DAW Workflow Questions

[Tony Monaghan]

Thinking about the way I work with my studio setup and how I may not be getting the best results.

Oldish PC with Sonar 8.0 and myriad VSTi's and effects;

I create a MIDI file with say 8 tracks
Tidy up the mix so it's how I want it
Add effects
Export the whole mix as a two track stereo .wav file

Looking back at other posts it looks like some people create an audio track for each midi track and apply any effects / mixing then master that to a final two track .wav

My PC just can't handle running multiple audio tracks with or without effects but I may have a solution. I have a Roland VS2400CD (http://www.musiciansfriend.com/document?doc_id=91546) and thought I might press this into service. It's normally used for live recording, voice and instruments.

My though was to sync this to Sonar and record each audio track one at a time with effects added. Once I have all the audio tracks recorded on the external DAW I could then mix to my hearts content and master on this or send a two track master back to the PC for any final tweaking.

How does that sound? How would you do it?

[MMOSC]

Thinking about the way I work with my studio setup and how I may not be getting the best results.

Oldish PC with Sonar 8.0 and myriad VSTi's and effects;

I create a MIDI file with say 8 tracks
Tidy up the mix so it's how I want it
Add effects
Export the whole mix as a two track stereo .wav file


Tony,

That's pretty much what i do. Large ensembles (in my case over 16 tracks) pretty much require creating separate audio tracks from the midi, I use Sonars freeze function for this.



My PC just can't handle running multiple audio tracks with or without effects


I'm not sure I understand, rendered audio playback requires far less power then the audio rendering of the midi with samples.

[karvasika]

I write the score in Overture and program all the midi there, including mod wheel and humanizing. I then export each track to Ableton Live (but the sequencer doesn't matter, really), possibly use up to 8 instruments (if building a string ensemble) on a single part and then add Altiverb (early reflections) to each instrument group with stage positioning. I then add the long tail to master and do any possible equing there. I also possibly tweak each individual channel with EQ or compression if needed.

This is my ideal setup, sometimes I also take shortcuts by adding just a stereo reverb to section strings, skipping the ensemble building.

[LFO]

Hey Tony,
I'm going to parrot what Mike said. MIDI tracks (and the associated sample playback) requires much more processing power than audio tracks. You should be able to record audio tracks for each MIDI track (muting MIDI tracks as you progress) and then add effects. Maybe you are running out of juice when you add effects, though it doesn't sound like you have taken it to this step.

A few questions:

1) What version of Windows are you using?

2) What make/model PC do you have? (Number of cores, speed of cores)

3) How much memory?

4) How many tracks do you typically have in a project?

The main problem I see with your workflow using the VS2400 is that you are using a VS2400. :D Sorry, those all in one units don't fit the way my mind works at all! Seriously, the main issue is that the reverb available on those kinds of units are usually very pop music oriented. Hence, not very good for orchestral work. What about selling the VS2400 and using the money to upgrade your PC?

These are pretty random thoughts because we don't know much about your system. The quality of response will improve with the quality of the data provided. :)

-Kevin

[Tony Monaghan]

Thanks for your thoughts guys.

PC is a PIV 3.2Ghz with 2Gb RAM XP Pro.

I typically work with 8-12 tracks sometimes more. I've tried recording audio tracks as LFO suggested but Sonar chokes on more than a couple of tracks without any effects - that's muting the midi tracks as well.

I wasn't planning on using the onboard effects on the VS2400 but to add conv reverb to one track at a time sending it out the external DAW. That way I can have as many tracks as I like recorded with effects on the DAW ready for mastering.

Will take considerable more time doing it this way but I'm hoping to get better results.

[LFO]

Hey Tony,
With a 3.2 Ghz box you should not have any problems. Am I missing something? You say it is an oldish box but the CPU frequency is not what I would call an older rating. :) 2 Gig of RAM should probably be bumped up to 4 and then your world would totally change. What motherboard are you using? Is PIV a brand name?

One thing I would do is take a look at http://www.musicxp.net and take their advice on optimizing XP for audio. I used their recommendations for three or four years when I was on a Windows box and my Windows box was rock solid and efficient.

