Shush Recording
Up until recently high quality recording studios had to contain expensive Tape Reels or ADATs, mixers, effect racks, patch bays, sound modules and mastering equipment. These days complete, digital, studio quality recordings can be done extremely cheaply with common PCs alone (well almost), the only restrictions being PC speed and the quality of software. By using software alone, a common PC can serve as sound source, digital recorder, mixer, effects rack, CD burner and internet connected MP3 distributor. Hopefully this will change the way people listen to music, knowing that there are good quality alternatives to the predominantly celebrity/image oriented shite on offer from the major record companies.
The Studio Setup
All Shush recordings were done with Cakewalk 9 running on a PC and SBlive Platinum sound card. The soundcard supports the importing of very large single instrument sample files (soundfonts) that have proved invaluable for drums, bass, piano, etc. Softsynths like Vaz 1.7 were also useful for some keyboard lead and pad sounds. External sound sources like vocal mics, guitars and midi controlled sound modules (OB2 Squared, Virus Rack) that were connected via a studio master 12/2 mixer.
The Sequencer
The sequencer software is the heart of any studio being the multitrack recorder for digital audio (vocals, guitars, etc) and midi controlled sound sources (synths, samples, etc). From the sequencer you record and edit individual tracks, apply effects to each and eventually mix all tracks down to a single final stereo track. This is then mastered for release on CD or as MP3.
At the heart of the Shush studio is Cakewalk9. This is a superb piece of software being a complete, virtual recording studio. Midi, Audio, Effects and Softsynths can managed side by side using one mixer screen or separate track screen. Multiple DirectX, VST* or Midi effect plugins can be assigned to individual tracks or accessed via virtual send and returns. Soundfonts are especially easy to use; desired soundfonts are loaded into the project and assigned to respective midi tracks. I used to use the alternative Cubase but found it so badly designed and bug ridden that I preferred Cakewalk even though it doesn’t have native support for VST.
(* VST plugins for Cakewalk via VST DX Wrapper Lite)
Freeware plugins
Though Cakewalk and Cubasis come with some effects plugins; extra equalisation, compressor and mastering plugins are essential. Although you could spend a lot on extra plugins there are some excellent freeware and shareware VST and DirectX plugins available. The quality of many is comparable to more expensive commercial stuff.
Maxim
freeware plugins
Fraser's freeware plugins
Vellonet freeware
plugins
Bo's freeware plugins
DS-GE1 Freeware Guitar
Equaliser
Freeverb
Digilogue shareware plugins
Smartelectronix freeware
plugins
The DirectX Files
Recording Vocals
Ive found recording with cheap dynamic mics a waste of time. I currently use an AKG c1000s capacitor mic which includes a couple of attachments to change responses, i.e. cardioid (near field), hypercariod (sound reinforcement) or presence boosted. Ive found that I need very little EQ to get a good vocal sound though a subtle 2kHz boost can add to clarity and presence. I personally like to use a lot of compression and tube warmth applied directly to the vocal track via DirectX and VST plugins. Subtle reverb and sometimes flange or delay are mixed into vocals as virtual auxiliary effects. No external rack effects are used.
Recording Guitars
Getting a decent guitar sound with PC based home recording can be a real problem, especially if you have neighbours who might complain when you mic record a properly cranked valve combo. The alternative is direct input (DI) guitar recording where you can record what you like when you like. However a DI guitar sound is never quite right and you have to colour it with effects. In theory this is possible with plugin based amp simulations like Revalver and Maxim Combo though the current unworkable latency between data going into the PC and processed data being heard means the guitarist is forced to play clean while recording. The future however looks promising as a new driver model for Windows (WDM) will optimise the transfer of data from input to output considerably improving the potential of for 'hear as you record' plugin effects.
In the mean DI recording is best achieved with the Line6 Pod, which provides remarkable modelling of famous tube guitar combos like the Fender Twin, Vox AC30 and Marshall JMC800.
