Kate Bush - Running up that Hill - main synth sound

[Thoracius]

I gave it a shot recreating this sound on my microKORG XL.

https://www.youtube.com/watch?v=y2c97a76NEI

I feel like I got pretty close, considering most of the character of the original comes from the weird pitch-shifting artifacts of the digital sample, which is hard to emulate with subtractive synthesis.

[muzines]

Not bad, if it’s just subtractive synthesis being used...

[Scramble]

If I was playing in a Kate Bush cover band I'd be pretty happy with that.

[ken long]

So glad to see this epic thread revived!

[muzines]

There's a new blog post on the sounds of Running Up That Hill, which seems to come to the same conclusions as we came to in this thread:

https://reverbmachine.com/blog/kate-bush-synth-sounds

[muzines]

Some additional info, which backs up most of my earlier findings (and confirms the Quantec frozen reverb) and is worth posting here.

From an interview with Kate Bush, International Musician, Oct 1985:

RUNNING UP THAT TRACK
(The making of the single)

Running Up That Hill all started off on the Fairlight. I just found this sound that I thought was a brilliant sound for the riff, and I knew I wanted a military drone-like atmosphere that would surround it all. I asked Del to set a Linn pattern—I sung him the part—and he got that together, and we set that to play the pattern round and round and I worked out the vocal that would go over the top. The lyrics for the first verse came straight away. I stuck it down like that with the voice, Fairlight and drum machine on my home eight track.

"Then we started work on the drone. We froze a three or four note chord from the Fairlight into the Quantec. Already it was quite there, so the next stage we transferred the eight track to the 24 track and we kept the original Fairlight; I added new Fairlight playing the same part because the problem was the sample was actually very noisy. It was called a harp, but it wasn't one. Eventually, although the Linn pattern was used for a long time, we actually ended up putting it down again over more tracks.

"Then we got Stuart (Elliott, drummer) in, and because the Linn pattern was so full in itself it was just a matter of him sticking a snare down. The original Linn pattern had a snare down, but it's not very good and needed bolstering up. We needed something at the end so he put down some very dramatic fills that were a combination of three different sounds, including something on the Fairlight.

"Then I put the backing vocals down, which apart from the weird ones at the end were done at demo stage, but we re-did them. Paddy (Bush, brother) came in and put down some balalaika on the choruses to give them sort of push, Del put some bass down that was very much in there with the toms to emphasise that rhythmic feel, then it just needed something at the end. I put some freak-out voices over it all and all it needed was a guitar to finally build it."

She also mentions in a few other places in the piece about adding real drums over the top of the Linn, including ethnic drums and other things.

[wdsteele]

Ive been reading a few other archive peices tbanks to the mu:zine site and it really struck me just how creative Fairlight users had to be .
Also interesting how they all seemed to express some level of frustration around the technological limitations , something I doubt is much of a concern now a days .

[muzines]

Ive been reading a few other archive pieces thanks to the mu:zine site and it really struck me just how creative Fairlight users had to be .

Indeed! I'm amazed anyone ever got it to sound good, especially the first one!

Also interesting how they all seemed to express some level of frustration around the technological limitations , something I doubt is much of a concern now a days .

If you used one of them, you'd realise how frustrating and limiting it really is - though at the time, it was a marvel because it let you do formerly impossible things!

[Tomás Mulcahy]

There's a new blog post on the sounds of Running Up That Hill, which seems to come to the same conclusions as we came to in this thread:

https://reverbmachine.com/blog/kate-bush-synth-sounds

Brilliant info, especially about the drum pattern. I struggle with those, he seems to have nailed the feel here. The clipping on the cello sound seems to have thrown the timbre off IMO.

I don't think he's right about the Fairlight sound on Army Dreamers, it sounds more like Cello 2 again, layered with something windy. It's hard to tell really because, as you say that's the thing with the Fairlight it's so limited. After a while everything sounds like the same "nasal honk" as Wiff once described it

[nathanscribe]

There's a new blog post on the sounds of Running Up That Hill, which seems to come to the same conclusions as we came to in this thread:

https://reverbmachine.com/blog/kate-bush-synth-sounds

Brilliant info, especially about the drum pattern.

It's good but his drums are way off. The original uses the low tom, with an occasional hit to the left only, and the kick/snare are on 1+3 and 2+4 respectively. The tom does all the filling in.