Last of all, and very important, are your hard drives. Are they 7200 RPM or faster?

I could do much, much more on a 2.4 Ghz XP Home box with 4 gig of RAM, so you should be able to get your Windows box humming along and you can forget (or better yet sell!) the Roland. ~|

-Kevin

[P.T.]

When you play back the audio tracks try raising the latency up to at least 40ms and maybe higher.

It won't affect the mixing and should give you more tracks.

I used to run more than a couple of tracks without effects on a 500mhz PIII laptop.

Though The Sonar demo ran very poorly on that computer. Very poorly.
I got 4-6 or so (I haven't used the laptop in about a year so I don't remember exactly) audio tracks on that computer using Tracktion 1.

[rbowser-]

Hummm, I can only echo perplexity already expressed on the thread. Audio tracks take much less CPU to play than MIDI tracks linked to soft synths. If you have 12 or more MIDI tracks playing - you should be able to have triple, quadruple that in audio tracks playing back with no problem. - I don't get it.

I work up to maybe 6 tracks then bounce to audio, mute and archive the MIDI tracks, then move on - ending up with 30, 40, 50 audio tracks. My old computer used for music, XP off-the-shelf Dell has a limit to live playback of MIDI, but never chokes on audio.

You're bouncing the tracks, right?--despite some people's misgivings about bouncing - you're not recording these in real time?

You can see I feel that it's really necessary to bounce tracks to audio before going to a 2-track master. There's so much more to be done after you've worked up the MIDI files.

Randy

[efiebke]

Tony -

Something to entertain for the future when there is some extra cash to spend: the use of slave computers to help distribute the CPU and memory workload. Even one "slave computer" could be helpful.

For my happy little studio, I have up to 3 "slave computers" networked together using midi-networking programs. To be honest, I rarely use all three. I usually use just one. But they're there if/when those big sequencing projects come along. (And, sadly, I haven't the time lately for writing and sequencing those big compositional projects.) It's a stable set-up and I can use dozens and dozens of software instruments from various libraries (GPO and JABB being the two main ones! ;) )

What was helpful in "creating" this set up was. . . . not having kids! LOL! :D I also didn't throw away my old Dell computers. So, if you ever decide to purchase another computer, keep the old one!!

Just tossing this idea out there to you (or to anyone interested). I'm sure you know lots of composers/musicians with similar set ups. I'm just a vocal advocate for the use of "slave" computers as one solution for managing those big sequencing projects.

Good luck! :)

Ted

[Michael_uk]

Hello Tony,

I agree with those who felt you should have no problems with your PC specs.

Kevin, PIV = Pentium 4 (processor).

I build my own DAWs and have not long rebuilt and updated mine. My previous DAW was also a P4, 3.2 processor with Sonar Producer and I had absolutely no problems at all with this. My scores often went to 16 or more tracks both MIDI and audio. The difference is that I had 4 MB RAM and you have 2 MB. I agree that adding another 2 MB would certainly be a good move although I'm not so sure it would make that much of a difference in your situation. I would certainly add the extra and try it; the cost of RAM is low at the moment and your computer could do with it. If the additional RAM doesn't spruce things up for you, as I suspect, then, without going into fuller details, your mainboard may be the weak link in the chain.

Regarding work-flow, I work with MIDI for a lot of the time bringing this to audio for final mastering. All of this is done within my DAW and, as I mentioned at the top, I had absolutely no problems with PC performance at all. Regarding bringing out tracks to an external source then taking them back into Sonar again. Personally I feel that this is not the best work-flow. I hasten to add that this is just my opinion and I realise people do work this way but it wouldn't do for me.

I hope you try adding that RAM and that you find the optimum work-flow that works best for you.

Good luck. :)

[rbowser-]

Hello again, Tony

I still feel there must be something incorrect in the way you're not having success with getting individual audio files for your projects. As everyone confirms, your computer should be able to handle it. What you describe is the opposite of the way things actually work - the number of live MIDI tracks you can drive should be much more limited than how many audio tracks you can drive.

