You been holdin’ out on me?

[solarbird]

j0, all you engineers and producers – you been holdin’ out on me. How come never any of you felt it appropriate to explain to me exactly how goddamn useful automation is? Huh? ‘Cause damn.

And Ardour’s automation tools are both effective and easy to use! It’s freakishly unlike Ardour to be this way, so I’m appreciative. Even if it did assplode my machine today for a while, about which I will now rant.

Because srsly, it was like some sort of goddamn movieOS crash. Windows popping up everywhere scrolling errors at machine rate? Persisting across reboots and hardware resets? That was neat. I figured either something very obscure but trivial had happened and I’d lose a few hours to figuring it out, or I’d just lost weeks of work, ’cause I couldn’t even get Ardour stable enough to export individual tracks – no audio sources or sinks would stay online. I mean nothing worked.

So I started googling and eventually I found a French website talking avout something similar in another JackAudio-enabled app environment and ended up reading up on PulseAudio, which Ubuntu Studio won’t let you not install and which plays very badly with JackAudio. So I spent a few hours in the guts of the install making it not launch PulseAudio. (You can’t just uninstall it. APT also tries to uninstall the desktop manager as a dependancy. NOT FUNNY!)

And that’s something I’d been wanting to do eventually, because it’s been problematic for me and causing performance problems, but I had it reasonably working some months ago – they’re supposed to work together – and left it alone after that while I tried to get actual work done. But I guess the Time had Come, so I read up on how to force Ubuntu Studio not to launch pulseaudio (nontrivial) and hid all its utilities (srsly it launches this shit like five different places) and reconfigured Jack and, as I put on moo:

You say, “realtime kernel active… pulseaudio confirmed nonstarting… JackAudio running…”
You say, “outputs ACTIVATED”
You say, “automation engaging… engaged!”
You say, “running… FUCK.”
You say, “wait I think I see it”
You say, “fuck. restarting”
You say, “godDAMMIT that was behaving SO much better.”
Anna aw
You say, “it _is_ behaving better tho – enough to show me something…”
You say, “I think I have a corrupt automation event… <deletes> <restarts>”
You say, “…and outputs up… ”
You say, “…dsp at 29% and stable…”
You say, “about to start automation…”
You say, “automation running…”
You say, “so far so good…”
You say, “Jack and ALSA reporting no errors yet…”
You say, “approaching fail point… still running!”
You say, “past crash point…”
You say, “end of song!”
Solarbird collapses. “Automation shutdown orderly, transport stop normal, JACK and ALSA report all is well. It’s fixed. That was scary and unpleasant! But at least repairable.”

An hour and a half actually working on stuff, three-odd hours of fixing the world, really, that’s about what I expect. I go back and forth on all this – I mean, in the middle of All That, if somebody had squawked “OPEN SOURCE IS THE FUTURE!” at me I’d have punched them right in the face. But once it’s working again, well, at least I have tools.

Even if they explode sometimes at random, like people on LOST.

• 

Mirrored from Crime and the Blog of Evil.

Comments

[hubbit.livejournal.com]

I never actually used automation, although my late lamented 24 track digital recorder/mixer was fully capable of it. :)

[solarbird]

it's official

it's awesome

[sunfell.livejournal.com]

I have a clueless no0b question:

Do you use ProTools? How do you like it?

I am going to purchase a Midi keyboard from M-Audio which has a 'light' version of this software on it. Apparently pros use it.

Sorry. I am totally at sea. I would do better learning Sanskrit.

[solarbird]

ProTools is what Ardour wants to be. It has some fantastic editing functionality I would KILL for. I don't know much about the "light" version but at free I think I feel pretty comfortable saying it'll be fine.

M-Audio's keyboards have a pretty good rep in small studios, from what I hear, so I think you'll probably like them.

