Here is a link to the RX Audio Cookbook for spectral audio correction:

Here’s a copy on the home page as well.

Welcome to the iZotope RX Audio Cookbook

The iZotope RX Audio Cookbook is designed to help you find solutions or ‘recipes’ for solving common audio problems. Each recipe includes step-by-step directions, before and after audio examples, illustrations and tips for solving stubborn problems.

As you work through the Cookbook, be sure to send us your feedback.

This info comes from Geoff the Grey Geek site.


Understanding Decibels (part 1)

Decibels are widely used in audio, and often misunderstood. These articles give a practical understanding on using decibels in audio work.  But first, some basic questions and answers about decibels.

What is a decibel?

A decibel is a tenth of a Bel, a unit of level, named after Alexander Graham Bell. A Bel is a very large unit, so the prefix deci (one tenth) is used. A decibel uses a logarithmic scale, not a linear scale like volts or watts – see below.

There is no absolute level called a decibel. A decibel is a ratio. It is relative to something. Unfortunately what it is relative to is often not mentioned when stating a decibel reading. For example, a washing machine might be said to have a loudness of 70dB, this usually means the noise it produces is 70dB higher than the threshold of hearing (0dB).

Why Decibels?

So why use a decibel? You might have noticed that the volume control on most HiFi amps is marked in decibels, as is the markings on mixer level controls. This is because our range of hearing is so vast, that to use a linear scale we would be using figures from 0 – 1,000,000!

Decibels are non linear

A secret to understanding decibels is to note that decibels are non linear. Another example of a non-linear relationship is between the side of a square and the area of a square.

decieble is non linear like square side and areaIn this example, you can see that increasing the side measurement does not have an equivalent increase in area, but a larger increase.  Also a doubling in the side length does not double the area – it is a lot more than the double! This is an example of a non-linear relationship – in this case, a small increase in the side relates to a differnet increase in the area. Decibels are similar. A small change in decibels relates to a different change in the ratio of the two levels being compared.

Decibels 101

When we are talking about audio levels we are looking at voltages, or sound wave amplitudes. (Note: power measurements (like the power of an amplifier or speaker) use a different formula). But without getting into formulas etc, we need to accept the following summary of linear ratios and decibels for voltages. (I’m not showing the formulas or calculations because I reckon most people skip over them anyhow, and if you like formulas any search engine will give you as many as you like)

Enter the voltage ratio to calculate the equivalent decibels

Enter decibel value to calculate the equivalent voltage ratio

Example One: you increase the volume on a mixer or amplifier by 6dB: this is actually doubling the voltage that will appear at the output (because 6dB is a factor of 2 – see table). So we could say the level is now +6dB, That is, relative to what the level was, it is now 6dB higher. Remember decibels are always relative to something, in this case, to what it was before we doubled (added 6dB to) the level.

Example Two: same scenario, but this time we lower the volume by 6dB. This means we are halving the voltage (because 6dB is a factor of 2). So we could say the output is now -6dB (below what it was before).

These examples also show that it is possible to have positive and negative decibels. If the voltage is +6dB, it is twice the size as it was before. If is it -6dB, then it is half the size it was before.

I know it can be hard, but stay with me here, a few more examples should help.

Let’s say we have a microphone that is giving out 1 millivolt (that is, a thousandth of a volt) when we speak into it. We plug this microphone into a pre-amp (microphone input on a camera or mixer) that has a gain of 60dB. From the table above, we see that 60dB is equal to a linear ratio of 1000. Therefore a pre-amp with a gain of 60dB is going to amplify our 1mV mic level signal by a factor of 1000, giving us a level of 1,000 millivolts or 1 volt. Behold, we now have a signal which is at line level (see article on Audio Levels).

Consequently, we could say that the level of the microphone signal is -60dB compared to the amplified line level. That is, the input is a 1000th of what the output of the pre-amp is.

