a.) Focusrite Saffire Pro40 audio interface (although you can use any audio interface provided it has at least two microphone preamp inputs)

b.) Rode NT1A Condenser microphone
c.) Reaper Digital Audio Workstation
d.) Reaper plug-ins (free along with Reaper)
e.) Focusrite bundle plug-ins – this comes free if you buy Focusrite audio interface.

This tutorial assumes your classical guitar includes an active pickup for additional DI recording and that you have a fully working DAW (digital home recording studio).

Step1.) Position the Microphone in the Quite Live Sounding Environment

It is important to put the microphone somewhere in the center of your room(away from walls or corners). It would be much better if you have a fairly large size room so that the microphones can naturally capture the reverberations.

In this tutorial, a sample classical guitar piece will be recorded in a 10ft x 15ft room with tiles but this is also a usual bedroom with furniture. If you want to know if the classical guitar would sound nice during recording; try to play it live in your room without microphones and check the ambiance and feel. If it sounds good, it would also sounds great during recording.

The Rode NT1A condenser microphone is position at this level:

Make sure the condenser microphone would be directly facing the guitar (not necessarily the sound hole) when the guitarist would be performing.

Step2.) Connect guitars and microphone to audio interface

The next step is to connect your guitar pickup output to the first microphone preamp input of the audio interface. And then connect the condenser microphone to the second microphone input. This is how it looks like in Saffire Pro 40 with the audio interface settings for recording:

Input 1(pickup output) has a low gain setting of 2 while Input 2(condenser microphone output) has a moderate gain setting of 5. This is because guitar pickup output is already strong and needs little amplification by the microphone preamp. Meanwhile condenser microphone output is still very weak and needs more gain.

+48V phantom power and instrument signal button (if supported by your audio interface) should be enabled for condenser microphone to boost weak signals from both instruments.

Step3.)Launch your recording software and configure to record

The recording software used is Reaper. You need to make sure that your software supports simultaneous/multichannel recording. In this case two tracks are added to the multi-track (one for condenser microphone output and one for guitar DI output). In advance, the first track is panned 30% left and the other is panned 30% right to have that thick sound. Then both are enabled for recording:

In this tutorial, Reaper is configured to record at 24-bits/48KHz (recommended).

Step4.)Position the guitarist in front of the microphone

The next step is to position the guitarist. Basically the guitarist would like to sit in classical guitar position. Aim the microphone face somewhere in the upper neck, not in the sound hole because it would sound boomy:

Then make sure the distance from microphone to audio interface would be around 6 inches at least. See the top view:

Step5.)Hit the Record Button and make sure they have similar levels

As you have seen above they have comparable recording levels; if the output from condenser microphone is too low (less than or equal to -48dB) then you need to increase the gain and re-record again. However ensure that the maximum peaks should not exceed -6dB and there is no clipped signal (red, beyond 0dB).

Step6.) Manually adjust the levels in Reaper until they have same levels

Listen carefully, if the first output (DI from classical guitar) sounds stronger than the microphone output. So lower down the DI output (which is in track 1) until the volume level is the same with track 2(coming from condenser microphone).

Track 1 is set to -6dB and Track 2 to 0dB. In your recording, you might have different dB settings, so use your ear to get the proper adjustments. You can solo each track and compare the level in volume for more accurate monitoring.

Step7.) Apply Effects

There are only two effects used, Reverb and EQ. This is how they are added in Reaper:

The purpose why the reverb comes before EQ is to have the EQ; filters the undesirable high end caused by the reverb tails. Take note the EQ settings strongly depends on the material, you may need to use a different EQ setting as shown below. But you can start with it, and then start tweaking until you get the best sound.

These are the EQ settings (using ReaEQ plug-in):

DI output (Track 1):

High Shelf 4000Hz, -3dB cut
1000Hz, -3dB, bandwidth 2.0 (a cut in the mid range since sound from guitar pickup has a lot of mid-content).

Microphone output (Track 2):

High Shelf 4000Hz, -3dB cut
1000Hz, +2.5dB, bandwidth 2.0 (a boost in the midrange since sound from condenser microphone has a lower mid-frequency content).

These are the reverb setting: Focusrite Reverb plug-in: Using medium jam room presets

This is the complete recording example:

Preparing for the recording session a day before

The most important thing is that you are prepared with the recording session. Supposing you have a friend that would like to record in your own studio; you have to check the following a day before the actual recording:

a.) Are the number of microphone and guitar cables sufficient for the recording session?

b.) Have you double check the computer for some software related issues? Of course, if you will not do this. The worst thing that could happen is that your computer might go down on the actual day of the recording.

c.) Are all recording gears ready, installed and working? For example audio interface, microphones, mixers, tuning of the drums, etc.

