alter (lower) the mixed or post produced level of that previously recorded track to any point you wish, so, once again, the dynamic range of the final production is also not sacrificed. Unless I totally misunderstood the poster's motives, he is wrong.
 
MB: "No, just the opposite. The dynamic range is the difference (in dB) between the noise floor and the maximum signal a system can tolerate without audible distortion.
(The term "audible distortion" requires a definition in the analog world, but with a digital signal it's clear cut.) Full 16 bit dynamic range requires peaks near 0."
 
BK: Correct, as I just said. But "Requires" is a dangerous word. For example, if I were recording a live performance direct to 2-track with a world-class properly dithered 16-bit A/D converter, and one peak in the entire hour hit 0 dBFS, but the second movement was all at -20 dBFS (the original range of the players), it would probably just be fine. The sound would be world-class. As long as you did not have to raise the gain of the second movement in post production or do any post production at all. The point being that in direct to 2-track of live material, if the original dynamic range is good-sounding, then you have made a proper recording. But if you have to do any post production on it, potentially needing to raise or lower some tracks, then it pays to record the 2 track (or in this case, more likely, multitrack) 24 bit with good ("24 bit") A/Ds, so that you have more room above the quantization noise floor when getting into post, where you might have to raise or lower gains. More on this in a moment.
 
OG: "It seems that your rationale forces you to constantly lower the music level in the mix anyway."
 
BK: That's fine, not a problem. By recording "hot", you have made a good, clean original recording which you can then later lower in level inthe mix or post stage. That's called "having your cake and eating it, too." It's a good idea: Maximizing the record level of A/Ds, especially the cheap, shoddy ones, is a very good thing. Another way of thinking of it is you are increasing the signal to garbage ratio of the original recording, where the garbage is at the bottom of the A/D converter. More on this, as we get into the "24 bit rules"....
 
MB: "My aim is simply to bring audio into the Avid at maximum resolution, and that means digitizing as hot as possible without distortion."
 
BK: Correct. But as ADCs have gotten better it is no longer necessary to digitize to full scale. Regardless, as I said, you are preserving dynamic range if you are simply recording the unaltered microphone output to the ADC at near full scale.
 
OG: "You are equating bit depth with volume. Are you sure that is really true? By definition, this would imply that a 24-bit system can achieve louder levels and a wider dynamic range. most commercial music is highly compressed to start with and really doesn't cover this full dynamic spread. As a result, if you have 20dB of actual (mixed) dynamic range with "loudness" that falls in the bottom, middle
or top of the available range, it really doesn't end up sounding different - or so it would appear to me."
 
BK: You are more and more correct for the more modern ADCs recording at 24 bits. Set your average level to -20 dBFS and let the peaks fall where they may. However, the older and more inferior ADCs may benefit from "maximizing".

MB:My understanding is that digital audio sounds best as you approach maximum level; it sounds pretty nasty down in the bottom of the range.

BK: Shall we clarify that to say: An Analog to digital conversion sounds best as you approach max. level..... (within reason).  On the D/A side and mixing side, it's a bit more complex, and it is possible to remain clean without needing to hit full scale. Otherwise you could never mix productions with reasonably wide dynamic range that are designed to be reproduced well (gotta sweat the soft stuff....).
 
MB: "I try to record with pre-mixing in mind. I have the playback levels all set at zero on the console (10dB down from the top of the fader.) I set my record levels so that I have a pretty good mix coming back with all faders in this ruler-straight line. If a guitar peaking at full scale digital is too loud, I can either pull down the record level or the monitor level.
 
BK: As long as you are talking about the A/D conversion (tracking) situation in 24-bit then I have to agree, you do not have to "maximize" your record side as long as your average forte levels measured with an averaging meter are working at or higher than about -20 dBFS.
 
OG: "Now... There is no difference between recording the guitar at -12 or pulling it down 12dB in the final mix.”
 