[Tomás Mulcahy]

I was trying to be nice . He's done it all on the kick, but has the right idea about the "aheadness"IMO. Reminds me that I must see if I can nail the portamtento with the CMIV. Kontakt did not.

[DunningUpThatHill]

The blog post is mostly drivel, the part about the drum timing is certainly no exception. The drums aren't "off-grid" at all. My guess is that the blogger has been analyzing the first 2 bars (as those are the closest you can get to a clean drum loop, disregarding the ever present drone of course) in isolation, not realizing they're wildly non-representative.

See, if you add a steady drum pattern that runs in perfect sync with the original, they line up all the way except for the first two bars. When the LinnDrum on the original kicks in, it's very late (approx. 1/48), skids in like a drunk on rollerskates, and over the first bar and a half it speeds up then slows down before settling into the grid. The first bar is all out of whack, on the 2nd bar it's a matter of milliseconds off. To complicate things, Stuart E is trying to play acoustic snare over the Linn, so his first hit is about as off as the sync, but much like the Linn he course-corrects quickly.

So what's going on here? Well, tape sync was no exact science in the early 80's, and most devices needed a lead-in of one or two silent bars. But Del hadn't composed a full song here, there was just a single pattern that probably started playing immediately, so what you hear on the first two bars is the LinnDrum struggling to home in on the correct tempo. Ergo, if you isolate that first bar and trim it so that the first bass drum hit aligns with the downbeat, and then look at this bar thinking it's representative of the entire song, everything will be off-grid and the snare will be miles ahead.

[DunningUpThatHill]

Also, about the CELLO2 sample: Close but no cigar, and here’s why.

There once was a Fairlight sample called TRAMCHLO. It co-existed with another sample named CELLO2 that has nothing to do with this song. TRAMCHLO is present on the earliest original disks available online, dated December 29, 1979. And it was still present in a version of the factory library from February 22, 1984. This may or may not have been Revision 1.2. At some point between that date and the end of 1985 (can't find an exact date), the factory library underwent an extensive overhaul and many samples were renamed and/or tweaked. TRAMCHLO is one of those samples. The second half of the sample was muted, likely to suppress the heavy aliasing noise. The attack may possibly have been tweaked too. It was then renamed CELLO2, and appears under that name in subsequent versions of the library, such as the widely available Revision 1.3 from 1985 (the one that ships with Arturia CMI V).

Now, it's documented that the RUTH recording was finalized on December 6 of 1983 and then sat in the vault for 20 months while she completed the rest of Hounds. So unless Kate owned a flying DeLorean, she can't have been using the 1985 Rev. 1.3 CELLO2 version of the sample. It must therefore have been the original undiluted TRAMCHLO from 1979. And that’s the key to finding the other important RUTH sample (The cello is only half the story), because that sample was ditched in future revisions, thus doesn’t exist on any disk sets where TRAMCHLO is named CELLO2. More on that sample later, but to summarize: Like I said, Cello2/Tramchlo is only half the story. Think of it as the sampled PCM attack of a D-50 sound. On its own, it's way, way too short to do all the jobs that it needs to do on RUTH, and looping it won't get you there. You need the other sample (that does a job similar to that of the subtractive synthesis half of a D-50 sound).

[Tomás Mulcahy]

By gum you're right. Cello2 is Tramchlo with clipping distortion and an earlier fade out. It sounds duller too. However, both are in the Arturia CMI V library.

They'll obviously both sound the same with frozen reverb though. Plus like I said earlier, the IIx is so lo-fi that after a while things start to sound the same.

I think when the library was re-jigged (by Peter Weilk I believe) he recorded all the sounds back to tape so they could be tuned to concert pitch.

[DunningUpThatHill]

They'll obviously both sound the same with frozen reverb though. Plus like I said earlier, the IIx is so lo-fi that after a while things start to sound the same.

By ”frozen reverb” you could mean two things.

1) The drone sound present throughout the song. Yes, that will obviously sound the same whether CELLO2 or TRAMCHLO is part of the signal fed into the reverb. The thread has it mostly right, except there’s more than just the two Tramchlo notes, and also that the drone with its infinite reverb tail is played back at half its recording speed. Play the notes one octave higher, feed that into the reverb and then play that at half speed — not only does it add the required dreamy, otherworldly lustre but it also gives the reverb algorithm a more chunky quality - the grains go from sand to about kitty litter sized, and that brings it closer to the original.