In practice, most microphones will produce about 10 millivolts (when spoken directly in to), so the signal only needs to be amplified by a factor of 100 to make it up to 1 volt (0.01 x 100 = 1.0). The table above tells us that a factor of 100 is 40dB.

Most mic pre-amps should amplify the signal by at least 40 dB (x100) and some will amplify up to 60dB (x1000). Some video cameras (semi-professional or professional) will have a switch which allows you to switch the gain between 40dB, 50dB or 60dB gain for the preamp. The other way of stating this is to say the input is selected for -40dB, -50dB or -60dB levels. Just remember it is all relative, and the greater the number, the greater the amplification.

Have a look at the manual for your HiFi, mixer or video camera, and find out what level your mic input it expecting. If you are putting more in, it will probably cause distortion.

If for instance you are trying to connect a mixer (PA board) tape out, into your mic input, what do you need to consider?  Well, first we need to know what the output level is likely to be. Let’s say that the specifications say the tape out (line out) is  -10dBV or 316 millivolts and your mic input is looking for -50dBV (or 3.16 millivolts).  This is where decibels makes it easy. We have a level of -10dB, and need to get it to -50dB, easy, use a 40dB attenuator to reduce the level. That is easier than calculating the ratio between 316mV and 3.16mV, (which when you calculate it happens to come to a linear ratio of 100 or a logarithmic ratio of 40dB).

Again, a 40 dB attenuator is indicating a ratio, in this case the ratio between the signal in and the signal out is 40dB, or a factor of 100. That is, the output is one hundredth of the input. If you want to look at other voltage gain or attenuation ratios and decibels, have a play with my Decibel Calculator for Audio

Decibels do make calculations easy, but it can be difficult to comprehend.

In the first article we saw that a decibel is not a fixed unit, rather, it is a ratio between two levels.  A decibel reading on its own is meaningless if it isn’t referenced to something. Unless it is spelled out, that reference something is normally 0dB. The mic input on a camera or mixer might say -50dB, that would mean it is 50 dB lower than 0dB. The line out of a mixer or DVD player might be -10dB, meaning it is 10dB lower than 0dB. The stereo out of a mixer might be +4dB, meaning it is 4dB above 0dB.

What is 0dB?

In simple terms, 0dB is the reference level. It is the level that is being referred to. When the measured level is neither above or below the reference level, then the difference in level is obviously zero, hence our reference level is called 0dB. So what is this level? In most situations with audio, 0dB will refer to Line Level. That is, the nominal level that comes from a mixer, keyboard, camera, or DVD player etc.

What level is Line level?

This is the $64,000 question, because there are various levels that are referred to as line level. A good specification table for a camera, amplifier, mixer or recorder will always say what voltage level they are using as 0dB. There are several conventional line levels that can normally be used if no other information is stated.  The secret is to pay close attention to the letter after dB.

  • 0dBV refers to 0dB being one volt (1V)
  • 0dBu refers to 0dB being 0.775 volts
  • 0dBm refers to 0dB also being 0.775 volts, but the power formula for decibel calculations needs to used, meaning 1dBm is close to 2dBu.

Don’t get too bogged down in all this, just make sure you compare apples with apples. That is, if you are looking at two cameras or recorders or mixers and comparing the microphone input sensitivity, make sure both specifications are dBV, or both are dBu or dBm. Before we look at some real world examples, we need to appreciate that 1 volt is a high level for line levels, and not commonly seen in consumer electronics. The “standard” line level in consumer electronics is normally -10dBV, which is 316 millivolts or 0.316 volts.  That is why the tape out of a mixer is often -10dBV. The same level is used for line in on digital recorders, TVs, VCRs and cameras with line in.

Camera Specifications

The popular Panasonic GS500 (similar to most cameras) states that the audio output level (line) is -10dBV (316mV). This is good because this matches the standard consumer line level found in VCRs and TVs. So when we plug the line out from the camera into the line in of a recorder or TV, the levels match and all sounds good.

