Hi! This is the first of (hopefully) many blog posts that will be appearing on my site. The plan is to release one monthly. Often these posts will follow what I’m currently working on, but depending on how things go I hope to cover a range of topics. Before I begin, I should say that I use Cubase, so a lot of the tips and tricks will be Cubase-centric, but other DAWs out there have similar features, so this should still be useful even if you don’t use Cubase.

For this post I’m going to be talking about the process of creating a MIDI Mock-up. That is taking a piece of music and imitating it as closely as possible using sampled instruments.

Why?

Why even bother doing this? It’s a lot of effort, and you don’t even end up with a piece of work that you can use commercially…

1) Ear Training

The first answer to this is that it is a great way of improving your ear. It is generally straightforward to work out the main melodies and bassline of a piece of music, but what about the middle parts? Often it’s the middle parts in your own compositions that end up letting you down, so being able to practice tuning into these bits gives a stronger idea of how they function, and also how much they are missed when left out.

2) Targeted Mixing Practice

There is so much choice when using sample libraries - what mic positions do you want to use? Tree, Close, Outriggers, or how about a mix of all available (if you have enough RAM)? Not only that but then there is the world of plugins – EQ, compression etc. Often this leads to difficulties when mixing your own music. What is great about this exercise is it allows you to use these plugins and mic positions with a target in mind. As you’re trying to replicate a specific sound, this leads to exploration with a purpose.

3) Template Balance

This exercise often reveals problems with your template. Your template is a vital part of the compositional process, and something I should write about in the future. Basically, you want your sounds easily accessible and sounding great within seconds. There is nothing that kills creativity more than having to load up multiple sounds, then balance and organise them. By the time you’re finished doing that, whatever inspiration you had has long gone. One of the great things about MIDI Mock-ups is they allow you to compare the volume of your sampled instruments with live ones. Are your Violin 2s too quiet? How loud should the Harp be?

The Process

Ultrasonic Flash by Michael Giacchino

I chose this track so I could really focus on the String section. It is a cue from the last season of the TV Series Lost. You can find the recording on most streaming services, but for obvious reasons I can’t post it here. Here is my finished mock-up:

1) Reference Track and Tempo Map

The original track needs to be as accessible as possible. Dom Sigalas has a video on a neat trick for switching rapidly between the original track and your version. If you do use Sigalas’ trick, a quick warning - if your instruments stop working at any point make sure you haven’t forgotten to switch back to your own mix. Seems obvious, but a surprisingly easy mistake to make!  

After setting up my reference track, I then create a tempo map using Cubase’s Time Warp feature. The purpose of this is to get my version living and breathing like the original. No live musician plays music at a completely even tempo, and part of what makes music great are the pushes and pulls that emphasise the key emotional moments. A good tempo map can go a long way to making your samples sound more realistic. My tactic is to work out the time signature and then drag each bar to the approximate position of each beat one. You can be even more detailed, by going beat by beat. To do this set the time signature to 1/4 and work out every beat of the piece.

This can be a tricky task, particularly with pieces that have lots of time signature changes. Thankfully for this piece, the music was mainly in 4/4, with only one bar of 3/4 at bar 18. I suspect this was originally a 4/4 bar that had to be edited down to 3/4 to fit what was on screen.

2) Marker Track 

Once everything is in place and I know exactly how many bars I’m dealing with, I then begin a basic analysis of the music. I like to put down markers dividing up the piece into sections. This breaking down of sections helps organise the workflow of what is to come, acting as motivation. But also gives you a clear idea of how the composer has organised their music. I would recommend using markers for your own music, as the process allows for a more coherent approach. It is worth mentioning that at the start the markers don’t need to be very detailed. You will have opportunities to move them around and rename them if you notice something in the future.

3) Transcription

It is worth saying that you don’t have to do this exercise by ear. There is still a lot of value in this exercise when using a score. In this case I don’t have the score to Ultrasonic Flash, so had to work things out by ear.

There are a few tricks that help this process. First off I like to have access to manuscript paper, and I also have a few tracks loaded into my template for sketching. I have 5 piano tracks labelled: Melody, Counter Melody, Harmony, Other, and Bass. This allows me to test out what I’ve heard in the music. I would strongly recommend using a sampled instrument that matches what you are listening to. For example if you are working out a Violin 1 line, load up a Violin 1 sound and try and play along with the music.

