About recording, mixing and mastering
Because I haven’t found a comprehensive online presentation about classical guitar recording, editing, mixing and mastering, I decided to make one for your free use. See also my new blog: https://mixhappens.blog (sorry, only in Finnish at this moment).
In case you would want to practice mixing and mastering with the original recording, you can order the original mix-ready file and use your own mix as commercial preference in your advertisement, as long as the basic information of my composition and performance (name, composer and performer), as well as the link to this web page is properly mentioned in the general information of your own mixing.
When it has to do with a hobby, people do not really differentiate recording and postproduction with its different phases from each other. Often the most important issue probably is to record one’s own playing for the enjoyment of oneself and one’s friends. It is possible to make a sufficiently high quality recording with an affordable microphone and a recording program, or even with a modern smart phone.
However, the so called home studio hobby takes most people with its flow. Subsequently, a problem will be formed by the excessive and confusing information available online, and the efficient online marketing machinery, that thinks that it is not possible to make a good recording without some very specific equalizer, compressor or a reverb. The best guidance I have received regarding the issue is, using only the software that comes with your own recording program, the ”plugins”, for so long, that you can use them skillfully and learn to hear the difference for example between two compressors, when the compressed audio has been adjusted into as similar as possible in terms of volume, so that the sound volume will maintain the same regardless of the compressor being enabled or disabled.
When recording, one should always use as high quality guitar as possible, one that has been serviced for recording. If the case involves a low quality instrument that is out of tune, has old strings, or the performance has not been trained for, it is impossible to make the end result sound good, even with professional audio post production.
There are several ways of recording with a classical guitar. If the case is about recording a solo performance, it is common to use two microphones for the recording, in order to achieve the natural stereo image of the audio. If the recording room is small and less than ideal in terms of its acoustics, during recording it would be advisable to use acoustic covers to reduce the sound of the room and its reflections being caught by the microphone. For this type of dry recording, it is significantly easier to later on add a high quality reverb with computer software.
The further away the microphones are from the guitar, the more the acoustics of the room are being caught by the microphone, and also on the other hand, the bass frequencies of the guitar are being significantly subdued as the distance grows. There does not exist a ”right” distance, but the usable distance can be discovered by making short test recordings by starting from a distance of 20 cm from the guitar, and then by always putting it 10 cm further away on the next test recordings, up until to one meter away from the guitar. When there exists an amount of bass frequencies in the recording that sounds natural to the ear, and it is possible to also hear some of the room’s own voice, the situation is already at a pretty good starting point. After this, the best possible distance, the so called ”sweet spot”, is being looked for by making small movements and additional test recordings. By directing the ”central area” between the microphones slightly from an elevated angle to the attachment point of the sounding board and the finger board, it is difficult not to end up with a pretty good recording, after the suitable audio distance has been found.
Later on in the text I have attached some photos of the positions of the microphones and the distance for clarification. You can be found several different ways of recording a guitar online, but the way I have presented above is a simple one and suitable for most situations. A good level of volume for recording can be found by taking care that the audio should not clip at any point. Usually audio equipment has some type of light or a meter that indicates the suitable sound volume. Personally, I leave at least a 6dB safety marginal (peak levels -6dBFS) when I am playing the loudest part of a song.
After this, the performance is being recorded several times by playing it as a whole and in parts for a sufficient amount of times, so that each single part of the composition has been recorded without one single flaw at least once.
After the recording phase, the several ”takes” will be combined into as flawless and coherent performance as possible, and an audio file will be prepared for the actual mixing phase, for example by removing the digital errors, and by naming the file in a comprehensible way (such as Feel-The-Flow-mixready-ddmmyy).
After the editing that has been done in order to create a flawless audio file, the aim is to enhance the good qualities of the audio, and to tone down the disturbing frequencies as well as the shortcomings that are typically involved in the audio. The recording is enhanced for the listener, by for example adding a pleasant reverb so that the recording sounds like it would have been recorded in a magnificent sounding concert hall.
Finally, the mixed audio file will be touched up both artistically as well as technically for the final distribution in the mastering phase, so that the audio would sound as balanced as possible in all types of speakers and earphones, and that it would sound good when being compared to other audio recordings. For commercial distribution metadata is added, for example the name of the composition, the composer, and the performer are being coded with so called ISRC code, as a digital series of numbers in order to make it easier to recognize the recording and the management of copyright based usage of the performance. My personal ”master” is finally a coded wav-file, that I have converted for my own distribution into an mp3 file.
I also asked the legendary Abbey Road Studios to master my mix, but eventually I decided to use my own version. It was nice to notice that my own master was fully publishable.
