Skip to main content

I produce lots of copies of songs mainly for theatrical or stage type shows. Over the years I have a pretty standard workflow, but while doing one today, I wondered if others do it my way or have their own system that might be better?

I always use the original song as a template. I tend to do the trickier bits in short chunks. So playing along with the real thing is pretty critical.

So here is how I tend to do mine - Ideas like to hear other people's differences.

Bring the original track in and create a tempo map. I never found Cubase's automatic system accurate enough - especially if I want to fix my pretty ropey piano playing, and especially my drumming!
The way I do it is to work out roughly the BPM and start with that. I use the visible waveform which usually has peaks on the beats. Then with the metronome on, I'll play the track and move the tempo track bmp up or down every time the peaks drift away from the on screen bar/beat markers. Takes ages of course, but eventually I get a decent tempo track, with changes maybe every couple of bars for some, but sometimes twice in every bar. I find it amazing that the BPMs speed up and slow down quite radically in many popular songs - up to the day they invented sequencers!

Once the tempo map is done - for my type of tracks its usually then keys first, then bass, then the other stuff.

My weak areas are drums and lead guitars - so I tend to do these in two bar chunks - until I get it right. Drums - I tend to do by kick and snare, then go back and add the hats, then again for cymbals.

Do any of you do this major differently?

Comments

audiokid Mon, 03/23/2020 - 11:08

Sorry for the confusion, I’m a bit confused with what you are “ultimately” trying to achieve in the long haul but I totally get what you are doing with Cubase and timing issues on older music that wasn’t ever tracked to click tracks.

If you were actually playing the drum parts in real time, an MPC is a much better drum controller to any midi keyboard on the market. But this doesn’t sound like it’s what you are doing or ever plan on.
However, if you ever are, then a great drum controller like an MPC is fast and very musical, the MPC hardware might be just what you have been missing.

If all you require is triggering off of prerecorded music, then your method or Kyles suggestions, (remixing tricks) work excellent.
I’d be using Transient detection software along with Spectral editing to strip out everything needed to get the audio to midi done ITB.

MPC are not your average drum box or like pads we see on some keyboards. They augment standard DAW and are in my opinion an essential midi hub for large midi production studios, live or in the studio.

Most drums in popular music are simple to play and in the centre where bass and vocals live so I simply avoid all the fuss of transient detection and just play drums which I usually do extremely well when I use an MPC.
If I am mixing music for people which usually has badly recorded drums screwing up the entire mix, I would use an MPC or transient detection and replace the bad drums (undetected)

To my experience, most popular music has drum tracks that are easier to play live than it’s worth trying to do what you are doing in Cubase or any DAW for that matter.

paulears Mon, 03/23/2020 - 11:34

NO - no triggering whatsoever. I think it's got a bit misunderstood.

We do show tracks - so they ask for New York, New York, and they send me the arrangement they want, usually the original track. I listen to this purely as a guide. NONE of it ends up in the new recording, so its a case of create a tempo track, so the tempo of the song allows the click to be accurate to play to. Then instrument by instrument we re-record it. When we're done the original track can be muted in or out and its exactly the same as the new one - then we mute it and it's never used again.

Obviously, without the click, we could simply just press record and play along, the recorded piece having no sync with the sequencers timing at all - but that means you cannot chop up the bass part and duplicate it. and drop it back in. You can edit it of course, but every note is in a random position and you have to keep the quantize grid function off, because the bar lines and beats on the edit screen are totally in the wrong place. If the drummer pushes a note in the recording, it would be nice to just move it a tick or two in front of the bar line, but with no time lock, it's a bit random to edit like that.

The idea of triggering came from the idea that if you removed everything bar say the kick and snare with eq, compression, etc, you'd have a drum track that Cubase could use to more accurately track the tempo, as other instruments like bass can confuse the timing engine. Good idea for some music maybe but I tried it and it failed.