This preparation may sound like a hassle but I do this all the time, to make sure that the recording session would be as smooth as possible.

Minutes before the Actual Recording Session

On the day the artists/producers would arrive, I do not immediately proceed to recording but I would rather talk about the expectation of the project. For example, how many songs are to be finalized and recorded? Also I would like to ensure that the artist and producers are in their good mood as possible.

So it’s not bad to give them a cup of coffee, tea, snacks, etc to refresh them. Also I would encourage the producer to rehearse MANY times before doing the actual recording.

One issue I observed is particularly with the singers not being ready for the session. They need some time to rehearse their vocals before the actual recording. Professional singers do some vocalization, review of the lyrics, tonal/pitch review, style review, etc. All of this work will be done by the artist/singers with the advice and instructions from the producer.

Then finally once I received a go signal from the producer to start recording. I would advise to make a test recording just for matching the levels (avoiding clipping).

Once the musicians are starting to play their instrument and monitored by your gears; I would carefully watched the following:

a.) Level meters on the audio interface (NOT the software/DAW level meters). Level meters on the audio interface shows you exactly how much level would be captured and to be converted to digital. To observed proper gain structure, I would be recording around -16dB to -6dB at most. For example this is the screenshot of Saffire Pro 40 audio interface (the one that I am using) level meters on the hardware:

It shows a captured level of -6dB. This value might be different as shown by your DAW recording level meters such as this one in Reaper:

b.) Of course the most important is that there is NO clipping in both audio interface/hardware level meters and the software level meters. If clipping occurs; adjust the gain in your hardware before officially hitting the record button.

The actual recording session

It’s good to motivate the musicians that they sound great in the rehearsal. So that this positive reinforcement should work in their minds and that they would at their best in the actual recording. Finally, be patient, your job as an engineer is not to determine that the artist is ready or not for recording. It’s the job of the producer. The producer would tell you that the artist is now ready; so hit the record button.

Again, I would watch and observed the following during the recording session:

1.) Are all gains properly set? I would look at the corresponding level meters to make sure they are within the -16dB to -6dB maximum and not clipping.

2.) I would as well look at the captured waveform to make sure no silly dropouts.

3.) I would also listen to the artist while they perform but I won’t look at them. I instead face the level meters, hardware and software settings to make sure they did a good job in capturing the excellent artist performance.

The art of re-recording (chaos is normal)

It’s normal for producers and artist to re-record a specific track MANY times. It’s because they want to make sure the recording should sound great according to their expectations. Sometimes the artist would make a mistake and then that section would be re-recorded. Sometimes also this mistake can be emotional and can be chaotic to the band/producers involved. The good news is that this is normal for all recording session.

Be positive, your job as an engineer is to make sure that perfect performance has been captured. So do not complain that the artist or musicians are not competent because they always re-record. It’s normal. Remember that you get paid by the hour.

Be creative with re-recording. Sometimes producers do not know a shortcut and they would like to re-record everything. This can be fatiguing to the artist/musicians. For example a certain guitar solo is still undecided after hours of re-recording. The main problem is that the producer does not seem to get the right sound appropriate for the song. So the guitar solo would be re-recorded a lot of times until the desired sound would be achieved. The guitarist fingers would burn out from repetitions. This is very inefficient.

Your job as an engineer is to suggest efficient recording process while not compromising quality. So I would suggest a reamping recording approach to lead guitar recording and I would strongly recommend this technique to the producer.

It is because with reamping; guitar solos are recorded clean (no effects). And then once it’s done. The recording producer, engineer can work together to get the perfect sound without having the lead guitarist to play/re-record the guitar a dozen of times. See the reamping process below:

This blog receives a lot of inquiry from interested readers on how they can create a short demo that can sound as good as possible. Although it varies a lot in recording equipments or gears, the techniques and processes stays the same. This will be illustrated in this tutorial.

What Gears do you need?

1 External Audio Interface (USB or Firewire) = with at least two pre-amp inputs. This audio interface should be capable to perform multi-channel recording and will be using ASIO drivers (most audio interface drivers used in home recording are now using ASIO).

1 Professional Vocal Condenser microphone = you will be using this to record your vocals. You can find a lot of lower cost microphones with superb capturing quality.

