Want Advice on Making Multi-Layer Recordings

[Piano Al]

I want to record some tunes by playing individual parts (e.g. bass, piano, vibes) on my Yamaha P90, and layering them together.

If you've done something like that can you give me some advice, or tell me of pitfalls to avoid.

To show you where I am, this is a recording I made this morning:

https://www.box.com/s/2v06ppukkrneygr62vpr

I did that by first recording a bassline in the P90's track 1, then the piano comping and piano RH in track 2, recorded at 100 BPM. I then played it back at 200 BPM and recorded it into the computer, sending the piano output into my laptop with Audacity recording.

Next, I overwrote track 2 on the P90 with the vibes, and recorded that into the computer.

Finally, I overwrite track 2 with the Hammond B3 sound, and combined that with the other tracks.

I'd like to avoid spending money on more hardware, but should I get a replacement for Audacity?

Thanks!

Al

[Jeff Smith]

What won't Audacity do for you? Why are you recording into your keyboard's sequencer first, instead of doing all tracks in Audacity from the start?

[iguessso]

Yes I would say record all (multitrack) in Audacity. You can have a metronome in Audacity by going to the generate menu "create click track".


Other than that you may want to look into Reaper as a DAW or even the free version of Presonus' Studio One...as they have a manageable learning curve and have much better user interfaces.

If you do that you may need to install ASIO4all which is a free soundcard driver that is more suitable than stock sound drivers for recording.

[alez]

There's plenty of top quality, free software for Linux. I would advise you to install some Linux distribution on your pc (dual boot possible), then you can use Ardour, which is a fantastic professional grade multitracker, or if you want your midi sequencing integrated in the same "digital audio workstation" (DAW), you can use Rosegarden instead. You could still use Audacity (since that's Linux native anyway) plus plenty other stuff such as virtual synths, music engravers (to produce music sheets) and so on. All of that you can do on Windows and Mac, but not for free.

If you'd rather use Windows, I recommend Reaper, which you can use for free for as long as you choose to, it's a really good DAW too, and my choice until I finally decided to go Linux all the way. It supports standard VST plugins, which means you can add all sorts of synths and effects.

Either way you would get stuff like a decent mixer from which you can do your levels, equalizers, reverb and so on in real time while you are actually listening to it.

[Piano Al]

Originally Posted by Jeff Smith

What won't Audacity do for you?

Nothing that I know of. I was just wondering whether things would work better with something else. I like Audacity fine, but I'll take a look at Reaper.

Originally Posted by Jeff Smith

Why are you recording into your keyboard's sequencer first, instead of doing all tracks in Audacity from the start?

1. I can have the metronome running, then when I have track I like, I just turn it off when recording into Audacity.

2. It's very quick to try a track, and reject it if I don't like it.

3. I can record at any tempo I want, then speed it up without any loss in sound quality (speeding up will be one of the features of these recordings).

Thanks for the help.

[alez]

I have read the method you use with greater attention. There's obviously nothing wrong with it, it works fine. But since your post is about how to do things in a better way, I'll tell you the method I'd be using, and the benefits of that compared to your one.

1. I'd interconnect the keyboard and pc to get them in sync. This has the benefit that you needn't manually sync your different tracks like I presume you must be doing now. If your sound card doesn't have a midi port, I'd get a cheap usb to midi thingy.

2. Now that I have those two talking midi, rather than use that comms for sync only (like you prefectly could), I'd use a pc midi sequencer to record my playing (at the slower tempo) instead of the keyboard sequencer. This has the benefit that you needn't do any midi to audio conversions to free up midi tracks like you are doing at present. You just record as many midi tracks as you please. Plus also you can see them all laid in front of you and correct (shift) any notes that you think are slightly too off, and whatever other editing. Your pc sequencer can play metronome clicks through your keyboard too.

3. Once I have all my recording done in the form of different midi tracks, I'd change the tempo to the desired one, I'd load a soft synth with a really good soundfont and make the sequencer send midi to there instead of sending to the keyboard as previously. I'd then do my mixing until I like the volume levels, reverb and / or chorus and so on. I'd then save the whole thing as standard midi file (smf), and use the soft synth to render it to wav. This has the benefit that your recording will be perfect in terms of sound quality since no digital to analogue to digital takes place (no hiss, hum or click noises, no distortion, no frequency range limits, etc). Therefore, it doesn't matter wether your soundcard recording quality is any good or maybe plain terrible, in fact it doesn't matter if you have a soundcard at all. Also it has the benefit that the process takes a few seconds only, since it's not done in real time.

3b. As an alternative to 3, if you really like the sounds on your keyboard and want to use those, once you've changed the tempo to the desired one, you can just do your mixing and have your sequencer play the tune on your keyboard while you record that into Audacity. Since it's one audio track only, you needn't sync the two applications.

[Piano Al]

Wow, that's good info, Alez.

What PC MIDI sequencer would you recommend?

And what softsynth and sound fonts?

Thanks!

