90s music production techniques vintage retro
Here is the completed track : ” A source of light “
This is a write up for the production of a Goa trance track. Initially I had in mind to make this as authentic as possible, as such I had a few reference tracks that I had been enjoying. Some for their musical qualities, some for the overall energy and some for their sonic attibutes. There was quite a range, I was noticing that some of them really did sound pretty rough in the upper mid ranges. Whilst this is an original trait it is something that I find a little bit difficult to suffer these days.
90s Goa trance production arrangement
Additionally, when you start to compose a track they have a great tendency to start asking for their own way, sonically speaking, and creatively and there is nothing wrong with this. As such maybe this track is not quite as authentic to the original sound as I originally planned, however it does indeed sound like late 90’s goa trance. If I would have to pick a year it would probably be 1998. There was quite a difference between a track from 1996 to 1999 as virtual analogue hardware synths started to become more commonly used. (Nord Lead, JP-8000 and Virus) One could say a slightly more modern sound started to appear along with what we might define as proto “psy kick and bass or “KbBB”. (which denotes the 16th note bassline we commonly hear)
However, there are a lot of analogue software emulations in the track.
All about the bass line
D-16 group Phoscyon 2 Roland TB-303 emulation
No such modern style kick and bass in this track though. This is classic, using a Roland TB-303 emulation, D-16 Phoscyon 2 which I recently upgraded to. And I am not disappointed with that upgrade. A vibrant, warm, powerful, gritty, bass line was produced. I also like many of the new features in that plug-in including the tweak section in the top right hand corner which allows adjustments of PCB level trim pots (which may have accounted for some factory variation between units depending on who set them up and component tolerances of the time). This under pins the entire track so I may as well give detail.
I intentionally kept the low end fat, thick and gritty for this track. Not all reference tracks had the same low end weight but some did and I followed suit and wanted to get that kick drum and TB-303 chugging along, so it is big on smaller systems as well.
I love delay on a trance bass, though it is very rare these days, it augments the bass line, it gives dimension, interest and makes it spatial and it inevitably makes it sound “of a time”. (All importantly it sums to mono nicely.)
Goa trance TB-303 bass processing
My CPU however was rather less impressed with the Phoscyon 2 update and my other choice of synths and I very quickly started to hit 90pct ASIO / CPU usage as the track progressed. Not uncommon with U-He Diva and Repro-1 and 5 in the mix ! (Despite “HQ” being off and “Multicore” being on). As such I had to bounce quite a few parts early on to free up some headroom. Fortunately Phoscyon 2 provides a draft mode and allows “Ultra” for rendering.
The bass line passes through the Trident desk emulation courtesy of Slate VCC with a very small amount of drive but has no channel saturation beyond this (it was merely on 2/10) which is very minimal yet it sounds really thick and full which is testament to Phoscyon 2. Though there was one not so secret tool that I used that is you were very lucky you may have owned in the rather more well equipped Goa trance studios of the time. The notorious pitch shifting algo’s of the Eventide H3000 in this case the H3000 Factory MK II plug-in which is good but does not have all effects algos of the entire H3000 hardware. I do own UltraReverb and Blackhole also by Eventide.
Eventide H3000 Factory MK II pitch shifting preset used as a bass line send effect
The H3000 Factory also had a nice update which I purchased for a reasonable price in a recent sale. A very interesting upgrade both in authenticity of the algos and the addition of an AD/DA model. And it just so happened I had an interesting plan, that did not go as expected this time for this ADDA model. More on that later.
For the bass line I set this (H3000) up on a send basis along with my chosen delay. For delays on bass lines I strongly recommend filtering out the bass and lows in the delay path either at plug in level or EQ the returns as this helps with mono compatibility of the bass line bass frequencies < 100Hz (and stability, as phase interaction between bass line and delays is minimized). In addition, I set the feedback time, HPF in the delay and the width of the “ping pong” so it was at just the right sweet spot to make the bass sound as spatial, driving and exciting as possible yet mono compatible, which is important.
I am very pleased with the growl the bass line has.
Retro vintage 90s synthesis
I pulled out all the synths I might expect to hear in a Goa trance track, SH-101 (LusH-101), TB-303, Repro 1 and 5 – Diva (all round amazing authentic analogue sound) TAL’s Juno 06 and a few other synths that I thought might fit the bill. Surprisingly ANA2.0 from Sonic Academy always puts in a nice performance (I especially like its filter options) for these tracks.