The specifications also say “Mic input: mic sensitivity -50dB (0dB=1volt)”. Knowing the reference is 1 volt, and that 50dB is a ratio of 316 (see table) we can calculate the mic input is expecting 3.16 millivolts (1 divided by 316). This is about the level from a normal microphone so all will be good when a microphone is plugged in.

When a sensitive or high output microphone (eg Rode Videomic) is used, it is useful to reduce the output on the microphone so as not to overload the camera input. Hence most users switch in the 10dB attenuator on the Rode Videomic.

Mixer Specifications

A common mixer used in churches and schools is the Yamaha MG series. Its specifications are also representative of many mixers. A quick glance at the specifications reveal both dBu and dBV are used. However at the bottom of the page, they do conveniently state that 0dBu=0.775 volts and 0dBV=1 volt. They also helpfully give the equivalent voltages alongside each dB value. Let’s look at some of the specifications that will interest us if we want to connect to a mixer output.

The stereo out has a nominal output of +4dBu (1.23volts) with maximum +24dBu (12.3volts). It should be obvious that we shouldn’t connect this straight into the mic in on our camera! If we need to connect to this output, we would need to reduce the level by at least 50dB.

The better output might be the tape out. The specifications state “Record out: -10dBV (0.316 volts)”. This is closer to -50dBV (3.16mV) but still 40dB higher than nominal level for the mic input . So we need to reduce even this output by 40dB, or by a factor or 100. The other common output often used from a mixer is the headphone output. Some calculations reveal this can have a level between -9dBV to +4.77dBV, depending on the signal level and headphone volume control. This should be used only as a very last resort, and then still using at least a 40dB attenuator.

40dB attenuator

Having a range of attenuators, or a variable attenuator is useful. However a 40dB attenuator will assist in most situations. So what is a 40dB attenuator, and how do you make one? Simply put, a 40dB attenuator reduces (attenuates), or divides the signal by a factor of 100. Most music shops will sell what is often called a DI box. The DI stands for Direct Input, or Direct Injection. Either way, it is simply a box which allows instruments (like keyboards and guitars) to be directly plugged into the microphone input of a mixer. These DI boxes also change the signal to a balanced line to be able to run a longer microphone lead (more on this in a coming article). Many DI boxes have a switch which will enable you to select -40dB, -20dB or 0dB attenuation.

If you need to buy a DI box, here is a link to some of Amazon’s range of DI Boxes.

If you are trying to connect from a mixer (line out) to a video camera or laptop with only a Microphone input socket, then it is easy enough to build a simple 40dB attenuator with two resistors, (a 10K and 100 ohm resistor) and a capacitor. The circuit looks like this:   understanding decibels, 40dB attenuator

This circuit simply divides the input which makes the output 1/101th of the input. It can be made up by anyone who knows a little about electronics and can solder. Similar circuits appear in any number of forums with slight variations. They should all work. The capacitor is used to block any DC voltage coming from the camera or from the mixer, it is not always needed.

Here is a drawing of how you might use this as an attenuator built into a 6.35mm (1/4?) plug (this plugs into the mixer – you’ll need a 3.5mm (1/8?) plug for the mic input). Don’t forget to fit the plug cover on the cable before soldering the resistors and plug on. This doesn’t use the capacitor, but you could fit it in the plug at the other end of the cable if required.

If you want make an attenuator with a different attenuation, use can use my simple decibel calculator for audio.

I hope this has helped in understanding levels and decibels, and the why and how of connecting line level to microphone inputs.

Funny story – I am transferring a bunch of old cassettes ( ! ) to my computer, and I came across this tape of my old friend Dave’s band.  On the flip side was a recording from WERS radio (Emerson College) with a bunch of circa 1983 songs.

There was this song called Pussy Whipped, and it sounded vaguely familiar.  I typed in the main phrase, and came up with this video.  Pretty funny.

I knew it had its roots in Gospel Music, but believe it or not I never took heed of most of the lyrics.