I should add, dovetailing from Ted's description of how he has 3 computers hooked up together - I've used a Seagate Terrabyte hard drive for a year now, and that's where my actual Sonar projects are. The audio tracks along with the Cakewalk files are run from the Seagate while my computer runs Sonar. I've expanded my capabilities with a number of external hard drives now.

Randy

[robh]

Tony,

One thing you haven't mentioned is whether you remove the VSTi once you've bounced to audio. I'm not sure how Sonar works, but only muting the track still means the VSTi and associated samples are still in memory - which the audio tracks still need (although not as much). Having said that, it does sound like your latency is too low (for mixing purposes), but you also didn't really describe what exactly happens when the PC "can't handle running multiple audio tracks."

Rob

[rbowser-]

"...you haven't mentioned whether you remove the VSTi once you've bounced to audio..."

Good call, Rob. In Sonar, in the Synth Bin, you click the power button to the left of each synth you've used, and it disconnects. Definitely need to do that. After bouncing to tracks, besides disconnecting the synth I also both mute and archive the MIDI tracks as well as the audio tracks the synth had been using.

Randy

[buckshead]

Tony, Don't know whether this applies to Sonar but in Cubase muting a track just makes it silent and doesn't stop it working. There is a separate "MUTE" that actually disables part of the track as distinct from making it quiet. This saves processing power.

[Tony Monaghan]

Rob, good point and probably where I'm going wrong!

I'm looking again at the setup and likely latency is too low as well. What actually happens when running multiple audio tracks is the processor maxes out and playback stops.

The way I've been working I never bounced any tracks at all, once I was happy with everything I'd just export to .wav. I'm definitley not getting the best out of the setup so back to basics again.

The VS2400 is still useful for live recording, I used to have a Delta 44 which I used to get audio in to the PC but abandoned that some time ago. No plans to sell the Roland, cost me £2,200 and I wouldn't get a fraction of of what it cost to sell it now. Still a great machine for recording live gigs, vocals and instruments.

Thanks all for your input! *()

[rbowser-]

Good luck with it all, Tony - To me it's still totally mysterious why your computer would choke on audio tracks like that. When you're still working in the MIDI realm, there's a big drain of power with the soft synths going Plus all the audio tracks for pumping their sound out. I'll start getting into the 90% realm of CPU usage when I realize I need to bounce tracks. After the bounce, CPU goes down to 2% to play back the audio. --That's the normal case.

In one of your posts you said that your machine would crash during the actual bounce - which sounds like somehow your bounce/record procedure is faulty. Latency can be changed if needs be, as noted on this thread, but I've never had to change it. To me, it's a weird problem you're having.

Randy

[buckshead]

To get the MIDI in I use an exterrnal Tone Generator, mines by YAMAHA. Its of course only 16 tracks but when running pure MIDI it doesn't use any processing power to speak of. Then you can use a library like GPO to play each track and render as you go.

[rolifer]

Tony

My old computer was a dual core w/3.0 AMD chips. I would have around 50 midi tracks going with very few effectsw added. That was just about the limit of that computer.

I now have a computer with the i7 which has 8 cores at 2.66 each. I now have 50+ midi tracks, a ton of effects going on in the outs and a few audio tracks going as well. The new computer is barely thinking about what it is doing.

It is 64 bit of course and that has created some problems, but all of them have been workable so far.

Just something to think about.

Ron

[Tony Monaghan]

Well, got it working sweetly at last *()

12 tracks of audio running with several conv's instances and no problems. This is a different way of working from what I'm used to but the difference already is very noticeable.

As usual excellent advice from everyone which sure helped me to figure this out, thanks again )(~

[rbowser-]

Excellent news, Tony - What did you set up differently to get the audio tracks working right?

Randy

[Tony Monaghan]

Excellent news, Tony - What did you set up differently to get the audio tracks working right?

Randy

Randy, I didn't have to change much at all. After bouncing tracks I switched off the myriad vst's and effects and of course processor usage dropped dramatically.