In addition Waldorf Largo makes an appearance. TAL Mod does a great detuned multi saw (that is one of the archetypal Goa lead sounds) and has a lovely band pass filter that I was suggested to try out. Hybrid 3 provides some workstation-esque “digital” pads, Predator II some occasional SFX.
Slowly I was at 80pct ASIO/CPU again, such is aging DAW life.
U-He Repro 1 software synth
TAL JUNO 06 software emulation TAL-U-NO-LX
Uhe Repro 5 sequential circuits
TAL Mod synth
Effects processes
I used RoomWorks/SE which is an old Cubase stock reverb that can sound rather like an Alesis Quadraverb and other reverbs from the 1990’s period and some basic delays. There was no complex or “expensive” effects processing occuring in this track (other than H3000 algo’s) just the basics to produce a nice wide rich sounding mix, reverb and delay.
I do use Slate VCC on the returns as well as all source instrument channels as these would have come back into the “Project mixer” (typically a 24 channel 8 bus type) via dedicated stereo returns in the master section or less commonly back into line input channels.
Alesis Quadraverb ? Roomworks does a similar task.
Resolving a mono incompatibility issue
I had some pad sounds (choir and gated) which I really liked but they were a little too wide and did not sum that well to mono. The level dropped off more than I was happy about. This is actually not that uncommon with 90’s mixes but this is something I did not wish to authentically replicate for this track. I resolved this by inserting Nugen Stereoizer (which is also a very good mono-izer and what I use it for most.) I reduced the image width by about 50 Pct on the pad group. This narrows the image and reduces the amount of L-R out of phase components in the L/R channels. (pseudo stereo or “sides”) which also makes it more mono compatible as a result. As an unwanted side effect it reduces the volume of the sound in stereo so I bumped that group up 3.5 dB. This was settled upon by listening (A/Bing by ear) the pad with and without Nugen inserted. I balanced the pad up by ear so it sat at the original mixed volume.
How to resolve a mono incompatible sound in the mix
Then as a test, I mono the master bus and I could hear that the sound was much more mono compatible whilst at the correct level and importantly still sounding similar to as it did before, just a little more narrow. An effective and straight forward problem fix.
Always keep checking your mixes in mono to ensure that the sounds hold up in mono. It is easy to forget as it is not part of the flow of creativity but really is worth checking. You know it has to be done so just do it, don’t delay.
A different approach to a kick drum this time
I had upgraded to Sonic Academy Kick 3 recently and wanted to give its analyse functionality a go. I normally design my own kick drums in Kick 2 and so I am very much aware of what I am doing to produce a good sounding kick drum. However, this time I tried to locate a kick drum from the era from a pre existing track. That probably took longer than making a kick myself ! It was fun listening though. I am very pleased how Kick 3 managed to recover a kick drum that could not be called “clean” by any means as there were other sounds layered over the top. It did a great job.
Sonic Academy Kick 3 analyze function tested
I cleaned the extraneous sounds away using the envelopes in Kick 3 and rebalanced the transient and sub content and also added a little mix EQ. I am pleased with the end result which broadly has the character of a reasonably modern sounding Goa Kick more of a 1998/99 sound. A little more “tennis ball bouncing” with juicy sub, than the very zappy, ‘muddy thuds’ of early Goa trance. I knew what kick I wanted to go for well in advance but it took a while to hunt down a track that had the right kick drum.
In all honesty, I could probably made a great kick from scratch quicker, but it was a satisfying process overall to test the new functionality. It marries really well with the bass line as I started with the kick and ensured the bass line gels well with the selected kick. This is all important in this music as the kick and bass drives the entire track along, you need a strong foundation to start and build from.
General writing and composition process
A slow and rather tricky piece to make this time for a few reasons, I had a stiff shoulder so this slowed me down in general, rather spending hours working on mastering as my sole source of income and doing some music making when my shoulder was feeling better. Quite slow to write after the initial inspiration. It is normal, so get over it and yourself and keep powering on.
It becomes more difficult to write in Phrygian mode after a while as so many melodies have been used before. I changed scale a few times when planning the track (which also slowed down the entire process with false starts) but as I am writing a single track not an album I did not feel the other scales quite right at this moment.
The usual tricky situations arose when extending and arranging, too little here, too much here, too repetitive etc. so worked to resolve these issues over some period of time. I find this reflection is needed to know what it is you have actually made. Tracks take the time they need, I have learnt this and they tend to guide me, not the other way round. As long as I leave time between changes and listen in other spaces with a pen and paper to take notes. I am quietly confident I will get through any challenges with tracks I start.