2) The theory put forth by blogger ”Reverb Machine” that the main synth sound on RUTH is CELLO2 plus its own reverb tail. No, that’s just patently and laughably wrong, just like the drum analysis. In the same blog post, this self-appointed expert on Kate Bush sounds bumbles ”Hounds of Love (…) shows the analog CS-80 side to Kate Bush”, referring to the main string sound on that song. Um… okay, except Del Palmer is on record in a BBC documentary talking about that very sound (which he refers to as ”the wash”) and how it was made with a Fairlight factory sample.

Like I said, the missing part is another sample from the same 1979 disks.
I tracked it down back in 2019 and was going to write a crazy ambitious exposé on the RUTH samples, but then this thing called a full time job got in the way… I’m gonna try to find time to dust off the project in the coming weeks.

[Arpangel]

Why?

[muzines]

Hey DunningUpThatHill - thanks for contributing to the thread with some excellent additional info!

The blog post is mostly drivel,

Yep I agree, and all the similar things this guy has done have been similarly, er, well not exactly close to the originals....

Also, about the CELLO2 sample: Close but no cigar, and here’s why.

This is great, and makes sense. So CELLO2 is a later slightly edited version of TRAMCHLO? But it's the same original sample recording? I'll have to have a play with this when I get some time (currently I'm in the middle of a deep dive into another very classic record, and revealing some unknown secrets about that... I love this stuff!)

Cello2/Tramchlo is only half the story. Think of it as the sampled PCM attack of a D-50 sound. On its own, it's way, way too short to do all the jobs that it needs to do on RUTH, and looping it won't get you there. You need the other sample (that does a job similar to that of the subtractive synthesis half of a D-50 sound).

Presumably you mean the main lower chordal part playing the regular chords is layered with something else? Very possibly, I wasn't really looking too closely at that at the time - I was really trying to answer my own long-standing questions about the source sounds, rather than explicitly trying to perfectly nail/identify everything (which I mention during the earlier posts) - but all the additional info is great - thanks!

1) The drone sound present throughout the song. Yes, that will obviously sound the same whether CELLO2 or TRAMCHLO is part of the signal fed into the reverb. The thread has it mostly right, except there’s more than just the two Tramchlo notes, and also that the drone with its infinite reverb tail is played back at half its recording speed. Play the notes one octave higher, feed that into the reverb and then play that at half speed — not only does it add the required dreamy, otherworldly lustre but it also gives the reverb algorithm a more chunky quality - the grains go from sand to about kitty litter sized, and that brings it closer to the original.

Yes, I love the character of the wash, and in my earlier demos once I found out what was doing it (which was the revelation), I did mention that the record is grainer and chunkier than my examples, and that I wasn't trying to *exactly* get the sound, but just reveal more or less what was going on to my satisfaction. I put that down to the inherent character of the Quantec's tail, but not having one I could try, it was obviously not possible to check (come on Relab, get your Quantec plugin finished!).

That it was half speed makes total sense to make it grungier. How did you come to this realisation out of interest - did you try it yourself, or find that info out elsewhere?

except Del Palmer is on record in a BBC documentary talking about that very sound (which he refers to as ”the wash”) and how it was made with a Fairlight factory sample.

Oh really - do you have any info/detail on this? I'd love to see it...

Like I said, the missing part is another sample from the same 1979 disks.

In the wash part, you mean? Presumably, if the TRAMCHLO part doing the main chords was layered with another sample as you suggest above, that's the same layer that was used to produce the frozen wash?

I tracked it down back in 2019 and was going to write a crazy ambitious exposé on the RUTH samples, but then this thing called a full time job got in the way… I’m gonna try to find time to dust off the project in the coming weeks.

Great! I was going to write what I'd discovered during this thread way back in 2010 up in my upcoming blog series where I'm uncovering the sounds I loved in my formative records, but if you're planning something similar, it probably doesn't make sense to duplicate the effort - and I have a few other record secrets up my sleeve too..!

[Drew Stephenson]

Welcome back Desmond.

[Kwackman]

Welcome back Desmond.

Indeed.
Great to see.
You've been missed!

[Dan LB]

Welcome back Desmond.

Indeed.
Great to see.
You've been missed!

I’d like to echo the above sentiments! Welcome back!