BK: There certainly is, from the point of view of the sound quality of the original recorded track, in a 16-bit recording, which I think your friend is referring to. As we moved to 24-bit, I would correct that to say, "there is no perceptual difference between...." meaning that there is enough signal to noise ratio in good 24-bit ADCs to allow more latitude in record level.
 
OG: "And, you are not really losing any resolution. If you are recording at 24bits, even if you record at -40dB you still have 24bits of resolution in the data you are recording.
 
BK: Depends on what you do with that data and how you gain stage it in the final. If that -40 dB represents a pianissimo passage that is going to remain that way and not be turned up, then you haven't lost any resolution. If instead that -40 dB represents the peak level you are putting on the 24 bit recorder, then you have lost almost 8 bits of resolution! So I agree with him if his forte passages are at -20 dBFS approximately, average-reading meter.
 
OG: In my definition of "resolution", resolution and noise are directly related. The closer the recording is to the quantization distortion noise) level (the lower the recorded level), the lower the resolution of the recording, and the fewer effective bits it has. If I record to a 24 bit recorder peaking at 48 dB below peak level, I have essentially made a 16 bit recording, from the point of view of resolution, and signal to garbage ratio. Resolution has nothing to do with record level. Resolution and noise floor both change values when talking 16bit vs 24bit, but they don't really have much to do with each other.
 
BK: My definition of resolution has EVERYTHING to do with record level. You can have a high resolution tape recorder but if the highest peak of the entire program is -24 dBFS, then you have probably lost 4 bits of resolution which you will never get back. If you have to raise it later in mastering it may not sound as good as if you peaked your max. peak to full scale. The difference will be subtle with modern ADCs, perhaps inaudible, but theoretically I would try to avoid making a recording whose highest peak doesn't reach at least, say, -10 dBFS peak. This is just a conservative application of the principles.... not too low, not too high  :-).
 
OG: "When you record at lower levels, you still have the same resolution, the only thing that changes is the noise floor. With 16bits the noise floor is always 96dB down from full digital level. If you record at -10 your noise floor is 86dB below the signal. Still very good. Since audio reference levels are at -18 or whatever it is for digital audio on video for the final mix, then it doesn't matter if you record at the lower level and mix with the master fader up, or record at the higher level on each track and pull the master fader down. The final noise floor is the same. Quantizing error is a function of the converters used, and does not figure in to the recording levels or resolution. So, it doesn't matter, and I would just get everyone to agree on one way to do it so that everything is easier and you don't have to switch back & forth between methods."
 
BK: I disagree with his stating that resolution is the same when the signal is closer to the noise. The closer the signal to the noise, the less resolution that low level signal will have, at least in terms of signal to noise ratio. Yes, the noise floor is the same IF in the end case you do not raise the gain. If the original soft signal remains reproduced soft in the final, then you have not worsened the perceived noise of the final product.

Now, if the original A/D and recorder was 24 bit, there is considerable leeway in how far you pushed the original track, absolutely, but it is still true if the original track was recorded rather low, and for esthetic reasons you must raise the level of the mix fader----then the level of the original quantization noise and distortion of of the original A/D conversion gets raised above the  noise floor in the final mixdown. This is true whether you have a floating point or fixed point mixer.

 
Hope this helps,
Bob
From: Dave Kirkey
 
Ok, so, I have time to read a little more information on your site and emails and it brings me to a pretty important question, (could be a dumb question but, the outcome could be important). I know you talk a lot in respect to levels, matter of fact, I believe you talk about not over compressing and over driving mixes or recordings, and in your book on written for the TC Electronics and the Finalizer you mention not going over -12db average, so, that brings me to the question that you said to mix as close to 0 as possible.
 
Hello, Dave.
 
Well, first, take a VERY DEEP BREATH. This is not a subject that is easy to explain at first, for a novice. It sometimes takes engineers who have been working for years a few days to puzzle it out. If you are just getting started, then it could take you a few weeks to puzzle it out. Knowledge is power, and knowledge comes with study.
 