1 Acoustic guitar with pickup = make sure the pickup is of high quality and does not introduce noise. You can check the guitar sound thoroughly with the pickup before buying one in the music store.

1 quality guitar cable = any brand will do, a shielded cable is better for lower noise.

1 DAW (Digital Audio Workstation) = there are lot of cheap solutions out there that can bring outstanding results at less than $100 licensing fee. Do not use free solutions such as Audacity.

1 working PC configured as recording studio = if you do not have a recording studio in your home, read this tutorial on how to easily convert your pc into a recording studio.

Case Example/Illustration

In this tutorial, a sample short 15 second acoustic guitar demo of the song “Forever and for Always” (written by Shania Twain and Mutt Lange) will be created. The following are the gears used:

Audio interface: Focusrite Saffire Pro 40
Vocal microphone: Rode NT1A
Acoustic guitar: Custom nylon guitar with pickup
Digital audio workstation: Reaper
Operating system: Windows XP 32-bit

Step1.) Connect the vocal condenser microphone to your audio interface input 1. Use the manufacturer supplied microphone cables which should be XLR from end to end.

Step2.) Connect a standard guitar cable to the acoustic guitar and connect one end to the audio interface preamp input 2.

Step3.) Position the vocal condenser microphone around 6 inches to 12 inches from the singer. Use a pop screen such as shown below:

Step4.) Turn on the phantom power in your audio interface. Vocal condenser microphones need phantom power to work.

Step5.) If the audio interface has some features that can recognize instrument level signals (like Saffire Pro 40), enable them. If you are not familiar with instrument level signals; read this post: Difference between line, instrument and microphone levels.

Step6.) In Reaper DAW; insert two new tracks. Configure the first track to receive the recorded audio from audio interface input 1 (vocals) while configure the second track to received audio from audio interface input 2 (acoustic guitar). This is how it looks like:

IP 1 stands for the audio interface input 1 while IP 2 for the second input.

Step7.) Set for optimal recording levels, it is suggested you understand the concept of proper gain staging.

Typically you would be aiming around -16dB to -6dB max. Some audio interface has some level meters so use them.

Step8.) Play your guitar and test your vocals. You should be able to hear them in your studio monitors. Look at the level meters to make sure you set it right (not clipping).

Step9.) Turn your studio monitor off so that it won’t interfere with the recording and make sure the entire recording environment (your room) is quite to avoid some leakage.

Step10.) Start the recording (record at 24-bit/48KHz). In this, you are recording the demo song live. But there are two waveforms (one for vocals and one for guitar) that should be capture in your recording software (in this case Reaper), see the screenshot below:

At this point, you have completely recorded the demo.

Some Audio Mixing Suggestions

Once recorded, you can now mix them. Since mixing itself is a very broad topic, you can refer to the entire mixing tutorials included in this blog (you can use the search function or the web sitemap) for details.

But the following are the effects and settings used for the tracks:

Vocals:
EQ: Waves Q3 Paragraphic EQ
Settings: Low shelf -6dB at 150Hz, 2000Hz Q=0.7, +3dB
Compressor: Waves C4 with pop vocal preset

Acoustic Guitar:
EQ: Waves Q3 Paragraphic EQ
Settings: 1000Hz -9dB Q=1.4; High shelf: 2000Hz -1dB

The guitars are doubled in the mix and panned at 50% left and 50% right. The vocal is panned in the center. Next render at 24-bit/48KHz, see sample settings below done in Reaper:

Some mastering suggestions

Mastering is simply bringing up the volume of your track. It is also a very broad topic by itself. But let’s make it simple if you have come up with a great mix:

1.) Launch Reaper, insert one track which is the mix down (rendered file in the mix)
2.) To normalize the waves non-destructively (which will maximize the volume without clipping); right click on the waveform, go to Item processing click “Normalize Items”.
3.) You can implement volume envelopes and automation to select which sections you would like to minimize the volume. This is how it looks like:

4.) Render it again as 48 KHz, 24-bit.
5.) Finally the rendered file can be down sampled to 16-bit/44.1KHz wav using Voxengo R8brain free software. See screenshot of the settings below:

This is a sample demo recording illustrated using the above processes:

Song Credits: “Forever and for always” written by Shania Twain/Mutt Lange
Vocal credits: Jeanine Maningo

Question on Sampling Rate and Bit depth

I have a lynx aes16e & aurora 8 and using Cubase 5 on windows 7,64bit. So my question is: Do u advise me to record with 48/32bit, or 48/24bit? Since as you know, Cubase is 32 bit floating…thanks.