Next I E.Q. the guitar track mainly to reduce
sounds below ~80Hz, this will help the guitar sit better in the final mix.
Double tracking with different guitars and plugin effects like reverb,
stereosisers and harmonic exciters can help fill out the mix, though it's easy
to overdo it.
There are a plethora of other freeware guitar effect plugins that do the job of
standard pedals like chorus, flange, phasers, autowahs, delays as well as more
adventurous sounds like resonant filters and ring modulators.
Midi tracks
Keyboard midi tracks form the bulk of the music especially drums, bass, piano, strings and synth sounds. Originally midi sequences were the preserve of the dance scene though these days more natural music styles can be created with the help of functions like randomise, swing and event editing.
I'm a big fan of soundfonts, especially the mulit-megabyte single instrument fonts that can be found on the web. There are some excellent drum, bass and piano soundfonts as well as very usable synth sounds. With drum soundfonts I'll add reverb via the SBlive's own environmental audio effects. The best bass soundfonts can be used almost clean. Synth soundfonts work particularly well with midi effects like delays and appregiators. The beauty of soundfonts is that they are very easily managed, processor friendly, multiphonic and extensible. As well as accessing thousands of high quality soundfonts over the Internet you can also create your own based on recorded samples or rendered outputs from softsynths.
Using softsynths in real-time can be a lot of fun though also very problematic with a sequencer due to heavy processor requirements. An alternative is mixing in external synth modules, though these are still very expensive. I'll often find myself preferring to use soundfonts to external synth modules.
Noise Issues
A key problem with home-based PC project studios is unwanted noise. As well as the usual domestic noise sources likes clocks, pipes, traffic, etc a key source of noise is the PC itself with includes whirring fans and disk drives. By using extension leads for monitor, keyboard and mouse the PC can be moved into a cupboard or separate room. Another noise source is RF interference from the Monitor. This can be a real problem when using outboard effects like the Pod though a monitor can always be switched off during takes.
Mixing
I use Cakewalk for both mixing and mastering. A real mixer is not essential, though they do look good next to a PC and so have some muso pose value. All mixing can and should be done virtually on a PC, though a real mixer is handy when controlling several external sound sources. It also saves having to plug and unplug things and reset input levels for separate sound sources.
Mixing is essentially a process of getting the relative levels right though it can also be used creatively to add to a song. For instance sticking a delay at the end of the vocal phrase or panning a guitar solo.
Setting instruments in the stereo field is one of the hardest parts of mixing though some of the job is already done when using drum sound fonts which usually have parts of the kit set in the stereo field appropriately. Low frequency instruments like bass sit best in the middle of the stereo field where as high frequency components like vocals, keyboards and guitar can be moved to the sides. I tend to keep the lead vocal relatively forward.
Another problem with mixing is trying to avoid clipping. Audio is recorded with 16bit numbers, which correspond to a dynamic range of 92dB. As it happens or ears can detect subtleties and sub harmonics corresponding to 118dB, more than the original designers of CDs allowed for. To have the largest dynamic range the audio signal must be as large as possible because quite signals can loose definition when boosted owing the limited resolution of numbers representing them. However, when the signal goes beyond the allowed 16bits horrible clicks and artefacts occur as the signal is clipped. This especially happens when gain boosting EQs and effects are added. Trying to get loud audio whilst avoiding clipping can be hard work. In the long term 24bit recording offers more dynamic range and headroom to mess around though can be a real drain on computer resources. Though bear in mind that most classic albums, including relatively recent ones were recorded with much poorer quality than available to average PC owners today.
Mastering
After the music is mixed down to one stereo track there is a final mastering stage that usually should be done in a separate session. Mastering involves setting the overall level and equalisation. As well as multiple band E.Q.s there are also plugins to maximise various audio components to give the final master more punch and depth. Master compressors or ultramisers plugins add more punch and authority to a mix. Stereo enhancers can add more spaciousness and depth. Tube and tape simulations add subtle warmth and compensate for the less classic and natural feel of digital recording. The final master can hence be burned to a CD or converted to MP3.
