Secrets Of The Mix Engineers: Peter Mokran

Take one Bollywood soundtrack, add a Pussycat Doll and a heap of new instruments, and you have a recipe for a big mix. Peter Mokran was the engineer who cooked up an international hit with 'Jai Ho! (You Are My Destiny)'.

Peter Mokran at the SSL desk in his mix room at Conway Studios.

For the normally rather world music‑phobic US and UK hit parades, the success of the 'Jai Ho!' single was a striking departure. The Hindi song was taken from AR Rahman's soundtrack for the movie Slumdog Millionaire, itself a major international success. Composer, producer, musician and singer AR Rahman is as big as they come in India, but less well known in the West.

The English‑language version of 'Jai Ho!' was the brainchild of Ron Fair, one of the US's foremost arrangers, producers, A&R men and record company moguls. Among Fair's many successes, the Pussycat Dolls are one of his pet projects, and he enlisted the help of lead singer Nicole Scherzinger for the remake, now retitled 'Jai Ho! (You Are My Destiny)'. Fair retained the most characteristic original Indian elements, notably the thunderous drums, some of the Indian voices, including the 'Jai Ho' chant, and the opening guitar/mandolin lick. He took these tracks down a semitone using Serato's Pitch 'n Time and Waves' Tune plug‑ins, and added various new instrumental tracks, including live strings. Fair and Scherzinger co‑wrote English lyrics and added a few new vocal lines. The singer, who co‑produced the new version with Fair, commented that theirs is "an interpretation rather than a remix”. This is seconded by mixer Peter Mokran, who says, "I see it as a pop song version of the track in the film. Ron put the tracks together, and came into my studio when he had finished the production part. We then worked one‑on‑one for about four days on the mix. We were aiming to create a good pop song: staying true to the original was not our major concern.”

"Ron is still a classic Quincy Jones‑style producer,” comments Mokran, "who will hire musicians, make arrangements, oversee the whole process, and bounce the background vocals together, like they did in the tape days, only Ron is doing it with modern technology. He will send me a very detailed rough mix, so he has a very clear idea of what he wants. But he then leaves the door open for experimentation. He'll listen to my mix, and he'll say, 'Wow, I really like that direction,' or 'Let's go with this direction but I also want this or that.' It definitely is a process.”

Three screenshots (above, below and opposite) reveal the entire 'Jai Ho! (You Are My Destiny)' Session.

'Jai Ho! (You Are My Destiny)'

Written by AR Rahman, Gulzar, Nicole Scherzinger, Ron Fair

Produced by Ron Fair and Nicole Scherzinger

The Pro Tools file with which Ron Fair walked into Mokran's studio one day last February was notable for a number of reasons. Consisting of 80‑plus stereo and mono music tracks, the Edit window looked like a patchwork of small bits and pieces. Only the first seven of the 17 drum tracks, as well as a synth bass track, were playing for most of the song. In addition, there were 44 stereo and eight mono vocals tracks, in effect making 96 vocal tracks. Many of them also occurred only for brief moments. Despite this enormous number of often bitty tracks, Mokran maintains that it wasn't an overly complex mix.

"Ron likes lots of little ear candies that only occur one time in a song. But it wasn't a problem to mix, because he had done his homework and organised the track well. Also, the drums kept everything together. A mix with lots of tracks that are playing from the beginning to the end is often much more complex, because not only do you have to make it work, you also often have to figure out what should be there and what should not be there. In some cases, people deliberately leave it to the mixer to figure out what should be left in. I've even had projects where I've had to replay or reprogram most of the parts. These are the projects in which I don't feel like a classic mixer, but more like a combination of A&R guy, arranger and producer. None of these things apply to working with Ron. Only tracks that will be used remain in the file, and the tracks are well‑organised and even colour‑coded, because he likes to be able to quickly see how things are grouped together. I may alter the colour according to how I want to see things. The only other thing I do in terms of organisation is to add track names in the comments box. Pro Tools tends to abbreviate the track names in a really strange way, producing lots of gibberish, but the comments boxes remain the same, so that makes it easier for me to orientate myself.

"When Ron came in, I first of all listened to his rough, which gave me an idea of the direction he wanted to take it in. There were lots of dense layers, lots of percussion, and several loops, and I wanted to make sure that each chorus was really building and that you could hear all the parts distinctly. The latter was a matter of finding the frequencies to scoop out and make things clear; if you have too many tracks you tend to get this weird, low‑mid‑range haze. Ron also wanted the same thing that we had done on a previous Pussycat Dolls song, 'I Hate This Part', which was that the main vocal constantly changed throughout the track. Having all the vocals split out over different tracks made this easier to do. Ron's ear candies also helped in making sure that no two choruses are the same.