In this presentation I am just going to be using the name ”mastering” for the final touches performed on the recording for the purposes of distribution.
In the next part, I will go through the different phases, from preparing to record my performance all the way up to making the finishing touches on the audio file in the order described above, because each phase has its own purpose.
I started preparing for the recording of my composition by changing the strings to my guitar, because only by using fresh strings it is possible to get the airy and sparkly sound quality typical for a classical guitar. When the strings have been used more, they slowly lose the overtones necessary for a high quality sound. In the recording, I used Corum Standard Tension strings from Savarez.
Because it is not a good idea to even start trying to record before the new strings have finished stretching, for the next few days I had the time to practice my performance and to take care that my nails are of the right length and shape before the recording situation. I finished the nails a couple of days earlier with a water sandpaper under warm water. Now I am just waiting for the best recording day for the nails. Even the best mixer cannot make a poor sound anything more than a poor sound.
Before the recording and during it, I was keeping note of the tune of the guitar closely, so that the tuning would stay as unchanged as possible. I was recording the composition for several times, so that I could later on compile that into a unified and flawless audio file for the mixing. I prefer to play the song in full without interruptions regardless of small mistakes, so that the performance would be a coherent interpretation. The more takes you need to record, the harder it gets to keep the tuning identical, so that the takes could be compiled into one coherent entity. After the bridge fell off on my Pappalardo guitar that has a beautiful sound, I was forced to use my old Jair, so with the tuning and sound quality of the guitar there were no other options, than to accept some compromises.
I recorded my playing with two microphones, as a so called x/y - crossed pair, of which the other one was pointed to the guitar bridge area and the other one to the finger board. In this case the microphones are as close to each other as possible, so that the sound waves wouldn't cause unpleasant comb filtering when they are at different phases. The distance of the microphones was approximately 25-30 centimeters from the guitar. The further away the microphones are from the guitar, the more the sound of the room / acoustics will be recorded to the recording and the bass frequencies of the instrument are subdued. I wanted that the recording would have a natural amount of bass frequencies and as little of the room's own sound as possible, because I would anyway add a reverb of a Spanish cathedral in the mixing phase to the rather dry recording. As a recording location I had my living room that is a rather open space and has a parquet flooring. The recording style presented above functions well, even if there exists other ways of recording as well.
I did a few test recordings, and came to the conclusion, that I do not need acoustic panels or a half circular acoustic cover to stop the voice reflecting back from the surfaces of the room to the microphone, that might change the sound unpleasantly due to coloration in a small room. It is recommendable to use one's own ears and not to automatically assume, that a usable recording would always require acoustic adjustments to the recording space. The aim of my recording is to primarily present my composition to guitarists looking for new material and not to record an artistically finished performance, so there wasn’t any need to make it "perfect". In the recording phase it would be good to always think about the intended usage of the recording.
Preparation for the mixing phase
After recording I compiled one entire version of the song that was as flawless and as coherent as possible. The errors in playing in the best take have been replaced with successful parts of other takes. Single takes are separated with light vertical lines in the image below, and a singular cut as a crossfade. The microphone pointing at the neck has been panned to the left and the microphone pointing at the sound board to the right, because I myself am used to hearing my playing from the point of view of the player. From the point of view of the audience the panning of the microphone to the right would be more natural.
Finally I printed all the takes into one new audio file.
Because in combined audio files, and in recordings in general there always exists different types of digital errors, such as clicks, pops, accidental sounds made by the player like the sound of the breathing and for example the sound of traffic. I fixed these to a certain extent with a plugin, that has been designed for the post production and fixing of recordings. I especially toned down some highlighted squeaks typical to guitar playing, that are created when the fingers slide on the string when the playing position on the neck is being changed.
In picture A, the software is correcting 66 small digital errors, also known as clicks, from the recording.
In picture B, I am slightly subduing the squeak of the string. It is sensible to avoid excessive correction, so that the final result would sound natural. After the correction the listener's attention is not drawn to the squeak anymore, but it sounds natural.
In picture C, I am removing the sound that was created by my chair moving during the end note of the song.
In picture D, I have adjusted the volume of single notes slightly (the light line down below), so I am correcting the tones that I have played with too much or too little volume 0,5dB-2,5dB in order to control the dynamics of performance. The aim is that the listener could hear all the notes, but not one single note would be highlighted too strongly.