A new lease of life for the old machine )(~

[rbowser-]

Great, Tony - I was never clear from this thread what it was exactly that was going to work it out for you. I had the impression you weren't being able to bounce at all. But the feedback here helped you out, and that's super. Hushing those tracks and processes once you've moved into the audio realm is the key for all of us users on "under powered" machines.

Now, go make some great music.

Randy

[LFO]

Randy, I didn't have to change much at all. After bouncing tracks I switched off the myriad vst's and effects and of course processor usage dropped dramatically.

A new lease of life for the old machine )(~

Hey Tony,
Congrats! If you want to get an even more efficient machine take a look at musicXP.net . It will make a difference.

-Kevin

[AlanPerkins]

Can someone explain for me (I am using Sonar by the way) Is it possible to bounce multiple midi tracks to multiple audio tracks in one pass or do you need to do it individually - one track at a time and mute or solo to achieve this as suggested by one poster?

[robh]

Randy, I didn't have to change much at all. After bouncing tracks I switched off the myriad vst's and effects and of course processor usage dropped dramatically.

A new lease of life for the old machine )(~
I won!:D

Rob

[LFO]

Hello Alan,
I don't know Sonar, but I think the basic principles would apply to Sonar as they do for other DAWS. I record multiple tracks at the same time by doing the following:

I set the output of my MIDI tracks to Busses instead of outputs. (Bus 1 for track 1, bus 2 for track 2, etc.)

I set the input for the audio tracks to the busses. (Bus 1 is the input for track 20, Bus 2 is the input for track 21, etc.)

I record enable the audio tracks, hit record and the audio from the MIDI tracks is sent to the audio tracks and recorded.

I put together a quick video to demonstrate what I described. It's in Logic, but it should be very similar in Sonar. I just did one track, but you just do the same for multiple tracks and you are good to go.

http://www.box.net/shared/static/f9ejzbcvv5.mov

-Kevin

[Tony Monaghan]

Hey Tony,
Congrats! If you want to get an even more efficient machine take a look at musicXP.net . It will make a difference.

-Kevin

I subscribed to musicxp.net - didn't learn anything I haven't already done ~|

Still some useful info though, good call )(~

[Tony Monaghan]

I won!:D

Rob

You did Rob and a very good piece of advice. Been using Cakewalk / Sonar since early days and never quite got away from using it as I started with it before audio was a real consideration.

I thought I knew sequencing and Sonar pretty well but it's all down to how it's used.

[rbowser-]

Can someone explain for me (I am using Sonar by the way) Is it possible to bounce multiple midi tracks to multiple audio tracks in one pass or do you need to do it individually - one track at a time and mute or solo to achieve this as suggested by one poster?

Yes, Alan, there is, and it's very straight forward. Also, the only way to fly. You don't need to go through the convoluted re-routing of tracks through buses or any of that.

Highlight the MIDI tracks and their corresponding audio tracks. In the bounce menu, on the right, change the default setting of "entire mix" to "tracks." In the larger window at the left, a list will appear of what tracks you'll be getting. Un-check anything you don't need, like bus automation, track automation if there isn't any yet and etc. Push the button.

When the files are cooked, you'll have new individual audio tracks, one for each instrument/MIDI track. They'll even have the name of the instrument at the end of the generated label, if you previously had named the original tracks. But I always erase that and type in the short name, "Piano" "Strings 1" etc.

Ta Da.

Randy

[AlanPerkins]

Thanks very much Kevin and Randy.

This provides one audio track per soft synth, since I have multiple midi inputs going to soft synth tracks.

I am about to install SOnar 8 where I understand things are a little different. I will see what I make of that.

I appreciate your input.

Alan

[rbowser-]

Hi, Alan

"...This provides one audio track per soft synth, since I have multiple midi inputs going to soft synth tracks..."

Well, no - it provides as many audio tracks as a synth is capable of running, and how many you need. So with KP2, you can yield 16 tracks this way.

When you first insert the soft synth into a project, just make sure you ask for "all audio outs" instead of the single stereo track which is default.

Randy