So all I really hoped for was that somebody knew of a hidden function that would help me speed up, or perhaps suggest a different way of doing what we do. Creating a manual tempo track is often quicker than editing the mistakes in the automatically extracted one, despite it having quite a few parameters to adjust.
What I'm doing in Cubase IS playing everything, track by track. That bit is our everyday activity - we don't use any stuff the computers generates, just what we play in as VSTi generated sounds or input from something else. For instance, I have a MIDI guitar, but using it takes longer than either playing the part on the keyboard, or recording real guitar. This is my least favourite option because it's less versatile for fixing mistakes.

I don't want to replace the bad drums, because if I play badly, I just edit. Best example I can give of things that cause me trouble are those songs where the music stops and the drummer does a big tom tom bum, bum, bum, bu-bum, then the song resumes louder and better. I find the timing of these bits quite tricky - but if you can see the waveform of the original you can then align the MIDI drum pattern to make it work.

I can't believe I'm the only person doing tracks? We've been doing it since around 94 I think. Sometimes we get asked for new arrangements but that's perhaps 30% of the requests.

https://www.limelight.org.uk/mbs.wav

Here's an example of the kind of thing we end up with.

audiokid Mon, 03/23/2020 - 11:52

After 20 years professional as a performer, I started RO and got into recording. Ironically I’ve never found many recordist open to remixing, programming and mixing tricks.
I used to talk about the things I’m sharing here, that I had been doing professionally but anything midi simply isn’t of interest to most recordist. Digital audio is really a bone of contention to those true to the art of recording. This is changing and it’s fun learning more about what you are doing, Paul.

Thanks for taking the time to explain more and to share the link. I’ll check it out later tonight and if I have anything more to share, I’ll chime in.

paulears Mon, 03/23/2020 - 12:18

I know what you mean - Mics are still important for us - our other field is of all things opera and live stuff! Choirs, chamber quartets, and of course 60s pop - for the band.

EDIT
I've just been watching the youtube videos on MCP X and I know that this is a product I'd have absolutely no use for at all. It's got a huge fan base but I just don't want to go down a sample based loop type system as a way of working. Just not how I work at all.
Neat product and well thought through, but totally alien to my way of production. In fact - it's a hinderance to my workflow, not an improvement. At least I now understand what it does -so always good to get a bit of extra knowledge.

audiokid Mon, 03/23/2020 - 12:47

paulears, post: 463748, member: 47782 wrote: I can't believe I'm the only person doing tracks? We've been doing it since around 94 I think. Sometimes we get asked for new arrangements but that's perhaps 30% of the requests.

https://www.limelight.org.uk/mbs.wav

Here's an example of the kind of thing we end up with.

Just listened. Nice
This is exactly what I did professionally everyday for 2 decades, Paul. I programmed everything, played guitars. You're doing a great job. However, the MPC X is what I would be using. The drums coming out of an MPC sound more live. Not sure why that is but I think its to do with how we can play feel into it apposed to a keyboard. Drums played on keyboards always sound like drum machines to me. Drums played on MPC can fool the best engineers. I've done that many times.
But, I don't think there is any one easy way to drag or speed segments that don't follow set tempos. I usually just turn off quantizing and fly at it until it sounds good enough.

kmetal Mon, 03/23/2020 - 18:12

paulears, post: 463748, member: 47782 wrote: If the drummer pushes a note in the recording, it would be nice to just move it a tick or two in front of the bar line, but with no time lock, it's a bit random to edit like that.

I once edited virtually every note, on every track of a 30 min death metal ep that wasn't done to a click, its amazing how fast you can get at edits without a grid! Lol never again tho.

paulears, post: 463748, member: 47782 wrote: The idea of triggering came from the idea that if you removed everything bar say the kick and snare with eq, compression, etc, you'd have a drum track that Cubase could use to more accurately track the tempo, as other instruments like bass can confuse the timing engine. Good idea for some music maybe but I tried it and it failed.

I think maybe your results were dissatisfying because you weren't using the most optimal tools for the job.

I did a quick test with RX7 and nearly completely stripped out vocals, keys, bass, and guitar, on one of my bands blues rock tracks, leaving just the drums. And this was my first attempt using RX7. The free m/s voxengo pluggin did a pretty good job muting the sides, with 1 button press.

When i get my system setup ill do a demonstration, but so far i think it works as well as i suspected.