With tools in place, I start with the parts that are the easiest to hear. Jot down the main melody, followed by the bass line. With this down, it frees the ear to concentrate on other things. Often our ears develop a form of tunnel vision, where we end up blocking out different frequencies, and it can be a surprise when you suddenly notice something that should have been obvious from the start! The bassline is particularly important for working out the middle voices. If you struggle with the bassline, you could put an EQ plugin on the reference track and either boost the bass or cut off the highs to help. If you notice any special techniques make sure you write them down. It’s important to mention that you shouldn’t put too much pressure on yourself to catch everything here. Often the actual mock-up will reveal missing voices. When you do discover things missing it’s worth keeping track of where they are. I use scientific pitch notation (C4 = Middle C) to reference these pitches, so if you notice that most of the missed notes are between C4 and C5, this can help reveal blind spots in your hearing that you can work on further down the line.

4) Optional – Create Score in Dorico/Sibelius

At this point, I like to create a score in Dorico of what I’ve noticed so far. Inevitably this a process that will work better for some musicians than others. For me it ends up being an invaluable resource for analysis, but is also incredibly helpful when creating the mock-up. I use this score to help plan dynamics in the music, and to try and spot gaps in my transcription. Not only that, but I can go back to the score when I encounter a musical problem in the mock-up. Having the notes lined up in this way helps me come up with good ideas of what the problem might be.

5) Begin the Mock-up!

Finally we can start! First off, you should have either a physical or digital notepad near you at all times. As you work through the mock-up you’ll have loads of observations and ideas coming at you at once. You will NOT be able to keep all of these in your head, so create a to-do list so you don’t forget anything. Good things to note down include: voice missing at bar.. balance feels wrong at… not liking the sound of the harp at bar… etc.

As you play around with the instruments available to you, make sure you experiment with mic positions (if you have them available to you). All the main sample libraries allow you to play around with mics. Usually you will have access to a ‘Close’ Mic (or ‘Spot’ Mic), that (as the name suggests) is close to the instrument and captures lots of detail. You’ll also have access to a ‘Tree’ Mic (Decca Tree), that is placed high above where the conductor stands giving an impression of the instrument placed in the room. It’s worth going through the user manuals of your libraries to see what options you have here. Spitfire have a good article about their mic positions here. The reason it is so important to experiment with the mic positions here, is that this is the best opportunity to get as close to the original as possible. You can mix and match microphones on your instruments but be warned that the more mics you use the more RAM you’ll need on your computer.

Make sure you work with the CC data of your samples. CC11 (Expression) and CC1 (Modulation) are the main ones, but Spitfire for example allow you to control vibrato on CC21. This can make a huge difference to the overall feel of your music. Finally, when using Strings, use ensemble samples blended with the individual samples to add strength to the sound. I found that particularly helpful in this project when the Violin 1 samples were sounding a little weak high up. Using the ensemble patch to reinforce them helped to match the original.

6) Audio Print

Once I have a completed version of the mock-up, I then print to audio. This is something that I’ve built into my template. It takes a bit of working around, but really helps with the workflow. The idea behind printing to audio (i.e. recording your samples into audio format) is that audio is much easier to do fine tuning to than midi is. With samples each time you hit play you get a slightly different performance as different samples are used each time. With audio you get the same performance each time, which is much easier to control. Being able to see the audio also makes it much easier to do fine tuning on the rhythm and volume.

The most important part of this stage is editing the volume of the audio. You can edit via a number of ways in Cubase. I use the pen tool on the audio to automate the volume on each audio clip, and you can also edit the overall volume of each clip by dragging the white box at the top centre up and down. I would say this stage made the biggest difference to my mix - much more than any plugins. Here is a comparison of before and after:

7) Mix and Master

This is a huge topic that deserves a lot more time. For now, here a few things I did on this project. I used EQ plugins to fine tune some of the instruments - for example I looked to match the Harp sound at the beginning of the track as closely as possible. Remember that the best way to do this is using the best mic positions earlier on in the process; all EQ plugins can do is fine tune the sound. One thing that gives a live feel to the original is that there is a lot of Room Tone in the track. Room Tone is basically any sound in the track that isn’t music. I tried to mimic this by using pink noise and an EQ plugin - I’ll let you decide how successful that was! I also used some compression on the opening part of the Harp to make the tail of each opening note louder. I then used Ozone 9’s Match EQ to compare the EQ curve of the original track to mine. For the reverb I used LiquidSonics Cinematic Rooms, again this is another huge topic that needs a separate article. Finally, I used Ozone 9’s Mastering plugin to match the volume of the original track.

This stage is hugely important for the learning experience. Try and be as experimental as you can in using the plugins you have at your disposal. You’ll almost always discover something new during this process!

And that’s about it for now, thanks for reading all the way to the end! Let me know in the comments what you found helpful, or any suggestions for future topics.

Daniel