Peter Mokran uses few plug‑ins: here the Waves Q4 equaliser is used to filter the feed to a slapback delay, and Digidesign's Digirack EQ III filters another vocal delay send."The drums were a challenge, because they were almost all loops. One of the first things I had to do was use ReCycle to isolate separate kick, snare and hi‑hat tracks from these loops, which allowed me more mix control, and also made it possible to trigger separate tracks with a kick and snare sample. Having said that, I began the mix itself with everything apart from the drums. In fact, this is how I always work. When I mix a track, I first need to hear what the instruments are doing. If I start with the drums, once I put the instruments in there I wind up having to completely change the balance of the drums. That low end that you feel when solo'ing the kick drum completely changes when all the other things come in. It also gets too boring for me to listen just to the drums. In addition, in these kind of productions the synths, and most of all the synth pads, are usually pretty close to the finished sound, so there's very little to do to them, whereas I usually end up doing quite a lot of radical changes to the drums, adding, triggering and replacing things. So I find that if I begin with the parts that will remain pretty much the same, and the densest parts, usually the instruments that provide the chords, it is a lot easier to make decisions later on about the drums.”

Mokran's approach is pretty unusual, particularly in the hip‑hop/R&B/pop vein, where almost all mixers hold that the beat and the vocals are the two most important elements, and so they start with either, or both. One wonders how he tackled 'Jai Ho! (You Are My Destiny)', because it doesn't really have chordal parts that continue throughout the song. "I can't quite remember,” replies Mokran, "but I think I started with the motif that starts the song, the Indian‑type thing with the guitar and mandolins layered and the intro drone, and then I brought in the strings, and then the vocals, at a low level, muting them in and out, and then the other instrumental tracks. Once I had all these elements balanced, I brought in the drums.”

Finally, as explained above, Mokran mixed the track to half‑inch tape on the Ampex ATR102, putting a Fairman tube stereo compressor over the mix, which is "made by a Danish guy; it wasn't really compressing but was just there to get a nice sound”, as well as a Neve 33609 compressor "to tighten up all the drum stuff and add a small amount of overall compression”. The mixer is happy to report that there was no brick‑wall limiting, observing that "people have recently come to realise that there's no good to be gained from it, because things become so mashed up that they're unlistenable”. According to Mokran, AR Rahman loved the results: "He had nothing to say but 'Wow!'”

• Drums

'Loop one kick': "This is the stereo kick drum that I took from loop one using ReCycle. Next to it is 'PM Kick Trig' a kick‑drum sample that I added and triggered from 'loop kick one'. 'Loop two — kick' is, of course, a kick I isolated from loop two. So I have three kick drums, and they are treated separately. I probably used an API EQ on each, to make sure that the three kicks are complementary; one kick would have the attack that I liked, but I'd filter out its bottom end. The one I triggered had the low end, and the third one added mid‑range. I split each kick off to a Distressor with a lot of compression, and they came back on a separate channel, which had a Pultec EQ on it to filter the overall kick sound.

"The sampled snare I added to the two snares that I isolated from the loops is on a track called 'PM Snr'. I put the sample in because I wanted a different snare in the choruses. The three snares would come up separately on the board and, as with the kick, I'd run them through a Distressor on a bus, and they came back on a separate channel with a Pultec EQ. I'd EQ the individual snares with an API 550a on the desk. I also added some reverb on the snares: there was a sort of inverse sound that's not really a reverb from a Lexicon, it's a very short room that almost sounds like the Phil Collins gated drum sound, and a short reverb from an AMS RMX. In the choruses, the snare sample would have had a longer reverb from an Eventide 2016.

"I isolated a hi‑hat from each loop: one was a pure hi‑hat, and the other also had other things on it. They just had some EQ from the SSL board. The three cymbal tracks have different cymbal samples appearing at different times. I didn't treat them.”

'Tribal drums': "That was a stereo loop from the original track. For the main sections I added lots of low end on the desk, so that the loop would be really deep and ringing. They were a cool part, and were featured quite a lot in the sections where the track gets very world‑music‑like and where the drones come in. For the rest of the time I switched the EQ on them off. I also added some AMS reverb on them for these main sections, and some SSL compressor. Like the SSL EQ, I like it on drums, but not on anything else.”

'Verby accent drum hit': "That's just a hit that comes in on beat four once in a while, and that I fed into an EMT plate reverb with pre‑delay. It's just another bit of ear candy. It also has some SSL compression and EQ.”