Digital audio workstation software basically emulates a multi-track tape recorder (so you can record sounds, or more broadly, so you can record several different instruments, simultaneously or separately, and play them back together), mixing board (so you can change things like tone and volume on each individual track, on collections of tracks, and on the whole thing), and, generally, effects boxes that do all kinds of different things depending on the effects box. (The most common I think is reverb, or that echoey sound that makes it sound like the instrument is being played in a big echoey room. But there are a lot of effects boxes out there.)

So if you've ever seen a recording studio, particularly an older one with the big magnetic tape reels and the control board with all the knobs and sliders, that's what you're getting emulated here. The knobs are things like left-right, equalisation (tone), and the big slider in the middle is volume, and there are also controls for whatever effects boxes are plugged in at the moment, and so on.

You don't actually have to use any of those if you want. The most important thing to start with is to get a good recording. You can always do effects play later. You want the recording to be at as high a recorded level as possible. (This is NOT the same as "play as loudly as possible!" You want to play at whatever tone is right, and have the recording gain turned up as far as you can without ever getting past the limit marker (red marker, 0db marker, whatever protools uses) even at the loudest part of the song.

Does that orient you a bit? There's a really steep learning curve here, I hate to say. But it won't be as steep with ProTools because you won't have to learn as much about audio driver software and conflicting sound systems and there'll be help/support and a manual.

[sunfell.livejournal.com]

I assume when you say 'high level' you mean high sample/bit rate. That makes logical sense. I am going to re-rip some of my source material at a higher rate.

I am seriously thinking about building a stand-alone computer for music. Right now, I'm going to use my main system- just to see if I can get the hang of stuff. Might start a blog, too. (or not- I have stuff scattered all over creation and need to consolidate it...)

It is a steep curve, but the results are relatively 'instant'. Thank you for letting me pick your brain.

[solarbird]

I assume when you say 'high level' you mean high sample/bit rate.
No.. Tho' that's obviously important too. I mean loudness level.

All equipment carries with it some fixed level of noise. There'll be some "noise floor" in every signal you record. So, let's say "signal" volume level goes from 0 (none) to 10 (loudest) and so does noise, and that your "noise floor" is 1. Every recording you make will have noise at volume 1. If you record your playing so that the volume is generally around 2, and you want to amplify that to make it louder (since that's quite soft), you'll be amplifying both:

Signal: 2
Noise: 1

So if you make it three times as loud, you end up with:

Signal: 6
Noise: 3

That noise will be highly audible and highly distracting! You don't want that. So let's say you record instead at 8, and decide that's too loud. That's a much better scenario, because you are also reducing both:

Signal: 8
Noise: 1

Drops to:

Signal: 6
Noise: .75

The same amount of final volume, but dramatically less unwanted noise.

See how that works?

[sunfell.livejournal.com]

Huh. You'd think that digital stuff would be 'cleaner' of noise than analog- but I guess that isn't the case, eh? I used to apply bandpass filters and notch-filters and other esoteric knob twiddling (more like pot-tweaking) to the voice circuits I worked with- additive and cumulative noise from thermal and other sources was our biggest quality control enemy. Heck noise is the enemy everywhere- our satellite receivers were cryogenically cooled to get rid of the thermal noise from the atmosphere. (The cooler the gear, the lower the thermal noise.)

You would think that over-sampling would cure some of that. At least where errors due to noise are concerned, but I can see where it could actually add to it. Volume would then be needed to drown it out. Of course, I would be wondering if the noise were consistent if a filter could be used to cancel it out.

So, step one would- at least for me- mean having as sonically 'clean' a sound source as possible. Sort of like cleaning up a layer in PhotoShop before applying it to a work in progress.

My problem with 'loudness' is that it reduces dynamic range. This is very obvious when I listen to my oldest CDs compared to the newer ones I have. I have to turn up the volume considerably on my older recordings. And I had to almost mute my new Balligomingo album I downloaded the other day because at 'normal' volume it sounded like I was standing next to a jet engine.