[alez]

I'm neither much of a midi guy nor a Windows guy I'm afraid, but I really liked Reaper very much. It's a DAW (which means it aims much farther than midi sequencing only), but its built in midi sequencer is a good one. As a side benefit, you also get audio support in case your sister in law turns up to help out with some lyrics. Or if you want to do "3b" on my previos post, you needn't a second application anymore, so it's good flexibility, and like iguessso said, the learning curve is actually great for the power and flexibility you get.

Since that's got standard VST plugin support, again you needn't a separate piece of software as soft synth either. Reaper may actually come with what's called a "GM/GS VST soft synth" in the package already, then again maybe not, I'll have a look to see what's available but those keywords should be enough to get you a free one that does the job. Any that can use SF2 sound fonts should be alright, since sound quality depends much on the soundfont, little on the synth (they all do a fairly good job).

For good quality, free SF2 soundfonts, google for the Hammersound website, a great resource that's been running for many years. You can download "collection" files, i.e. all instruments in the same file, or individual instruments. Being a pianist, it's likely that you will want to download one of these massive piano sondfonts with great sound, plus a GM/GS collection for vibes, bass, funky organs or whatever else.

In the end, if you like your keyboard synth sounds better, you can do "3b". Or maybe you can buy one of these super good quality piano soundfonts and rely on a free soundfont for the rest of instruments. Using your keyboard synth for those other instruments would be very tricky due to the latency (delay from command to sound out) inherent to soft synths. You can do audio + soft synth no problem, or soft synth on its own, but hardware synth + soft synth is a no-no generally speaking, also soft synth + soft synth is often tricky due the different latencies, which is a shame because it often makes sense to use a GM/GS synth to emulate real world instruments plus some other(s) for '80s fat analogue synth emulators or whatever other spezialized sounds.

[alez]

A Google search for "free gm/gs vst synth" gives plenty of links for research.

BTW, I didn't mean to imply that Audacity is poor, in fact it's my wave editor of choice, I'm a fan of it and use it all the time for many different tasks such as transcribing and transposing. It's just that there are tools much better suited for the task of multitrack recording than a wave editor.

[shiatoru]

Originally Posted by Piano Al

3. I can record at any tempo I want, then speed it up without any loss in sound quality (speeding up will be one of the features of these recordings).

Good time-feel is tempo dependent, meaning that if you play a track with great time-feel at 120 bpm, speeding up the track to 220 bmp (or any appreciable amount) is going to result in not so great feel. If you merely speed up a track, the time-feel of your playing is going to be sped up linearly, which isn't how it *should* happen, if you were actually playing at the sped-up tempo. You have to be feeling the particular tempo that the track is going to end up in. Of course, I'm not talking about a minor adjustment, but even that....

Hope I'm making some sense here....it's late...

[Piano Al]

Originally Posted by shiatoru

Good time-feel is tempo dependent, meaning that if you play a track with great time-feel at 120 bpm, speeding up the track to 220 bmp (or any appreciable amount) is going to result in not so great feel. If you merely speed up a track, the time-feel of your playing is going to be sped up linearly, which isn't how it *should* happen, if you were actually playing at the sped-up tempo. You have to be feeling the particular tempo that the track is going to end up in. Of course, I'm not talking about a minor adjustment, but even that....

Hope I'm making some sense here....it's late...

I have to say that I've found that the exact opposite is true. Even if the time feel is mediocre at a low tempo, when I speed it up, it sounds great.

What do you think of the time feel in these?

https://www.box.com/s/gmqxqknlrhl4ia8vbpsr

http://www.youtube.com/watch?v=gtWBGlXTjs8

[Kevin Murray]

Originally Posted by shiatoru

Good time-feel is tempo dependent, meaning that if you play a track with great time-feel at 120 bpm, speeding up the track to 220 bmp (or any appreciable amount) is going to result in not so great feel. If you merely speed up a track, the time-feel of your playing is going to be sped up linearly, which isn't how it *should* happen, if you were actually playing at the sped-up tempo. You have to be feeling the particular tempo that the track is going to end up in. Of course, I'm not talking about a minor adjustment, but even that....

Hope I'm making some sense here....it's late...

That is an amazing point. You never play the same at breakneck speed as you do at a nice, relaxed swing. One thing I have noticed when changing tempo live (not via computer) is that the faster you play, the calmer you need to be. If you look at videos of the greats playing, they'll get really physically into it on medium/slow tunes, but on the up tunes, they'll lay back a little bit on "moving around" and physically "grooving." This is not, in any way, saying that they play less intense at uptempo. Just an observation.

And personally, I'm against everything multi-track. We're talking about jazz here, for god's sake! Its all based on originality, spontaneity, and the groove that occurs in the moment, not in the moment that a guy drags some tracks together, edits some waveforms, and hits "Play."

Now, overdubbing can be helpful for the solo performer with no band, or for the guy trying out a new piece to see what it will sound like with a band, but I think its only good for getting a general idea of things, not the actual groove... I don't play the same when listening to a click track as when I'm listening to my bandmates.....

But, great point above, and great thread.