'Exploding drums': "There are two stereo pairs of these, also from the original track. I added a lot of compression and reverb to kind of get the John Bonham reverb sound on certain accents that I wanted to bring out. I also used a Pultec plug‑in on them, just to get a bit more mid‑range. These tracks had lots of effects and noise on them, and that's the reason why I faded them out in several places.”

'Claps': "They also came from the Indian track. It seemed like all the original tracks had really bathroom‑like reverbs, so in many cases it was a matter of carving out frequencies to take out the dull, swimmy edge. There were so many effects on the claps that I wanted to replace them, but Ron said that they were a feature of the original, so we kept them as they were. I still compressed and EQ'ed them on the board. The 'arab march' was another element from the original, which only occurred towards the end, and that I only EQ'ed.”

• Keyboards & strings

'Synth bass': "This filled out the bottom end, rather than being an actual part. I treated it with a Urei 1176 blackface compressor, to get that second and third harmonic, so you could hear it a little bit more in the track. I usually don't hit the compressor too hard on bass. I don't even look at the meter: when it sounds right, I leave it. It's more a matter of tightening the bottom and getting the right amount of tone and grind out of it. The blackface 1176 has a very grindy character, more so than the new silver ones. I also used a Lang EQ, which I think was made in the early '70s and which acts a bit like a Pultec.”

In addition to a number of hardware Pultec equalisers, Mokran used the Bomb Factory plug‑in version to process the 'exploding drums'.

'Noisy stabber': "There are two stereo tracks, and Ron wrote a comment about cranking the second one up. He really wanted to hear it! They are just tiny synth parts and I probably EQ'ed them on the desk, both for the sound and to make them sit in the track. But in general I don't treat synth sounds that much. It was the same with the other synth part, 'space noise', which was a hazy, droney thing, and which had some Behringer stereo enhancement and EMT reverb.”

'Pad piano': "Again, this was a synth, and I would not have done much to it, other than some EQ. However, I wanted to spread it out more into the stereo field to try to make it fit. I also used my Behringer Stereo Enhancer box, which is modified by Jonathan Little of Little Labs. Without affecting the phase too much, it does more or less what the QSound plug‑in does.”

Strings: "Ron will normally keep all string tracks separate, including a room mic, but in this case he had bounced the strings down to stereo. I usually run strings through my two Neve 1084 units. They have just a three‑band EQ, but seem to do the trick. I usually pull out some low mids and add a little bit of sparkle, trying to get the strings to speak in the track. I also added reverb to the strings from an Eventide 2016.”

Synth strings & 'string stinger': "The synth strings played a counterpart to the real strings, and I would have used an API 550b EQ and the same reverb. The 'stinger' was just a little part that comes in and out, and that I had separately on the board and would have applied different EQ and panning to.”

• Guitars & mandolins

Main acoustic guitar: "This was a nylon‑strung [guitar], on which I used a Distressor compressor to help even the notes out and make them pop out a bit, and a Neve 1084, to get rid of the woolliness in the low-mid that you often get with nylons and add some high‑mid and top end.”

'Indian reinforce main lick': "As the name indicates, that was taken from the Indian original. Again, all the reverb was dulling it, so I used a graphic API 560 EQ to really get it in place.”

Mandolins: "These also were from the original track. They are playing that intro theme and the fast licks that lead up to the verses. Mandolins can be too piercing and bright, so I used an [Empirical Labs] Fatso, because you can adjust how much top end it takes off from the spikes while leaving the brightness the rest of the time. The main thing with the mandolins was panning them to the sides and getting them to sit right without them being too piercing.”

Acoustic and electric rhythm guitars: "The electric plays a clean arpeggio in the choruses. I added a detuned stereo delay to give it a little bit of bounce. One side would be up eight or nine cents, and the other side would be down eight or nine cents. When a track is as dense as this one, you never hear a delay on a guitar, so detuning it slightly and adding a reverb to the delay means you can actually hear it in the track. I used an Eventide H3500, which is cool, because you can do the detuning, delay and reverb all in the same preset. I probably also used my Distressor, just to make the notes kind of pop out and to give me a little bit of that '80s, clean, compressed guitar thing. On the acoustic I used a Fairchild compressor and an API 550, to thin it out. The Fairchild has a thick sound, which means that it didn't sound too thin after applying the 550. The wah‑wah downbeat is a really subtle electric guitar that didn't make the final mix.”