There are a few techno style sounds in this track, maybe one or 2 even slightly cheese oriented, maybe a few, not entirely suited to Goa Trance but this adds a little character and oddity and they work for me. Goa Trance was an immensely creative genre at the outset and so the occasional incongruous sound works all the better, many made this music with what they had available. Goa trance was quite raw sounding and as such I did not correct everything into cleanliness or perfection and left the sonic elements with a nice grit fitting of the time but without pushing that too far and let it be tasteful.
I thought it worth mentioning how time line markers can help you to remedy problematic parts of an arrangement. I usually have a marker track or 2 for just this and add markers for “something new to happen here” this is when part of the track is maybe becoming a little repetitive, more work yes but release once, take your time and make it count.
DAW arrangement time line markers
Other drums
I mainy used NI Battery with 2-3 instances of various kits, generally Roland 909 and 808 as would be common, there may be a 606/CR-78 hi hat in there. Also self resonating filter “lazer zaps” were common to MIDI roll in Goa trance. These sounds still sound incredible and simply beg to be programmed into nice rhythms and rolls. The snare is a selected sample which I narrowed as it sound a little too wide for my liking. I worked quite a while on closed Hi hat velocities to get the pattern sounding nice with a good groove. Same with snare “riddims” as I labelled them. (a little like Astral projection tracks) And also a new subtle hat pattern was added later on. Little variations add some interest.
A snare must land correctly
Another important aspect of this music is the interplay between the open hat and snare drum landing. The key I have found is to allow the open hat to lead a little as this sets the snare landing up just right so the groove moves forward nicely. If the snare leads (with volume) sometimes it can create a groove which subjectively I like less. I did not mix the snare that loud (often in Goa trance it would be a clap as opposed to a snare) this allows the open hat to lead. Often the snare is not so loud in Goa trance.
There is always a bit of to-ing and fro-ing of the individual drum relative levels (all one shot individual drum sounds) until you have the groove interdependence that you are listening for. It’s all in the details, fortunately the more you do it the more you know what you want and it tends to fall into place with the correct shuffling around of volumes against the kick and the bass. It is often not as easy as one might initially think. I am trying to build grooves not simply accepting whatever happens first when you place and balance drums.
Mixing and mastering
The following is rather technically indulgent and if you wish to skip it, pass down to the + sign a few paragraphs below.
I approached this a little different this time. I initially considered to use the default empty state of the Eventide harmoniser as a stereo output bus processer. I asked Eventide if the AD/DA model was in tact (which it immediately sounded like it was) in this “effectless” default state. It is and I initially liked it. It immediately helped soften some of the highs and upper mid range ills that come from unrestrainted analogue and virtual analogue synths when playing squelchy filtered lead lines. Back then these were not all equalized into submission and softened, smoothed and de-harshed, they were left to let rip. This can be good as it gives the sound a very wild “punky” feeling.
However, it can also be quite unpleasant on the ears so this H3000 as a bus processor helped. However, despite the obvious roll off something was bugging me so I checked what was happening in Plug In Doctor and at the 0dB Gain input setting I could see what was “going wrong” or according to Eventide going right. As I had gone somewhat down the rabbit hole I recreated this EQ curve in Pro Q without the L/R mid range disparities.
H3000 MK II Factory can exhibit a strange L R channel FR disparity
All in all though it was a creative experiment gone a little too far for myself and I ditched the Pro Q4 LPF, benign though it was (phase response) and instead left the master bus flat and adjusted my Mastering EQ a little to compensate, doing the least possible is always the best plan. Whilst it did not sound bad the original idea in my mind was to go some way emulate a 16 Bit DAT Tape ADDA (The timeline works as the H3000 and DAT was 1987 as well). It was pretty clear that this had gone way beyond what I was expecting and was audible across the entire mix. The H3000 had bespoke AD DA as off the shelf chips did not exist at the time so they were “imperfect” but worked, they did what was required in such a digital device.
Subsequently I found that the frequency response linearity is greatly affected by input level. And I mean greatly, up to 4-5dB differences per channel in the mids through upper mids. This was explained as potential clipping. This may have been correct, although the input metering on the H3000 MK II Factory plug in showed nowhere near clipping (all in the green and just 4 LEDs showing lit) in both music making and plug in doctor. The uneven L/R frequency response was there though, with a much lower input level the FR becomes much more flat, although never entirely flat. Eventide did say this was a very colourful device.