So.... here is a start. Take a deep breath again, work patiently through this and you will become a better engineer because of it.... Here goes...
 
I never said "mix as close to 0 as possible", exactly. If I did, then I would like to know where I said it because I should correct that.
 
In my article "Levels Part II: How to Make Better Recordings in the 21st Century", I cover this in great detail. It describes an advanced method of metering and monitoring and in reality, it will not be necessary to mix with peak levels as close to the top as possible if you are using an RMS metering scheme. The RMS metering scheme I propose places the ZERO at 12, 14, or 20 dB below the top... In summary of my article, the principle is to just work to that ZERO and ignore the peaks for the most part. You have to see such a meter in action to realize that the RMS levels are far below the peaks, and the meters that you have been accustomed to seeing only show you the peak levels, which tells nothing about the story, or about how much compression you are applying!
 
I know there is a difference in the metering of Analog vs Digital and I think there is a difference in references in actual meter calibration, somewhere I read that a -12 digital is actually 0db on an analog meter, is that correct?
 
This varies all over the place. There is no standard. I discuss this in my article "Levels Part I", but as a novice I don't suggest you even worry about this question if you are mixing totally digitally. Only get involved with this if you have to set the gains of an A/D converter or use any conversion between the multitrack and the mixdown recorder. I think you are mixing totally digitally. Someday you can revisit this question.
 
I have done mixes in the past, used the Waves Ultra Mix 16 bit master resolution setting and the final results are either over compressed, or, the songs sound dull and low.... I know in mastering you can do wonders, but, that all takes time, I want to send you a set of songs that is right from the start!
 
That's great! Yes...you want to do the best. I have had the best work from mixing engineers who work in my proposed K-20 format. Or who mix on an analog console with VU meters and adjust it so 0 VU with sine wave yields -20 dBFS digital---which is essentially the same as K-20. This approach largely frees you from questions of too much compression; allows you to concentrate on getting a great mix, and then in the mastering we can work wonders. K-20 means the following:
 
---Your monitor gain is very high, sufficiently high so that you will not be tempted to use compression to "make it loud", but only for esthetic purposes or for part of the mix. The method of calibrating the monitor gain is described in Levels Part II. If you don't want to bother with that...then work to the K-20 meter and your monitor will probably fall in the right spot.
 
---You are using a meter which has an RMS zero at 20 dB below full scale. Are you using a Mac or PC? Do you have a sound card? There are a couple of such meters available and if you let me know what you have, I can point you to where to get one at reasonable cost. I think it is an essential tool for you at this point in the game. There are only a few experienced engineers who do not need such a meter, but it really helps, anyway.
 
Anyway, if you send me one of your first mixes, I'll be happy to give you pointers on how it's sounding. Yes, we should wait until you've got this concept under your hat.
 
We are starting our premixes on a project and can you direct me as to the signal level with the yamaha board that I should drive the output to? It seems when I start to push to 0 on the digital console meters things get a bit odd sounding.
 
It's hard to diagnose this at a distance. This could be because you are using a bus compressor and pushing it too hard. I would take the bus compressor away and mix without it. It could be because your D/A converter for monitoring doesn't have enough headroom.
It could be because you are using some of the compressors in the Yamaha and they are not particularly good sounding. It could be a number of possibilities.
 
Do you want me to push the mixes to "Digital 0" in reference to the Yamaha board?
 
In a single word: NO. Instead, I would like you to get a metering system that is separate from the Yamaha board, since as a relative mixing novice I think a K-20 or K-14 meter will help to guide you tremendously.
 
Yes, I have found the bypass of the Finalizer, set the sample rate converter off and set the dither to 24 bits, that mode was also verified by TC Electronics. The computer is recording the 48k, 24 bit information using Soundforge 5.0 just fine.
 
I hate to be a bother, but, I want to be sure I am using the right level reference to mix to, I'd hate to think I am mixing at a level that will in the end not be maximized or worse, distorted.
 
