Khiyo's new music video 'Bareer Kachhe Arshinogor' is out now, please check it out!
https://www.facebook.com/video.php?v=10153021640400390&set=vb.153064630389&type=3&theater
or
https://www.youtube.com/watch?v=d3JqMmVA7XA&feature=youtu.be
As I did with 'Murshidi', I would like to do some mix analysis to show you what I've done during the mixing, things like choice of plugins, effects and master bus processing etc.
19 audio tracks, 39 including AUX, parallel processing, effect return, sub mixes and main master bus.
It's relatively a small session compared to some of other songs from their debut album.
I recorded the drums and tabla at London School of Sound, and all the rest were recorded by Dave Mountain at the Attic Sounds. All the session were kept 24bit 48kHz throughout.
My choice of DAW was Pro Tools 10. If you want to know the basic setup of my mixing template, please see the previous post of Murshidi mix analysis.
You will notice that on this mix session I've used Focusrite's Liquid Mix 32 quite a lot.
Let's start from looking at the kick drum.
Kick:
Kick was recorded with only one mic. It sounded good during the recording session but once I started mixing, I felt that it needs more definition/clarity especially when the song hits choruses.
I didn't have any re-triggering plugins in RTAS/AAX format at the time of mixing, I had to export the kick audio file to Logic, use their audio to MIDI function, then brought back the MIDI file into Pro Tools, using NI's Kontakt to audition my kick samples and finally I chose one that blends well with the original kick. I know it was a very tedious and time consuming work... but worth it!
The original kick had a gentle compression from Liquid Mix (LM), the second kick (sample one) had Massey's Tapehead driven moderately to add some dirt since it was a very clean sample. Then both kicks went to Kick bus and processed together. Gentle compression (LA-2A emulation) and EQ (EQP1 emulation) from LM, another compression form NI's VC76, very tiny adjustments from Digirack EQ3 and peak control from Waves L1.
On this song, the difficulty was to control the frequency relation between kick, tabla and bass to be heard nicely without over crowding the low-end. More on this later.
Small amount of kick bus were sent to drum reverb.
Snare:
Snare was also recorded with only one mic at the top. During the mixing I divided into 2 separate tracks, one for normal hit, one for rim hit (shot) so that I can set the EQ, compressor and reverb send differently.
As Murshidi, there were too much hi-hat and cymbal bleed on the snare track so first insert was the Digirack Dyn3 expander/gate to clean it up. Then I used Dyn3 compressor to shape and even out the dynamics of each hit. EQ3 was used for HPF and a small cut at 284Hz. API 550A adding nice top end at 12kHz by 4dB in shelving mode since they weren't cutting through the mix enough especially at those busy part of the song. L1 to control some of the peaks during fills. Both tracks were sent to drum reverb with some automation; snare send increases during the chorus.
Hi-hats, Floor tom and OH:
Just a HPF for the Hi-hat. For the floor tom I've used Massey's CT4 to enhance the sustain of the sound and also to control the dynamics. CT4 is one of my favourite compressor.
For the OH, LM with Joe Meek SC2 emulation selected doing very gentle compression and Neve 1073 EQ emulation, DDMF 6144 to cut some low-mid, then L1 for the peak control.
Drum bus and parallel process:
Gentle compression by McDSP's 6030 in MOO TUBE mode and L1 to control the peaks.
For the parallel channel, Dyn3 compressor to squash quite hard then Tapehead driving quite hard to add more saturation. The drum reverb was PSP's EasyVerb, one of favourite drum ambience plugins.
Tabla:
Tabla was recorded in 2 track one mic for Baya, one mic for Daya.
I have used standard Digirack delay to make Daya stereo by adding 38ms delay to only right side, then both tracks were routed to Tabla bus. First plugin was Massey CT4 doing nice mild compression, EQ3 for some minor cuts, HOFA IQ-EQ in dynamic EQ mode to tame around 79Hz.
API 550A to boost 10kHz and cut 100Hz. Then Dyn3 compressor was used with side-chain key input from Kick bus to duck whenever kick hit happens by small amount (2-3dB max).
Then the usual L1 to control the peaks.
Only the Daya track was sent to the room reverb.
Upright Bass:
Recorded by Dave Mountain at his Attic Sounds and came in 2 mono tracks, bass bottom and bass top. I assumed one near the bridge/f hole area and one up the neck somewhere. Both track were routed to Bass bus. Tapehead for adding some thickness and compression, EQ3 cutting 75Hz also cutting at 416Hz to clean up some room resonance. LM dbx 160S emulation doing moderate compression. IQ-EQ in dynamic mode to tame further at 75Hz, and Expander mode at 114Hz to boost that particular note. It looks like there was a strong peak at 75Hz and a weak bass notes at 114Hz. Then 6030 in Opto-C doing another mild compression and L1.