You would be well warned it can sound quite different depending on drive levels irrelevant of what the meters on the plug in suggest. Use your ears and let them determine what drive levels sound right. I suggest setting up the gain structure with great care.
H3000 lower gain settings flatten things
Frequency response becomes more flat still with ever attenuated drive levels
This proves technical creativity does not always go quite as planned and I was pleased to work out the issue and come to a more satisfactory and logical result. Of course the Eventide was still used as a send effect on various sounds but not on the stereo master bus which I felt was a step too far.
As such I remained with Slate VCC Stereo bus summing and the Sontec EQ set neutral (there is slight harmonic enhancement).
+
Project mixer channel EQ emulation for the 90’s
Mackie 8 bus EQ preset for Fabfilter pro Q4
Most EQ was broadly based on my Fabfilter Pro Q4 “Mackie 8 Bus” preset which you can read about on other blog articles I have written. (in essence low and high shelf, 1 semi and 1 fully parametric band and HPF at 80Hz which could be inserted into the line signal path.) Other projects desks were available and used, Soundtracs, Studiomaster, Soundcraft Spirit, Allen and Heath etc.
Contrary to the above statements, which openly shows creation is not always entirely logical and smooth, the track was rather dark tonally when I brought it to mastering. I had to back track and make some adjustments which complicated things slightly before I could master it.
Goa Trance often had quite a present upper mid range, often slightly harsh and scratchy, this was wide spread amongst references and not something I really wished to overdo for the sake of authenticity. I want it to be fairly nice to listen to. I chose the middle path and left it slightly less smooth than I otherwise would do. At times it was also quite wide sounding (maybe slight over use of an original Behringer Ultrafex back in the day.) Not always though, though it does suit the spacy sound. Many tracks still worked well in mono (which is just as well as many beach sound systems would have been exactly that, mono !), as such I did a little widening with my favourite widening tools, nothing too extravagant but tastefully so. (Incidentally this is where it was reinforced that it was important I had resolved those pad mono incompatibilities.)
I EQ’d the track in mastering to bring everything into balance tonally. The beauty of this is the kick and bassline presence is also brought up a notch or 2 (through, said equalization) along with all the leads and other drums. The instruments were all well balanced with each other just the tonality needed presence and a little extra highs.
Rather than go back and start changing mix levels (or EQing the individual mix elements) which potentially (and often) affects both feel and groove, EQ can be a supremely powerful tool to keep “what is” and yet very effectively enhance the overall tonal presentation. This saves a lot of painstaking work which could open up many problems of groove change and feel change. That is the power of EQ knowledge, very valuable enhancement with very little drawback on the compositions feel. (groove/instrumental balances and important feel)
We are talking 2 or 3, 2dB to 3dB boosts in the 1-6kHz range across the mix which is quite a lot but shows that one does what is needed, not what is typical, as typical does not ultimately exist.
Limiting was virtually zero, 1 limiter doing nothing just to protect -1dBFS, this is 1997/1998 afterall, so it is a nice “hairy” mix, in that I refer to the waveform of intact transients. I will keep a 24 bit version of the mix in case it requires a louder perceived volume master at any stage. Though of course it would have been played off the project desk straight to 16 bit DAT tape back then, maybe using a TASCAM DA-20 (I owned one of these and still do). I did add a little suitable saturation across the entire mix as the 90’s trance was a little ragged sounding along with quite a thick low and lower mid range. I took a while to ensure it sounded about right, trying various options as this is not common to do in mastering. Though this is quite a different approach as I am actively trying to sound like the times gone by.
Slow progress but a decent track was made and I am satisfied with it. It is always good fun and nice to bring it over the finish line. Always finishing tracks using techniques you write about online, feels like a respectable approach rather than just words or a “do this, do that” video on the internet, of which there are so many.
I master all genres of music as a sole income
Whilst being intimate with music production of all kinds and as my spare time hobby, I am a mastering engineer for a living. If you would like your music mastered and taken to its highest sonic potential for a great price please get in contact. I work with all genres of music happily and with ease (electronic is just my hobby music making, as an audio engineer for 25 years I have recorded many types of bands and acoustic music, so it is all familiar. )
1T £30 | 2T £60 |4T EP £120 | 10-12T Album £230 – See pricing page for full details:
Pricing and payment | Mastering costs
Contact email
Barry safeandsound123@googlemail.com