I used "manual" volume correction during mixing instead of compressing which is more popular. In this way the final result will sound as natural as possible. The aim of compressing is to narrow the dynamics of the performance in a way that the difference between the most silent and the strongest notes would be evened out. Often compressing is also used for adding a slight distortion to the audio file, even if by running the audio file through an analogue compressor or one modeling that without compressing or by only doing it very lightly.
By compression it would be possible to for example soften sharp sounds, by using a very fast compressor (fast attack). For this purpose I used later on what is called dynamic eq and de-esser. This type of manual volume correction is most often done in the mixing phase. However, the correction happened smoothly and naturally already in this phase.
Next up, I opened the above mentioned prepared file in a software called Cubase. Into the same session, I added two commercially mastered famous guitar compositions ”ref 1 and 2", to which I compared my own mixing.
By comparing the clarity of the audio files or the amounts of bass and reverb, it is easier to ensure that the mixed recording sounds good when being played on different types of audio equipment. I adjusted the volume of the audio files to as similar levels as possible in order to make the comparison easier. As a primary preference I used the wonderful "Recuerdos de la Alhambra", that has a balanced sounding recording. Nowadays commercial recordings are often being mixed into being brighter, meaning that they have relatively more high and high-mid frequencies.
From the actual "mixer" view, it is possible to see the plugins that I used for the adjustment of the audio file located on the left.
First, I adjusted the recording with two equalizers (eq). I sent a copy of the audio file to the reverb plugin. Then I adjusted the reverb track with another eq, so that the reverb would sound natural and would not muddy the audio.
In this way, the corrected audio file and the reverb are being combined into one stereo track. I further adjusted this one with the eq, because the recording still sounded too harsh and muddy and somehow honky in comparison to the reference recording.
I added two different plugins to the reference track for analyzing the recording that I was comparing my own recording to. I used the same plugins also for the analysis of my own mix on the mix track, after which I deleted them from uselessly eating up the processing power of my computer.
First I cut out the lowest frequencies (< 30hz), that cannot even be heard on regular audio equipment, computer speakers or earphones. Removing the lowest frequencies clarifies the recording. Often when mixing / mastering the frequencies below 30-35Hz are being removed. For the same reason I have removed the highest frequencies as well ( >20kHz). Other than those I reduced the typical problems from the recording like muddiness, boxiness and honkiness or otherwise edgy and hard noises. In mixing descriptive terms are often used in order to make it easier to recognize certain frequencies. In addition to the above mentioned, I have used eq only for reducing problem frequencies (subtractive eq). Oftentimes mixing is started by attempting to remove or to weaken the frequencies that are reducing the audio quality of the recording. For example the free strings of a guitar resonate more than the other notes, and that is why the free strings are often enhanced when playing and thus covering over the following notes, which turns the recording unclear. Also the acoustics of the room cause some frequencies to be distractingly enhanced. Analysis software are making it significantly easier to find these problematic frequencies.
Eq is also used for enhancing the frequencies that sound good in the recording (additive eq), just the way I did myself in the mastering phase. At the side I enhanced with the additive eq the 1kHz frequency, which brings the guitar somehow closer to the listener.
I toned down sharp noises with the so called dynamic eq, that tones down the frequencies that are being processed, only when they are being over enhanced, for example when a string hits the frets of the upper parts of the neck of the finger board too hard. Dynamic eq is often used, if the frequency is only randomly present in the recording. Traditional, Oxford Eq I used for toning down qualities that were constantly present in the recording, such as a honky or a nasal voice.
After this, I used the reverb plugin to add reverb to the recording, so that it would sound like I would be playing the guitar in an old castle chapel instead of my own living room. Often reverb is only being added so much that it adds some three dimensionality and realism to the sound. Then I delayed the reverb to start 20 ms after the original recording, also known as pre-delay. In this way the reverb does not cover up the beginning of the notes that are being played and the sound of the guitar is closer to the listener.
From the reverb I first reduced frequencies below 80Hz, so that the low frequencies (impulse responses) taken from the chapel echo audio sample wouldn't blur the recording. Correspondingly, I cut the higher frequencies, so that the echo wouldn't cover up the recording itself and the higher frequencies would stay clear and bright and the reverb would stay slightly warmer. After that I reduced some frequencies from the reverb that were needlessly covering the recording or somehow enhancing the sharpness in the sound of the guitar needlessly. Thanks to the Eq, the reverb sounds more natural and blends into the sound of the guitar in an unnoticeable way.
I continued to enhance or cut some frequencies, so that the eq balance of my mixing would be slightly closer to the song that I was comparing my own mix to.