Am I over-thinking this? I have a great interest in both audio recording and music creation. And I have training in electronics- including audio circuits, and analog and digital radio systems and computers. I wonder how much of my previous training and experience will help- or hinder me. I haven't thought about a lot of this stuff for nearly 20 years.

[solarbird]

You'd think that digital stuff would be 'cleaner' of noise than analog
It is. And if you're doing exclusively MIDI for everything then really it's not a big deal. But if you're dealing with anything analogue, and in particular, anything with a microphone, all this becomes very important.

Of course, I would be wondering if the noise were consistent if a filter could be used to cancel it out.
Yeah, you're thinking of this in DSP terms, and you need to think of it in music terms! It'll be a big switch for you. On the whole, you want to do the least to a signal you can, unless you're going for an artificial/shaped feel to the sound.

My problem with 'loudness' is that it reduces dynamic range
No, no - I'm not talking about reducing the dynamic range, or the final result product. I'm talking about the initial step of capturing the sound signal, and I'm talking about where the peak levels are. If your intended dynamic range is 1-10, then, well, there's nothing to be done; don't compress it. But if your intended dynamic range is, say, 5 wide, on our made-up scale, then you want to record it so that the range on your track recording is 5-9, not 1-5, even if you want the final product to come out 1-5. (You'll pot it down when making your actual master recording.)

This is all, again, for noise control. If you have to amplify a signal, you're also going to amplify noise. If you're going to reduce a signal, you're also going to reduce noise. You always want to be in the latter position if possible, and not the former.

[sunfell.livejournal.com]

Sounds like I am going to have to damp down my old schooling and seek a genuine Beginner's Mind about this.

I can do that. It'll make it far easier.

[sistawendy.livejournal.com]

What exactly are you automating here?

[solarbird]

Levels, right now. But it'll automate everything it can do within a track. It's kinda awesome. I've been using compression but this is so much better.

Criminal Studios is also gonna need an icon.

[sunfell.livejournal.com]

Your icon should be a microphone behind bars.

Just sayin'...

[phillipalden.livejournal.com]

I would recommend making data back-ups at the end of each day of work.

[stickmaker.livejournal.com]

Sorry, I thought you knew...

[djrock3k.livejournal.com]

You raise an embarrassingly good point. I'm an Ableton Live user, and like most DAW's it has automation built in but I never use it. Glad to hear you are making the tools work your way. Ardour has been beckoning to me, but I resist so far. It's awfully pretty.

There are instructional videos for Pro Tools (and more) on YouTube, a heck of a resource. As Ms. Solarbird says, everything can be your little robot friend. The problem is I get so lost in the sonic tweaking. I get a sound I like (after playing around for hours...)and just sit on it.

I own some M-audio gear and it's pretty good, Prosumer I guess.

You are going to have a lot of fun :)

[solarbird]

Ardour is pretty. And when it works, it's awesome. But there's All This Shit you just need to know (or learn the hard way). I have a real love/hate relationship with it.

I guess what I'd say in particular is to remember to treat the tracks as tape decks and the "playlists" as tapes, but also remember that you're editing views onto a sound file and adding external modifiers to the views and not the underlying sound files, which is decidedly un-tape-like. If that makes sense. "Tracks" are expensive; tapes (playlists) are cheap. Um... Also the whole use-ALSA-and-Jack-and-NOTHING-ELSE thing detailed in this post. And a dedicated partition booting to Linux, and watch your USB interrupt allocations or you'll have performance issues even on awesome hardware. And NO WIRELESS with the realtime kernel omg! The wireless drivers are NOT well tested against realtime and it shows. Paricularly not if you use Windows drivers with the thunking layer. Use native Linux drivers and wired connection and an external wireless bridge if that's what you need, that'll solve a lot of problems.

(See what I mean? Seriously, what.)

I'm also using some M-Audio hardware. It's a little flaky occasionally but the quality for price is just excellent.