My answer:

Technically, 32-bit float has the same resolution as a 24-bit as I have stated in this post:32-bit-float recording bit depth vs 24-bits. The only difference is that 32-bit is a floating point system used internally by the DAW during mixing, etc.

I would advise the following:

a.) Make sure your audio interface hardware is configured to record at 24-bits (not 32-bit float). Depending on the audio interface, this can be set directly on the hardware or using the software that comes along with it (the drivers).

b.) Set Cubase to record only at 24-bits. You do not need to worry about this internally since any great DAW can process/represent a 24-bit audio into a 32-bit float data or even 64-bit float (such as used in Reaper).

Recording at 24-bits save you a LOT of disk space usage while not losing the quality you need. Again take note that once you start processing these 24-bit audio in your DAW, the software won’t be processing it as 24-bits. Instead, they will use either 32-bit float or 64-bit float to represent the 24-bit audio for better accuracy and lower errors in digital processing internally. There are some DAWS that are an exception to this (e.g. some versions of Pro tools works at 48-bit fixed internally).

If you are interested in this topic, I would suggest reading the advantages of 64-bit float DAW over 32-bit float.

c.) Much better not to alter or edit the 24-bit audio recordings “destructively” so that you would preserve the original recordings which is very important for future remixes, etc. One mistake I’ve done in the past is that I record at 24-bits but I edited it destructively for some reasons and then save it as 32-bit float.

As a result of this procedure, I could hardly trace the original 24-bit audio recordings and consumed a lot of hard drive space in the process since 32-bit float is enormously larger than 24-bits.

I regret doing this; and then I only edited waveform non-destructively using envelopes, automation techniques, etc. This will preserve the original recordings.

Question on Acoustic Guitar and Bass Guitar used

Hello,
Thank you for your informative website about (home) recording.
Your recordings sound very good. What bass guitar do you use/ and how do you record it?
And what acoustic guitar do you use?
Regards

Hi,
Thanks for liking the sound of my bass and acoustic guitar. Anyway I used custom/handmade nylon guitars. I bought this in Mactan Guitar Store, Lapu-Lapu City Cebu, Philippines in 2003. It still sounds great today.

Unfortunately, it seems they have no online store so you would need order personally. And then I use a D’addario nylon strings. This is a picture of my guitar:

For my bass guitar, I use two. In my very early recordings, I record it using Fender Jazz Bass guitar. And then now, I use Fernando Bass guitar which I bought from JB music store SM City Cebu, Philippines.

Question about Music Licensing

What license do I need to publish a songs that is already out there for a Kendall. The page may include the already published lyrics and the Mp3. Is there a license that I could get to cover me and what would it cost and would there be a royalty that I would have to pay as I sell one, should someone want to buy one? Who would I contact to find out if this is even doable?

Your situation is that you want to license an existing song to make a cover version. And then you would like to sell this cover version. Yes, you need a license for this so that you can cover the songs legally. Here are the options:

a.) Check who are the songwriters and music publisher of the song. You can find this information in the song credits. Make sure you obtain the correct contact information. This is often the hardest step. If it’s not provided on the material you have, I suggest searching on the Internet to get some information.

b.) Contact the music publisher. And ask that you want to cover the song. You need to provide more details if they ask some questions. They would need this information in the formulation of the license.

c.) If there is no music publisher information. Contact the songwriter or the artist. They would most likely know the music publisher of the song and can give contact information.

d.) Or if the song is a mainstream hit. It is most likely the songwriter/music publisher is affiliated with Harry Fox agency. Go to this page:

http://www.harryfox.com/public/Licenseehfa.jsp

You need to find out if the song is represented by Harry Fox agency and then ask for a license.

e.) Regarding the cost, I could not give you an accurate figure but it depends on the number of copies you are planning to sell as well as the background of the song (it’s obvious more expensive to cover hit songs because they are already established songs).

f.) Once the music publisher or the Harry Fox agency give you the licensing quote (the fee that you need to pay for a license), you need to pay for it and then the official license would be issued. You can start recording the song cover and releasing it as you have plan.

“Hello, 16 bit (2 bytes) can hold frequencies up to 64 KHz (65,535 bits), so I wonder why anyone would use 24 Bit when a whole byte is being wasted.

You will get the exact same results if you record at 48 KHz for both 16 bit and 24 bit. So why waste the space?

I recorded everything at 16 bit to save space and don’t see the need for 24 bit until I move up to 96 KHz.”