Finally my own mix (on the left) sounded rather balanced with the reference song. As it can be seen from the pictures, both recordings have relatively equal amounts of high, low and mid frequencies. This is how an analysis software makes the mixing process significantly easier. After this I saved my composition into a new file that I then listened to on the speakers of the laptop, phone, home stereos and in the car's audio system. Good mixing sounds balanced and listenable in different systems.
Finishing touches for publishing
Next up I purchased three albums for reference, and downloaded a few songs I had chosen from each album and added them into their own tracks. After this I ran the tracks album by album through the analysis plugin, in order to get an idea of the typical eq division and width of the stereo image on a full guitar album. On the left on the top, my own mix can be seen. The starting point seemed pretty good, even if my own mix seemed a little bit thin and edgy. Some single frequencies were enhanced too much creating distracting sharpness. Additionally I had cut slightly more of the lowest frequencies.
Sadly the sound quality of my spare guitar does not quite compare to the top notch guitars used in the songs and albums that were used for reference, so I needed to mostly bring out the best qualities of my own recording. This is a good example of how the most important thing in recording is the quality of the source of the sound. It is impossible to mix an average guitar into a high quality one. Because the recording was meant to mostly function as a support for the notes, and not really as a guitar performance, I didn't really worry about this issue too much.
In the attached picture, the audio file that is to be mastered is the one on the top and the three reference songs I chose are each under each other in a new Cubase session.
Next, I slightly increased the width of the stereo image (131% in the picture below), because the stereo image of the reference albums was wider than in my own mix. After this it was easier to compare the recording to the reference recordings in order to fine tune the eq.
Usually already in the mixing phase I increase the width of the stereo image slightly (8-12%), because practically in the mixing phase, or at the latest in the mastering phase the width of the stereo image is usually increased anyway. The picture below on the right shows the final width of the stereo image. The stereo image was also very centered, which should always be checked, and fixed if needed.
After this I used dynamic eq to decrease frequencies that resonated too much and decreased distracting frequencies that sounded somehow too sharp, honky or otherwise distracting.
For example, approx. 220Hz repeating of enhancing caused muddiness in the recording.
In order to sharpen, warm up and to thicken the warm image and to make the higher frequencies softer, I added mild distortion, odd and even harmonics to the recording with two tape and distortion emulators.
In this way I attempted to copy the warmth and analogue sound of analogue equipment that makes the recording a more pleasant one to listen to.
The recording sounded still a little bit sharp, which was why I decreased the distracting frequency with the de-esser. The idea was, that I would decrease the frequency only when it is being over enhanced. In here, I have set the de-esser so, that it will mute the 2600Hz frequencies temporarily at most by 3dB, so the eq balance doesn't really change at all. The same spike is still visible in the following pictures, even if it is slightly less enhanced.
Then I chose one of the reference songs, that reminded my own composition the most in terms of style and I adjusted the eq balance of my composition with two consequent eq's slightly closer to the eq balance of the reference song.
Finally my composition sounded sufficiently balanced and listenable when compared to the reference song. Underneath on the left there is my first version and the reference song's eq graph. After having listened to my "master" with different headphones and speakers during the next day with rested ears, I slightly enhanced the 200-1000Hz area more and correspondingly reduced the 2-7kHz area, so that the recording would be slightly warmer in tone and less sharp to the ears.
I ran the audio file through Neve compressor and API's equalizer without processing in order to add the analogue shade to it. I used a limiter to make sure that the audio wouldn't clip when the volume is at its loudest (output trim -dBFS). After this I normalized the audio, meaning that I adjusted -1dBFS as the maximum audio volume, so that the audio file would be played with as high volume as possible. I additionally added metadata that contained the information of the work and performance as well as the ISRC code. Finally I printed out the audio file both as wav and mp3 files. After this I checked once more, that my playing begins and ends without an empty beginning or ending and that there would not be noticeable or visible faults or errors in the files.
In this phase it is good to remind oneself, that only very few listeners will pay attention to the "fineries" of the mixing. None of my friends have ever marveled at a fantastic stereo image. The listener is mostly interested in the composition itself and the performance of it. Excessive self-criticism and endless fine tuning is best to be forgotten.
Although I have used professional level software, plugins and hardware, the mixing and final touch ups can be done with basic software and plugins. The significant part is to be able to utilize the available software and plugins as well as possible.
Above I have compiled all the different phases of recording my own composition Feel The Flow, all the way from the preparation to the final audio file. In this way I want to encourage every single guitar player to get involved and fascinated in composing, recording one's own playing, doing the post-production on the audio and on analytical music listening in general. Listening to one's own playing for example from earphones or composing are great ways of finding new enthusiasm into one's own playing.
In Tampere 2.3.2019.