Wednesday, September 9, 2015

The pseudo-science of digital audio recording

It wasn't so long ago that digital recording didn't exist or wasn't accessible at all to anyone with any conservative budget. Because it's so new, it should've been a red flag to me to see so much "expertise" in the area of advice to young "recording engineers" (of which I've been) and I should've done a bit more research into the history of the art-form before jumping right in. I started recording music in 2001 or so and I started digitally. My first recording device was a Fostex MR-8 Digital Multitracker. Only last year did I ever experiment with using analog tape to record live music and the experience has changed my approach, to say the least. Not because it sounds "better"... but because the way you get signals to tape (or in the case of digital, hard-drive) is, in my opinion, overlooked.

Let me elaborate... When you record to tape the audio signal needs to live on the physical magnetic tape, whereas with digital the signal is converted to digital information (i.e. 0s and 1s) and this difference has an effect on how you should be recording. Now, the size (width) of the tape is finite and so is the size of the digital storage space, but the sizes are different.

The big mistake I was making was that if you read any beginners guide to recording they tell you this: Get the loudest signal you can get without going into the red (clipping) for each component of your music, then once it's all recorded, go back and "mix" your audio and adjust the volumes. I blindly followed this advice for years but now, I don't anymore and this is why...

When you record to tape, if the signal is too hot you will achieve tape "saturation" which is generally an unpleasant sound of distortion, but before hitting this "point of no return" you will hear the tape gladly accepting the signal and letting it fill out all of the space it is alotted: generally this is what's referred to as "warmth," i.e. natural tape-based signal compression. Now the physical size (i.e. width) of the tape is only 2 inches or less, so it's generally a small space to live. I think the idea of capturing  the most digital info as possible in terms of the various components to a mix is mis-guided, almost by definition. This is because the point isn't to hoard all the sonic information just because you can, the point is to record the sound you want to hear on playback. The luxury of "unlimited tape" blinded the neophyte digital recordists ability to properly engineer mixes. The analog workflow has proven to be essential to a natural sounding recording. When you dive deeper into this subject it becomes a combination of philosophical decisions and assessment of history, but ultimately it's about one thing... what sounds good?

Of course we aren't taught this concept of commitment and technique to recreate analog warmth on a digital audio capturing device when it comes to digital recording expert advice and that's where my issue is. If you apply this strategy of get-the-loudest-signals-you-can-without-clipping-and-then-adjust-in-mixing to your digital recording, you'll make a recording where it will come out sounding not necessarily over-saturated, but somehow wrong. It's too many hot signals competing with one another... and only one or two break through at a time. It's a lot for the ear to handle. The processing power of digital recording is mis-used in this context and the results are less than good typically. At long last, clarity?

So how do you achieve tape warmth on a digital recording interface? Change the way you take in audio signals. Just take in the signal at the volume it will be heard at in the mix (i.e. an appropriate and accurate reproduction of what you're actually hearing in the room) and don't go as high as the red, don't even go as high as the yellow. Don't have to adjust any volume knobs at all in your DAW. They should all always be at zero. I personally record to peak signals somewhere between -24dB and -12dB depending on the instrument. I assess this input level as I mostly focus on what I'm hearing through the monitors. If you're using a mixer, put your faders all just below 0 and adjust your input signal volumes with gain (and EQ) only. Your results will be better because you are following an analog workflow for your digital session, and the unity gain simulates the concept of natural tape compression in that it limits the input depth? Something like that... I'm still working on articulating this properly.

Monday, September 7, 2015

Tutorial: G-chatting in the Terminal

This is a Mac OSX tutorial to install Gchat in your Bourne again Shell with bitlbee+irssi via MacPorts

Step 1. Install MacPorts, which is a package manager which allows you to easily install utilities.
In this instance, we'll use MacPorts to install bitlbee and irssi, see Steps 2 and 3.
Go to The MacPorts website and click on "Installing MacPorts"
This tutorial is for Mac OSX, specifically, I am running Mavericks, which is OS 10.9.
During the MacPorts installation they may ask you to also install Xcode, but I've skipped this in the past with no real consequences.