First, the reader ask some few questions pertaining to 24-bit recordings comparing it to 16-bit, so this is a topic relating to analog to digital conversion or PCM (pulse code modulation).

Don’t confuse bits with frequencies

In PCM, it is a standard of representing analog signals in the digital domain. This is a sampling technique; using a high resolution sampling method results in a more accurate digital representation of the analog signals.

Now to sample an analog audio to digital, your converter needs two parameters:

a.) Bit depth
b.) Sampling rate

Analog is represented by continuous signals such as voltages. After all when the sound wave hits the microphone, it is first converted into microphone levels (weak millivolts) then it will be amplified by audio interface or mixer pre-amp into line level signals. Line level signals are stronger voltages which are then inputted to your analog to digital converter that happens inside your audio interface.

Continuous signals such as voltages vary a lot continuously but digital signals are not. Digital signals are square waves which have only two possible values: 0 and 1. Continuous signals such as voltages at the input of the audio interfaces have infinite possible values. They could take just any value of voltages as long as it’s within the resolution of the converter for example 0.324 volts, 0.345 volts, 0.232 volts, or even negative voltages -0.656 volts, etc.

These voltages hold up the “INFORMATION” of the audio content in analog domain. Remember that the microphone is a transducer that converts sound pressure vibrations into voltage levels. High sound pressure results to higher voltage induced at the microphone output while low sound pressure will have smaller induced voltage

So you could imagine a singer with dynamics (low to high pitch or volume levels) can induce infinite possibilities of voltage levels at the microphone output. These voltage levels would make up the singer audio waveform which is to be converted to digital (if it’s to be recorded in your DAW).

Representing these analog levels in digital domain

To represent analog levels, you need sufficient sampling rate and bit depth to convert the waveform accurately. The sampling rate required is twice the highest frequency to be converted, so for music, it would be around 22050Hz. It is why the most common sampling rate for music would be 44.1 KHz because:

Sampling rate required for accurate reproduction = 22050Hz x 2 =44.1KHz, This implies that there are 44100 analog voltages samples taken per second.

The number of bits you need can also affect the resulting representation. Supposing you want to sample a sine wave voltage signal using 3 bits at low sampling rate, these are the output:

Take note that the sine wave is jagged and not a good looking sine wave. This is not an exact replica of the analog sine wave because it has only been sampled using 3 bits. The maximum possibilities of 3-bits are: 2^8 = eight possible representations of the analog signal voltages. These are not enough, considering that the analog signal is continuous.

However, if you are using 24-bits and using a reasonable sampling rate such as 44100Hz it would become very accurate as there are 2^24 = 16,777,216 possibilities that a voltage levels can be represented and then there are 44100 samples taken per second. The resulting digitized waveform would now be very smooth looking much like the original analog sine waveform:

You cannot see anymore those jagged or ragged corners on the sine wave. Those dots are the samples and there are sufficient samples taken per second at reasonable bit depths; thus making the digitized representation accurate.

Always record at 24-bits for better resolution

The reader is confused that 16-bit is too much because it can store 64KHz. This is wrong, because the bit depth does not have anything to do with the frequencies of the digitized samples. In fact, the bit depths tell you how much resolution you have in your analog to digital conversion. So if you record only at 16-bit, you only have 2^16=65,535 levels. Supposing the analog voltage levels to be coded is from -20,000 mV to +20,000 mV then the resolution would be:

Resolution = [+20,000mV – (-20,000mV)]/65535 =0.61mV per sample

If you are recording at 24-bits, this resolution would be:

Resolution = [+20,000mV – (-20,000mV)]/16,777,216 =0.0024 mV per sample

Now you can accurately represent an analog audio signal if recorded at 24-bits because of this very high resolution. With 16-bits example above, you cannot resolve voltages significantly smaller than 0.61mV resolution so it would simply be round off to 0.61mV thus causing what is known as “quantization error” because of the significant difference between the analog input and the digitized output. This will have a significant effect on the resulting recording quality.

To make this clearer to you, the resolution affects the granularity or steps in the digitized signal, a high resolution results in a more smoothly looking converted/digitized analog signals, see screenshot:

If the resolution is to be made to be smaller, those “steps” would become smaller also until it looks smooth. The resulting sound would also be less “digitized” making it sound exactly like analog. But with 24-bits example above, with 0.0024mV resolution, even smaller changes in the analog input voltage can be represented, thus resulting into a more accurate reproduction of the analog input. It is why 24-bits are a standard in music production for recording. You should be recording at 24-bits.