Some notes were sent to drum reverb with automation.
Acoustic guitar:
Also recorded by Dave Mountain. came in 2 mono tracks, one brighter sound (AKG C414) and another one slightly darker sound (it was labelled Ribbon mic). After tried many configurations (2 mono tracks at same panning position etc), I've decided to pan them 100L and 100R plus delaying the C414 track by 1472 samples (approx 30ms) to create a placebo double effect.
Both tracks had CT4 to reduce the dynamic range before they were sent to AGtr bus.
At the AGtr bus, LM LA-2A was used to control the dynamic range and 1073 EQ for general tonal shaping. Brainworx bx_control to reduce the stereo width and used mono maker to collect the low frequency to the center of the mix. The placebo double effects made the guitar sound too wide on the stereo field so bx_control will bring the width back but the delay effect is still audible. IQ-EQ was used on 3 frequency areas to reduce some sharp whistling resonances. IQ-EQ is great for this kind of job, then L1.
At the second chorus, instrumental section and the last chorus, a parallel channel will come up via automation adding more excitement. On this parallel channel, heavier compression, heavier Tapehead were used. Please note that the pan pots were both set to 0 (mono). This is to fill up the center space during the instrumental secion and also to compensate the signal loss caused by the placebo double effect when it's played in mono, though, I did tweak the delay time sample by sample to choose the least loss delay amount when it's combined to mono.
AGtr bus was sent to room reverb.
Viola:
Viola was recorded by Dave mountain. It looked like they had 2 separate sessions for Viola.
One session was recorded with 2 mics (second half of the song), one session was recorded with only one mic (early part of the song). I duplicated the single audio and made a double for the early part. On the duplicated track I have delay the signal by 1497 samples to create a placebo double (just like the AGtr) and used Tapehead to change the tonality to beat the mono summing issues. The other viola session were fine as they were and sent to Viola bus. R compressor was used to gently control the dynamic range, EQ3 for minor sculpting, API 550A adding some air at 12kHz in shelving mode.
Viola was sent to room reverb and also to a longer reverb (see effects section).
Vocal:
Recorded by David Mountain and came in 2 mono tracks. One I don't know what mic was used and another one was labelled SM7B. After auditioning those 2, I've decided to use the first one as the main and SM7B as a parallel boost. First plugin was Waves RVox. I don't usually do this but probably it felt good! But then there is CT4 after it. I must felt it wasn't enough dynamic range control done by RVox so I inserted CT4 for further control. EQ3 for some minor cuts, DDMF's 6144 doing some nice high-mid to high boost adding some nice air to Sohini's voice. IQ-EQ doing some frequency control on low-mid to mid range.
Vocal was sent to a dedicated vocal reverb (Waves TrueVerb) and a stereo delay (Waves SuperTap).
Effects and side-chain ducker:
Lead vocal was sent to TrueVerb. It looks like I chose New York Plate preset and I tweaked from there. Also sent to a stereo delay and this stereo delay is also sent to Trueverb to blur the delay sound. IK's CSR Room was chosen for the short ambient type reverb, Air Reverb was chosen for the longer reverb adding some tails to viola. Side-chain ducker was setup using Dyn3 expander/gate. AGtr and viola were sent to this AUX and whenever Sohini is singing, AGtr and Viola ducks down in level by whatever I feed to this AUX (please see Murshidi analysis for more details on this technique).
Master bus:
Throughout the mix I've used Slate Digital's Virtual Collection in Brit 4k. Then LM with SSL FX G384 emulation doing some 'Glue' compression 2-3dB max. bx_control making anything below 152Hz in mono, and finally Slate Digital Virtual Tape Machine doing the final finish touch.
Mix summary:
This was one of the longest songs on the album and the dynamic range was very wide throughout the song so I had to be careful not to lose the natural dynamics that musicians intended in the studio.
There are some favourite plugins that I tend to go to but as you can see, I like to try many different plugins to see how they sound. In this session Focusrite's Liquid Mix 32 was heavily involved despite there are some known issues with the low frequency with impulse response problems, but I really like the sound and didn't hear any artifacts so I guess all is good! Also when I started the mixing on this song, I was still using a Mac with dual core 2.16gHz so the DSP on Liquid Mix really helped me getting the mix to progress.
Thanks for looking!
YS Sound Production