Finally, I have asked others the same question, even the manufacturers and you can't believe how many answers I get, I don't think any of them have been the same...why the heck is this question so difficult?
 
Because it is complex, and requires good education. I've written a book on mastering and it's over 300 pages. I'm trying to simplify this concept into one web page!
 
Yes, I want to send you a song or two to have a listen, but first, I want to get a sound and a mix that I feel is finished enough for you to hear.
 
I'll check it out when you have it done. Be patient with yourself, it takes time to learn these concepts. Some people go to school four years to learn this stuff well,
 
Bob
From: Jim Schley-May
10/28/00
 
My comments are:
Bob,
 
What is the precise and accepted definition of 0dBFS? Perhaps this is common knowledge, but I haven't seen a definitive reference.
 
Definition 1: the mathematical evaluation of root mean square on the signal, with it's values normalized to the range of +1 to -1. This method yields a result of -3dB for a full scale sine wave and 0dB for a full scale square wave. Sound Forge uses this method.
 
Definition 2: as in definition 1, but raised by 3dB. This method yields a result of 0dB for a full scale sine wave, and +3dB for a full scale square wave. Cool Edit Pro uses this method.
 
Which is it? I'm anxious to try calibrated monitoring levels as you've recommended, and I want to roll my own pink noise reference.
 
Thanks in advance for your reply, and I really appreciate your ongoing efforts to educate the audio world.
 
Hello, Jim...
It is a standard, as set forth (I believe) in AES-17.
Sound Forge is not following the rules of the standard AES-17 as set down and they are in error from the official standard by 3 dB. A couple of manufacturers have made this serious mistake. Basically the rule is as follows: The 0 dB reference for either peak OR RMS measurement is that of a sinewave at full scale. Or, to put it another way, if you wish to work with RMS measurements, the 0 dB reference for that is that of a sinewave whose peak value is full scale.
That's the way the rule works! Even if it doesn't seem logical to you; just think of it as a reference, and that it is IRRELEVANT that the RMS value of a sine wave happens to be 3 dB below its peak level. So what... you can (and the AES standard does) define your reference as 0 dB.
 
Which is it? I'm anxious to try calibrated monitoring levels as you've recommended, and I want to roll my own pink noise reference.
 
Gotcha. In that case, start with a very accurate peak-reading meter, and calibrate the sine wave to 0 dB. (measured on the peak meter). Then read the RMS value, and calibrate the RMS meter to 0 dB. This is the absolutely correct method for measuring the pink noise. At that point, if you wish to roll your own pink noise, then set it to -20 dB, RMS measured, below that full scale level.

We have a download with this pink noise signal at our downloads section!

Hope this helps,
 
Bob
From: James Trammell
 
My comments are:
 
Bob, I have some instruments I want to sample with my sampler. It has an AES input, so instead of using the cheap A to D on my sampler, I'm using my Apogee AD-1000 with UV22. I have my sampler normalize my samples as I take them. Because I'm normalizing, should I sample with UV22 off and use the AD-1000's flat dither instead? Or do you think it's ok to use UV22? I know additional processing of digital data after UV22 encoding is frowned on, but does that include normalizing? If your answer is "don't use UV22 if you plan to normalize" thats fine with me. I just want to do it right and be mathematically correct. Please keep up the good work informing us all on digital matters.
 
Dear James:
 
Hi.... Thanks for your comments.
 
It's a good idea to use the superior external A/D. That's what you're doing right.
 
But you suspected correctly. The rest is actually backwards! The last step in the chain of processes should always be the wordlength reduction, along with dither. With a 16-bit sampler you're damned if you do, damned if you don't, because you will eventually be using the samples again in your digital mixer and will eventually be adding another stage of veiling dither to it. Instead of normalizing in the sampler, which makes the sound grainy and harsh, and loses depth and stability, you should raise the gain of the source within the A to D converter until the highest peak hits zero. Then dither, then feed the sampler, and don't change the gain or process again until you have to. If you had a 24-bit sampler, you would not need to dither except at the 24th bit level, which barely changes the sound.
 