• Vocals

Lead vocals: "Nicole's lead vocal occurs in the verses and is split over three tracks: a main lead vocal track and two support vocals, 'lead vox L' and lead vox R', that double up towards the end of the verses. As I explained before, we varied the sound of her lead vocal throughout the track, using different effects and EQ. There is an AMS delay and a TC 2290 delay on her vocal that feeds an EMT 250 reverb to get a very washy, airy sound. The 250 is the very first digital reverb, it's a big, giant monster! The 2290 was set to a dynamic delay, according to where the word stops, so you get a rhythmic effect. The left and right tracks had similar treatments, but with different EQ from a GML parametric EQ, carving out a little of the low‑mids, which generally tend to build up on vocals. I also used a [Digidesign] Digirack EQ plug‑in, cutting below 200Hz and above 6k, on a certain phrase in the vocals on which I wanted a filtered delay.

"I didn't use a de‑esser. In general I don't use them. Instead I have a Fatso across the vocal bus, and the Warmth control on that takes the spikes and the esses out. If it's really bad, I'll use a plug‑in, but that's rare. I used one other plug‑in on the vocals, the Waves Q4 parametric EQ, just for one phrase, where Nicole does some sort of Indian ad lib, and I wanted to have this Elvis Presley‑like tape slap effect, just for that one lick. So I made a separate track, and used the Q4 to filter it very dramatically, and then I sent that through a Marshall Tape Eliminator, which is a kind of tape slap box, to make the effect even more radical. I created an auxiliary track for it, which doesn't show the waveform, but just a line to indicate the level. Nicole did another vocal track called 'opera vox hi', which is a really high vocal that we buried in the background with tons of EMT 140 reverb, with a pre‑delay on it, so it sounds more like an instrument than a vocal.”

'Jai Ho' and male hook vocals: "The 'Jai Ho' chant had a lot of EQ on it. All the male vocal parts had lots of effects on them, so it was a matter of trying to carve out the mid‑rangey, bathroom‑like 'verb. On the 'Jai Ho' track I then added my own reverb, which I filtered below 800Hz, so it had a very sparkly sound. There are another six tracks with male hook vocals from the original, and I used just a few choice licks from them. Also in this case I tried to take off a lot of the effects. Funnily enough, they had used Auto‑Tune on these vocals, really set to stun, to create this vocoder‑like effect. It made me laugh, because it was like: 'Wow, they're doing this in India as well now!' I basically EQ'ed these vocals, and then I used a Tube‑Tech CL1B compressor and an Avalon 2055 EQ, plus another TC 2290 delay which fed the EMT 250.”

Chants and drones: "Nicole did a lot of vocal overdubs, including three stereo tracks of chants and four stereo tracks of drones. I just tucked all these elements in, really, really low, and I used a Pitch 'n Time plug‑in to move one of the drone tracks an octave down. It's not very accurate, but it's not meant to be, it's just a drone. There's a little bit of the Eventide 2016 reverb on the drones, as well as the EMT 250 going into a 2290 delay. The reverb is quite long, so this delay/reverb washes in and out, rather than it being a defined sound.”

Backing vocals and 'ear candy': "Nicole did many tracks with little ear‑candy snippets of vocals, like 'aaah‑breathy' and 'Hut‑groan' that come in at various points and are panned off to left and right. Mixing them was mainly a question of finding space for them and not putting them through the middle. All the red tracks, from 'Verse Answers' to 'Hook "oh‑oh answer lick”' are what I consider the main backing vocals. They would come up on the console in pairs, and I would then bus them to two tracks on the board, doing my thing with the Fatso compressor on the bus, and then to a Tube‑Tech CL1B compressor on the insert of a separate channel, which also had an Avalon 2055 EQ on it, all just for the sound. Those units created the overall background sound. The panning on these tracks is pretty basic, because they were all stacked up so much that they had to be left and right to some degree.

"In addition, there are three stereo 'catch me melody' and 'four part vox' tracks which came up on separate channels on the board, with maybe slightly different EQ on each, and then to the same effects as the main backing vocals. Finally, there's a stack of nine stereo 'oh…' tracks, which provided another layer that happened once in the song. Together they almost work like a synth pad, done by the human voice. My work there would have been to really thin them out, and get them as airy as possible, using an EQ. They would also have had the Fatso, Tube‑Tech and 2055, maybe with some more delay.

"Do I have a mix recall sheet in front of me while talking to you? No, I can remember all these details from memory. I can usually remember what I did on tracks I mixed, and especially in this case, because we spent a lot of time on it. As I said, this mix took four days! Ron is a real perfectionist, and we were exploring different avenues to get all that we could out of this track.”