Step 2. Install irssi, which is a chat client that supports IRC protocol.
To install it... open up a shell and run $ sudo port install irssi
Along the way, it asked me to get jiggy with something called javac, I obliged.

Step 3. Install bitlbee, which is an IRC client that interfaces with irssi and conveniently lets you use multiple IM services at once.
So you can login in once and have your MSN, AIM and gchat all ready to go.
To install it... open up a shell and run $ sudo port install bitlbee

Step 4. run $ bitlbee
Here was my result:
$ bitlbee
Warning: Unable to read configuration file `/opt/local/etc/bitlbee/bitlbee.conf'.
:localhost.localdomain NOTICE AUTH :BitlBee-IRCd initialized, please go on
:localhost.localdomain NOTICE AUTH :If you read this, you most likely accidentally started BitlBee in inetd mode on the command line. You probably want to run it in (Fork)Daemon mode. See doc/README for more information.
ERROR :Closing link: Ping Timeout: 180 seconds

so then I ran...
$ cd /opt/local/etc/bitlbee
$ sudo mv bitlbee.conf.sample bitlbee.conf
$ sudo emacs bitlbee.conf
...Then I went in and changed the line
#RunMode = Inetd
RunMode = ForkDaemon
now I rerun
$ bitlbee

Step 5.$ irssi
Now once inside the irssi interface, type in "/connect localhost" this will connect you to the local bitlbee server.
Now type in Ctrl+p to shift over to the bitlbee-connected window.
Now you can just type in "help" without the leading forward slash, to read out the basics of bitlbee.
You will need to synchronize your IM information into bitlbee so it will work, that is done like so...

Step 6.
Add your gchat account with "account add jabber" Then type in /OPER and hit enter.
It will prompt you to add a password to you account, type it in and hit enter.
You can type in "ac l" at any point to list the accounts and their statuses in your session.
To turn on your gchat, type in "ac 0 on". Each account corresponds to a number, in this case, starting at 0 your gchat account will be the first account added so it should be "ac 0 on."
Type in "blist all" to see your buddy list.
For good measure, let's also add an AOL Instant Messenger account...
"account add oscar Your_sn" should work, then simlarly, use /OPER to add the password and then "ac 1 on" to log in.

Step 7. Register your settings so you won't have to do all this everytime you log in.
Type in "register" then add your password with /OPER
At this point I got an error message of
03:06 <@root> Error while opening configuration file.
03:06 <@root> otr save: /opt/local/var/lib/bitlbee/username.otr_fprints: Permission denied
03:06 <@root> Error registering
...So I had to run...
sudo touch /opt/local/var/lib/bitlbee/username.otr_fprints
cd /opt/local/var/lib
sudo chown -R username bitlbee/
sudo chgrp -R staff bitlbee/
sudo chmod -R 775 bitlbee/
That just search did it... then once I ran register and then entered my password via /OPER prompt I got back...
03:15 <@root> Account successfully created

Step 8. Testing it out
To exit bitlbee, type in "/exit" then back at your new shell type in "killall bitlbee" This completely ends your session.
Now let's test out what we've done...
Go to a normal shell and follow this flow...
$ bitlbee
$ irssi
/connect localhost
identify your_register_password (i.e. this is what you typed in at the /OPER prompt after you typed in "register" in Step 7.
blist all

That should search wrap it up. As far as passwords and plaintext, you can google oauth and check that out and there are some other settings in certain config files that are worth checking out and understanding, you can find info online. There are other commands you'll see in there, but that's generally how it's done. Oh yeah and to chat with a buddy you type something like, hmm... /msg smarterchild hello and then Ctrl+p your way over to the chat.

Good luck!

Classic VHS Review: Interview with a Vampire

I first heard about this movie when Richard Christy (Howard Stern) talked about it on the radio and how he had a "moment" with ano...