So many of the other manipulations within the samplers (e.g., pitch shifting) also affect the quality of the sound. But if you can't avoid that, that's life. Nowadays many plugin samplers use double-precision 48-bit internal calculations and internally dither to 24 bits on their output. This will do the least damage to the sound.
 
Hope this helps,
 
Bob
From: Francisco Domingo
 
Hello Mr. Bob!. It's the second time we are in contact. Now I've a doubt about digital audio transfers with different resolution. For Example: If I'm working with a keyboard with "light pipe" digital outputs (Alesis QS-8/16 bits-48Khz), and I want to do a digital recording thru one interface but with a 24 bits resolution. These 8 Bits stay without use I think?. Is this correct?. Can I transfer whithout any problem or in the soft levels could I hear some "stair" levels?. I know that in the opposite way, I mean, 24 to 16 bits, the 8 bits word is truncated. If you can help me about this topic I'll appreciate it a lot!. THANK YOU VERY MUCH!
 
My regards,
 
Francisco Domingo.
 
Hello, Francisco.
 
You are correct, the top bits get transferred, and the bottom 8 bits become zeros. You will not hear any "stair" levels... it's just a 16 bit word riding in a 24-bit space.
 
Hope this helps,
 
Bob
This is a cautionary true story from the naked city.
 
Yesterday I mastered an acoustic music album for a fine instrumental artist who tried to do it right the first time. He went to a recording studio to get it done (I think it was a "project studio" because they used Cubase), and then went to one of the best-known mastering houses (lots of gold records on the walls) to get it mastered.
 
The artist was dismayed. The master which came from from this top-notch mastering house was (in the artist's own words) "to my ear, the tone seems "cold" and somehow less compelling than what is available even on the mix-studio ref. In some cases there seems to be a loss of "presence". It is not clear if the problem stems from the method of transfer, or from the method of dithering."
 
As you can see, this artist has educated himself on the vagaries of digital. He even asked the mix house to supply 24-bit AIFF files on CD ROM from Cubase. He asked me to consult on this and see what I could find. I listened to the mix reference CD and compared it to the master. I agreed that the master was grainy and unresolved---I agreed that even the original mix reference CD sounded clearer and more real than the master he had received. I concluded that something had gone wrong in the mastering.
 
Ironically, very little was done in the mastering, according to the artist, who attended the session, no limiting, no compression, simply EQ (the artist thinks a 1 dB high shelf at 10K, which in most cases he decided he didn't like) and (I think) UV-22 dither. I am not sure if it was digital or analog EQ. Clearly this is a purist project done by an artist with good ears and by the way, the mastering engineer involved has a good reputation with acoustic music.
 
So, at his request, I began to remaster his project. I put up two of the songs on the original 24 bit AIFF files. The first thing I discovered is that all 8 bottom bits of the 24 bit mix (source) files are ZEROs! (looking at my bitscope). So, that means that the original mix house truncated the Cubase mix output somehow when making the AIFF files. That's loss number one which can never be restored without a remix, automatically his mixes have some unwanted grunge and loss of depth due to the truncation to 16 bit. Even though it was in a "24-bit container". A bitscope really helps weed out the problems.
 
I applied very simple digital eq (1/2 dB boost at 20 kHz Q 1.0), some other subtle processing, and 16-bit dither with a shape that sounded good. And I believe the sound that I got is more open, livelier, even clearer, and no colder, even warmer than the original. I'm not sure if what I did was any different than the previous mastering house. But we should consider these questions:
 
1-The original mastering house could have noted and alerted the artist that his source mix files were really 16 bits masquerading as 24. This might have been correctable at that time at the mix stage while the mix engineer still had his session, if the mix engineer had used his workstation incorrectly.
 