Oh, when the saints Go marching in
Oh, when the saints go marching in
Oh how I want to be in that number
When the saints go marching in
Oh, when the drums begin to bang
Oh, when the drums begin to bang
I want to be in that number
When the saints go marching in
Oh, when the stars fall from the sky
Oh, when the stars fall from the sky
I want to be in that number
When the saints go marching in
Oh, when the moon turns red with blood
Oh, when the moon turns red with blood
I want to be in that number
When the saints go marching in
Oh, when the trumpet sounds its call
Oh, when the trumpet sounds its call
I want to be in that number
When the saints go marching in
Oh, when the horsemen begin to ride
Oh, when the horsemen begin to ride
I want to be in that number
When the saints go marching in
Oh, when the fire begins to blaze
Oh, when the fire begins to blaze
I want to be in that number
When the saints go marching in
Oh, when the saints go marching in
Oh, when the saints go marching in
I want to be in that number
When the saints go marching in.

Often the first two words of the common third verse line (“Lord, how I want…”) are sung as either “Oh how”, “Oh, Lord” or even “Lord, Lord” as cue notes to the simple melody at each 3rd line.

Arrangements vary considerably. The simplest is just an endless repetition of the chorus. Verses may be alternated with choruses, or put in the third of 4 repetitions to create an AABA form with the verse as the bridge.

One common verse in “hot” New Orleans versions runs (with considerable variation) like thus[citation needed]:

I used to have a playmate
Who would walk and talk with me
But since she got religion
She has turned her back on me.

Some traditional arrangements often have ensemble rather than individual vocals. It is also common as an audience sing-along number. Versions using call and response are often heard, e.g.:

Call: Oh when the Saints
Response: Oh when the Saints!

The response verses can echo the same melody or form a counterpoint melody, often syncopated opposite the rhythm of the main verses, and a solo singer might sing another counterpoint melody (solo soprano or tenor) as a 3rd part in more complex arrangements.

The song is apocalyptic, taking much of its imagery from the Book of Revelation, but excluding its more horrific depictions of the Last Judgment. The verses about the Sun and Moon refer to Solar and Lunar eclipses; the trumpet (of the Archangel Gabriel) is the way in which the Last Judgment is announced. As the hymn expresses the wish to go to Heaven, picturing the saints going in (through the Pearly Gates), it is entirely appropriate for funerals.

The first one is an oldie, yet awesome. Paul is definitely on Mary Jee Wanna

This is a well done eye-opening look at what happens to us all. Some less fabulously than others 😀

  1. Morgan Bay (Dexter's Song) Steve Keith / Darren Garrett
  2. East St Louis Toodle oo Steve Keith
  3. Fall Apart Blues Steve Keith
  4. Words @doesdwars, @kiwichrys, @steve2k2
  5. Well Wishes @kiwichrys, @thelonewulfproject, @steve2k2
  6. Every Night @kiwichrys , @steve2k2
  7. Harvest Moon @mojoespage, @kiwichrys, @steve2k2
  8. Always Remember Us This Way @kiwichrys, @steve2k2
  9. I Can't Make You Love Me @steve2k2, @cocobark, @kiwichrys
  10. Night From Day @eungang, @steve2k2
  11. The Other Day @eungang, @steve2k2, @textured_grooves
  12. San Tropez @eungang, @steve2k2
  13. Lucky Man @eungang,@steve2k2,@textured_grooves,@the_mooney_blues
  14. From the Beginning @steve2k2, @thelonewulfproject
  15. The Nameless Heights @eungang, @steve2k2, @textured_grooves
  16. Norwegian Wood @steve2k2, @eipi, @eungang
  17. Imagine @steve2k2, @eipi
  18. Wild World @steve2k2, @eipi, @eungang
  19. Dance Class @paranoidcitizen,@gbailey,@steve2k2
  20. We'll Get By @jannasyn,@steve2k2
  21. Vibe Girl @steve2k2