But again different strokes for different folks. I just want to clarify between a technique not being effective, vs just not being a prefferred method.

paulears, post: 463748, member: 47782 wrote: When we're done the original track can be muted in or out and its exactly the same as the new one - then we mute it and it's never used again.

This is where im confused. If the point is to make the track as true to the original as possible, wouldn't re-amping/sampling the actual track be the way to go? Why exctract the exact tempo, match the sounds, then perform the parts? As midi your really just altering the velocities, since the timing is the other main parameter.

Im unclear how lifting the track vs matching it identically via editing and grids is any different?

Im curious not trying to criticize.

audiokid, post: 463750, member: 1 wrote: Ironically I’ve never found many recordist open to remixing, programming and mixing tricks.

About 10 years ago i made a concerted effort to listen to electronic music to expand my musical and engineering. There is a remarkable amount of technique involved in some artists workflows, and the approach to mixing and arrangement is an interesting lane compared to 'rock band in a room' methods. Both can be very cool, and very deep.

MPC seems like a cool world. I just installed the mpc studio software the other day, its funny how recognizable the sounds are. It was just like using a neve/apogee the first time, like aha, that's the sound ive been hearing on records my whole life, without trying!

Mpc and Reason take some interesting approaches to creation i think will be fun to explore.

paulears Tue, 03/24/2020 - 00:03

I wonder if it’s just a history thing. To me the idea of stripping out the drums then using this to replace the original drums with high quality new samples runs against the grain. However I’ll admit that this would be more accurate once you’ve cleaned things up. I’m not quite sure why I’d not wish to do this apart from knowing this isn’t a process I want to do. The drums do take me disproportionate time so it could be a time saving idea, which is the point of this topic. I don’t think it’s an ethical thing, as in stealing the real drummers performance and using it, but it does seem more complicated and time consuming? Is the time to extract the drums then edit out the inevitable mid-hits then trigger the samples longer than my current process or shorter? As I don’t do it I don’t know?

paulears Tue, 03/24/2020 - 08:52

I think we have to agree to disagree. There's something musical when you play something - using samples triggered by a technical process is great music technology and doable - but many of the clips I watched on youtube were music created by random button prodding, and then the machine doing the work - which it does admirably clearly saying to itself "he really didn't mean that, I'll fix it - he won't notice". A great product I think for DJs, and the technologist dance music creators, who wouldn't (and needn't) know a minim from a crochet . Just a totally different process to mine. I'd like to think that even when pressing a button produces a great sample based 2 bar sequence, the musical skill is in picking the next starting note for the next sequence. I do understand the product - and agree how clever it is, but it's a product for people who don't need to be musicians. Clearly many decent musicians use them because it speeds that kind off workflow, but I've been sequencing in a linear manner since 94, and I don't want to change that.

audiokid Tue, 03/24/2020 - 09:09

(Edited to explain better)

I agree.

But then I guess we are all cheating if we are using a keyboard to reproduce an orchestra, brass etc.
We are all hypocrites.

I use an MPC and actually play drums on pads when its for the right reasons. Other times, I use technologies that yield whats whats best for the song, the goal or clients best interests.

audiokid Tue, 03/24/2020 - 09:14

Regarding stripping and replacing and doing everything we do to recreate music.
What are the differences between writing and performing?

I’d say do what achieves the best results. The new world does not care who played what on a remix, reproduction of some muzak. I'd rather make a living in this business sounding awesome apposed to sounding old and dated. I say... use the tools you need to produce the best sounding music you can. Digital technology is amazing and a wonderful part of our music world.

Tony Carpenter Tue, 03/24/2020 - 10:03

I upgraded to a Native instruments Maschine Studio some months ago. I didn’t like the pick a bar amounts then tap it in approach. They just changed maschine vet 2.10 to allow free form input. Along with my owning the full collector edition of the latest top of the line NI sample instruments and other libraries I’m excited to have a shot again at it. I also play drums keyboard and guitar. That being said some things MPC and Maschine etc do is very fast, and more natural now.

kmetal Tue, 03/24/2020 - 15:41

paulears, post: 463762, member: 47782 wrote: but many of the clips I watched on youtube were music created by random button prodding, and then the machine doing the work - which it does admirably clearly saying to itself "he really didn't mean that, I'll fix it - he won't notice".