Recommended Software
You don't have to spend a lot of money to get good results. Although some audio software can be ridiculously expensive, there are developers who provide high quality applications at affordable prices or even as freeware. Most people start with a basic midi/audio sequencer application like Cakewalk, Cubase or Logic Audio and add all sorts of extra effect 'plugins' as they go along. (N.B. ‘Pro Tools Lite’ with Plugins will soon be free from www.digidesign.com).
Software based sound sources are also a useful addition, notably software synths that run alongside, and are controlled by the sequencer via midi channels. Software synths are very cost effective relative to their hardware counter parts (e.g. JP8000s, Nords, Moogs, etc) and many are fantastic fun and can sound surprisingly authentic (notably Dreamstation and Reaktor). However processing demands and Window's poor I/O latencies currently restrict their use in real time when together with digital audio, effects, etc. There are workaround solutions like recording synth output to audio tracks and shifting the wav data into corrected time however this can be a real pain relative to controlling processor independent outboard synths. Hopefully the new Windows Driver Model (WDM), Cakewalk's promise to support DirectX plugin synths and the appearance of >1GB CPUs will improve things.
Vaz 1.7 Softsynth
This is a real time virtual analogue synth that
cost me about £20 from Software Technology
three years ago and is still my first choice virtual synth. Its has two main oscillators,
multimode filters, envelopes and LFOs can be routed in loads of different ways
to give very strong, fat sounds ranging from classic vintage to completely
weird. As well as the usual saw, pulse and triangle waveforms one oscillator
can import and loop wav files which makes it superb for home made vox sounds.
It's relatively low processor overhead makes it easy to implement within
Cakewalk using a loopback driver.
I use Vaz on the intro to 'Lucid Dreamer' and various other tracks.
The are a number of other softsynths and samplers out there ranging in price
and quality. The best sounding my may not be the easiest to use with a
sequencer. (Dreamstation, Dynamo
and SynC are shithot). For
more try sonicspot, shareware music machine or synthzone.
Recommended Hardware
OB3 Squared - Organ Module
An excellent Hammond Organ module from Oberheim. I also have a real Hammond X5 and Leslie 820, however the OB3 squared sounds as good, you can D.I. it and is much easier to lug around.
Virus Rack - Synth Module
I put a deposit down on one of these mother f***ers in June 2000 and I’m still waiting. Access are renowned for producing some of the best sounding virtual synth engines. The Virus has loads of class and you can put a guitar through it, so they’ll be lots of wacky Steve Hillage-esque weirdness in our next set of MP3s.
Epiphone Sheraton
This is my first choice guitar for recording being an ideal compromise between the rich, tarry-ness of Les Pauls and lighter, more percussive sounding Strats. The hum-bucking pickups prove invaluable when recording in an environment where there lots of potential RF interference from monitors, adapters, etc. A similar guitar, the Epiphone Casio features on many Beatles recordings being an ideal rhythm guitar. I also use a Gordon Smith GS-2, Burns Marquee, Clarissa UK 100 and some learner stuff. Ive done some stuff on British Guitars via
Line 6 Pod2
The Pod is ideal for DI recording. The Pod uses physical modelling to emulate famous guitar combos including tube overdrive, speaker and miking. The quality of the simulation far exceeds anything Ive heard before being a serious alternative close miking of real combos. Although in the long run real-time versions of amp-sim plugins will be used, in the short term the Pod is easily the best solution for recording guitars in a project studio.
Links
Soundfonts
Hammersound
Melange
Sound Source
The Sound Site
Vintage
Dreams
Music Software
Sonic Spot
Hit Squad Musician's Web Centre
Recording Tutorials
Home Recording Site