2-Why did the original mastering house not carefully compare its master with the original and make sure that the master sounded at least as good as and no worse than the original, even with minimalist processing? What aspect of the minimalist processing was shrinking the sound of the original?
 
I'm not perfect (far from it). If at some time you work with me and you find that one of my masters sounds worse than the source you sent me, please CALL ME and tell me about it. Maybe I made a mistake that could easily be avoided. One slip of the mouse in this digital world.... I accept constructive criticism (usually :-), and I promise I will listen to you. Anyone can get hit with the dreaded "digital bug". As you can see from this precautionary tale, it can happen to anyone.
 
Should I have called up the first mastering engineer and let him know about his problem?
I do not know him well, and I felt it was not my place to call him. But maybe he'll see this message and check his gear with a few tests to make sure that a bug has not crept into the digital system. I hope he finds his problem! We've all been there at some time.
 
 
Bob Katz

From: Richard

 

Bob,

How's it going? Hey, I have a quick question for you regarding PEAK levels at K14 or K12. I'm getting pretty good at shooting right between the two, almost like a K13 (I'm using Klanghelms VU meter aligned to -13 mostly). This seems to me like a good compromise between a musical sound and the fact that one way or the other I will have to limit or hard-clip a client reference to about -9 on the VU meter to sound in the ballpark (until it goes to mastering of course).

 

One thing that I'm noticing though is that the highest peaks on kicks and snares sooner or later do hit over 0dbFS. I tried to go a little lower (arriving at K14) and it still is there. The only remedy is to put a limiter on the drumbus and take off a db or two.

 

What gives? Shouldn't I be able to run K14 and not run into overs? Do you see that issue? Do I have excessive transients that I should be more careful about?

 

Hi Richard,

 

Depends on your goals. Of course if you are working in 32 bit floating point, then for the time being the question is academic. If you are trying to make a nice mix, then it's not too much of a problem because you can give the floating point file to the mastering engineer and he should be smart enough to know what to do with them. However, you may not like what he does with those peaks! It will change the sound of what you are hearing. Remember that your DAC is clipping if your file is going over, even though the file is 32 bit float, so you are hearing the "bad" results of those transient peaks clipping. If it's strictly occasional percussion peaks the clipping may sound benign. It is only when you have to meet the real world, specifically, conversion to AAC, that the overs have serious meaning in this narrow case of occasional "innocent" percussion peaks going over.

 

However, if you are trying to make a finished master, then good fixed point behavior is your goal. And with very clean material that has not been peak limited sometimes percussion peaks can occur that are higher than 14 dB above the 0 point. And if you decide to peak limit it to get a higher RMS you may or may not like the sound of the result!

Inspect the true peak with a true peak meter. If it goes over zero, really if it goes over -1 dBFS, then you should be worried about AAC conversion or bad behavior by DACs, SRCs and other systems. At that point you have to make a wise decision as to whether you should reduce your overall level (which would be the nicest thing) or add a peak limiter to soften the true peaks (which can easily degrade your sonic quality, by losing transients and somewhat reducing the soundstage depth and imaging). That's your tradeoff if you are making a master at this level with very little peak processing.


Hope this helps,

 

Bob

Name: Richard Furch

Message: Bob, I have a quick question after aligning my monitors for the K system you outline in your (outstanding) book. I'm a mixer though. I don't want to waste your time, so here goes:

1. I run ProTools with Apogee converters aligned at -20dBfs=0VU (I'm looking for more headroom than normal). I inserted a signal generator plug in at 0 on the fader and Pan to one output at -20dBfs.

 


Don't trust the signal generator plugin. Unless it is guaranteed RMS-calibrated, it will probably be less than accurate. Download the -20 dBFS RMS pink noise from our website. It's a stereo file. Play one speaker at a time and don't play with pan pots, play it in stereo but mute one speaker at a time and tell me how different that is than your own generated pink noise.

 

I aligned the monitors with Pink noise/C weighted/slow/one channel at a time at 83 dB SPL and marked the position. (a C24 with db readout).