Isn't this pretty much the same as when you extracted the tempo, and snapped the midi to the grid? If im understanding correctly, the problem you experience with auto mapping is its snapping the piano to the wrong parts. This wouldn't be an issue if quantization wasn't being used.

Ive always hated tempo mapping even just stopping the band to insert a tempo change. Its the worst part of what i had to do as a staff engineer.

I think this conversation is just showing differences in preferences in manual vs automated, when essentially performing the same task.

paulears Tue, 03/24/2020 - 15:46

I've got Maschine and perhaps should try it - one of the Kontakt things that you have but never fire up.

I thought about trying out some of the drum stuff - I figured I would also try more of the guitar processors, effects and gizmos too, and set myself the task of writing a song outside my comfort zone, with drums and for good measure, a key change in the middle. All inside Cubase. Two hours end to end to this stage - which doesn't;t have any fader movement yet - so still a work in progress although I doubt I'll ever finish it off.https://www.limelight.org.uk/City-Streets.mp3

paulears Tue, 03/24/2020 - 15:57

kmetal, post: 463766, member: 37533 wrote: Isn't this pretty much the same as when you extracted the tempo, and snapped the midi to the grid? If im understanding correctly, the problem you experience with auto mapping is its snapping the piano to the wrong parts. This wouldn't be an issue if quantization wasn't being used.

No - I've not explained properly. The tempo map allows the ebb and flow of the musicians playing free and not to a metronome to be edited. So all I'm doing is making sure the bar lines align with the music - so each bar has it's own BPM setting, or sometime the bar might have two or three changes within it. To be able to copy and past, accurately, you need to be able to make each bar group - 2,4, 8 or whatever be able to paste somewhere else, and have it play at the variable tempo. I'm not quantising anything, the notes are all as they would be played in free time and often a bit loose - but the timing is cured by the tempo map. I'm NOT making every single note lay on a bar or beat - but just locking the tempo to the original. I don't think I'm comparing manual or automated that's not what I'm doing. Music from the 80s onwards tends to be played to a rigid click, so every chorus is the same length and the same tempo - very precisely.

kmetal Tue, 03/24/2020 - 18:04

paulears, post: 463768, member: 47782 wrote: No - I've not explained properly. The tempo map allows the ebb and flow of the musicians playing free and not to a metronome to be edited. So all I'm doing is making sure the bar lines align with the music - so each bar has it's own BPM setting, or sometime the bar might have two or three changes within it. To be able to copy and past, accurately, you need to be able to make each bar group - 2,4, 8 or whatever be able to paste somewhere else, and have it play at the variable tempo. I'm not quantising anything, the notes are all as they would be played in free time and often a bit loose - but the timing is cured by the tempo map. I'm NOT making every single note lay on a bar or beat - but just locking the tempo to the original. I don't think I'm comparing manual or automated that's not what I'm doing. Music from the 80s onwards tends to be played to a rigid click, so every chorus is the same length and the same tempo - very precisely.

Oh ok i see what your doing know.

audiokid Thu, 05/07/2020 - 10:44

paulears
Hey Paul, I just finished watching Samplitude's new features, specifically video on tempo for you. Maybe this is helpful and possibly something Cubase may do or will improve upon?

New! Tempo Track Tempo now has its own track. Use automations to create elegant transitions between slow and fast passages. Set additional buttons as helpful markers for beat and tempo change.

https://www.magix.com/int/music/samplitude/pro-x/new-features/

paulears Thu, 05/07/2020 - 10:52

That seems to be very similar to how Cubase works - I've done a few new little jobs and I've resigned myself to just carrying on as I've done for a few years now - manually pushing and pulling by ear seems the only solution to following the music. I'm doing one now - music from showboat and have a choir who sung to one track, but can't get the copyright cleared properly, so I'm trying to fit their singing to a new track, which is the same, but varies greatly in tempo. Not really fun.

x

User login