2. I checked that 6 dB down from that actually meant 77dB SPL and it did.

3. So far so good (I think).

 


Assuming it was accurate pink noise then you're right "so far so good". Except if your monitors are very close then they can end up 3 dB OR MORE louder than mine regardless of the accuracy of your calibrations.

 

4. Using K14 (6dB down)

 

Meaning that you set your monitor gain as in the paragraph "I aligned..." and then left it at that monitor gain (6 dB down from the 83

setting) and mixed to that, right?

 

I mixed a contemporary rock/pop singer songwriter track by ear only (without checking the meter too often).

 

A good goal!

 

5. Mix finished, I'm at -8 to -10 VU, even though I would say that sounded pretty loud while mixing (probably a 2 or 3 dB louder than I would normally mix).

 


What was the peak level of this mix?  How far did its highest true peak get on the peak meter? You're saying that its VU level (on which VU meter?) is what?

Now in the end, regardless of the miscalibrations or assumptions, your method of monitor calibration and working to that monitor gain while largely ignoring the meters will probably produce a good, clean mix that sounds great and is VERY ready for mastering. It may not be ready to send to a client who expects something hotter, but it is ready for the next stage and you can never say you ruined the mix  :-).

Now for possible explanations and tools to trying to get you closer to your stated goal of producing a true K-14 mix:

First of all, try to get the LM5D meter from TC for your Pro Tools Rig. If not, the UAD limiter has K-metering, for example. Could be a combination of the inaccurate pink noise and the position of your loudspeakers. When you said "I'm at -8 to -10 VU" what meters are you using to measure that? Are those accurate VU meters that were calibrated for -20 dBFS = 0 VU?

 

6. Confused, I played the same mix releveled to 0VU

I'm curious, using what meters?

(obviously I still have miles of headroom in the DAW). At 0VU I could positively not take the volume in my studio for more than a minute (and I have a hiphop/rnb background as an engineer check my site emixing.com).

 

7. What gives? At the 83dB alignment I can positively not get to a 0VU mix without killing myself.

 

You betcha!

Trying to put things in perspective, keep in mind that your 45 inch speaker distance completely changes the landscape, raises the subjective loudness above the typical points by as much as 3 dB compared to, say a 9 foot speaker distance. That, combined with a possibly mis-calibrated pink noise signal and we're probably on different planets.

 

First, compare your pink noise signal and method with mine, just to see how far off we are.

 

Then, assuming your pink noise signal was accurate, I'd say a true K-14 mix (forte passages at K-14 Zero) with your speakers at the 45 inch position will need to be reproduced at around -9 or -10 dB! Compared with my -7 to -9 dB for the same musical source. Keep in mind the "-6 position" for K-14 was based on the theory of a mono speaker and mono pink noise, but two stereo speakers raise the loudness 2-3 dB so a true, conservative K-14 at about 9 feet speaker distance will probably reproduce around -8 or -9 dB on the monitor gain, so -8 or -9 dB monitor gain is the recommended gain for a K-14 with 9 foot speaker distance.

 

And then your speakers are yet 2 to 3 dB louder than that by virtue of their physical position. So a true K-14 (conservatively measured) will probably reproduce on your setup at -11 or -12 dB, which is 5 to 6 dB lower than the gain you were running!!!

 

Does this help?

 

You have done well and you're using the right language and speaking clearly, just fine so it's just a matter of sorting it all out. I must say that using a hand calibrated monitor gain is iffy in itself, you cannot easily move it off that -6 position and readjust it to, say -9 without jumping through hoops. Ideally your monitor controller should have 1 dB steps so you can help debug this. With the monitor set at a fixed -6 dB position I'm trying to get a handle on what you meant by "-8 to -10 VU", whether that was on peaks or average, loudest passage, etc. etc. etc.  It gets complicated to debug. We'll sort it out, be patient.

 

Bob

Social