Audacity digital editing software

Audacity logoAudacity is the name of a free digital audio editing software package made distributed by Sourceforge. It is distributed under Version 2 of GPL without exceptions.   It does require an .mp3 plug in to generate mp3 files.  According to the Sourceforge website:

Audacity was started by Dominic Mazzoni and Roger Dannenberg in the fall of 1999 at Carnegie Mellon University. It was released as open-source software at SourceForge.net in May of 2000…

Audacity is a free, easy-to-use and multilingual audio editor and recorder for Windows, Mac OS X, GNU/Linux and other operating systems. You can use Audacity to:

  • Record live audio.
  • Convert tapes and records into digital recordings or CDs.
  • Edit Ogg Vorbis, MP3, WAV or AIFF sound files.
  • Cut, copy, splice or mix sounds together.
  • Change the speed or pitch of a recording.

The full list of features is available here.

So, I have download a copy and installed it on my test machine in the basement (hardware requirements here).  My test machine is a stripped out P4 2.4 GHz Windows XP box that I can isolate from the network and experiment with.  On that machine with a digigram VX-880 soundcard, Audacity did very well.  I did not record multi track, but with 24 bit sound sampled at 48 KHz, the computer kept up nicely.  The basic editing features are intuitive and easy to manipulate with mouse and keypad.

For a quick to install downloadable program, it does very well.  Does it do everything like Adobe Audition or other professional editing software suite does? No.  But for the price, it can’t be beat.

RCA receiving tube manual, 1964 edition

RCA receiving tube manualI found this in the great clean out of 2010, Bridgeport, Connecticuit.  Once upon a time, I had a slightly newer version of this, I think from 1972 or so.  This version is from 1964 and gives a complete run down of most small tubes that were manufactured back then.

There is something about a well designed, well maintained piece of tube gear.  I remember an old Collins tube console that was in a production room at a small AM station.  The console went dead (paper clip shorted the B+) and I fixed it.

I recall listening to the test recording of my own voice from a reel to reel machine when I fixed the console.  It sounded better than I’d ever heard it, not that I have a great radio voice, by any means.

A tube is a voltage amplifier versus a transistor, which is a current amplifier.  A tube does not have the same fidelity as a transistor, as the voltage reaches it’s peak, it gets a little fuzzy, adding some distortion and harmonics.  Tube gear adds warmth, what a musician might call Timbre. The combination of fundamental frequency and varying amplitudes of harmonic frequencies that allow a listener to tell the difference between a piano and a guitar playing the same note.

This is what the current crop of tube mic preamps and other tube products tries to reproduce.  Several companies have come out with an amplifier design that has mostly transistors and one tube, usually a 12AX7.    Unfortunately, at least in my opinion, the fall a little short.  If you want to have the “tube sound,” it needs to be all hollow state.

What is intriguing to me are the schematic diagrams in the back of the book.  There is one for an audio amplifier:

RCA receiving tube manual audio amp schematic diagram C 1964
RCA receiving tube manual audio amp schematic diagram, C 1964

This is a single channel unit, for stereo, it would need to be duplicated.  Also, I would loose the tube rectifier in favor of a solid state full wave bridge, that would simplify T2 somewhat.  The OA2 could also be changed to a diode.  Looks like unbalanced audio in, which could be modified with an input transformer.

Another interesting diagram is this one, which can be used as a mic preamp:

Microphone Preamp schematic diagram RCA receiving tube manual C 1964
Microphone Preamp schematic diagram RCA receiving tube manual C 196

That looks like a pretty solid design, a few tweaks here or there to add some gain reduction and some type of output level adjustment and I would be a really cool piece of gear.  Again, the tube rectifier could be replaced with something solid state.  The output transformer would likely have to be changed to something like 600 ohms.

Couple that to something like this, the Collins 26U compressor/limiter, and one would have a great sounding microphone processor:

Collins 26U compressor limiter
Collins 26U compressor limiter

Looks pretty cool.  R10 is used to balance the two plate currents.  I would be interested in the transformer values, input/output impedances, voltages, etc.  The 6386 tube is very hard to find these days, a good substitute would be a 5670 which are still made by several manufactures.

Update:  This is a picture of a Collins 26-U sitting on my bosses floor.

Collins 26-U compressor limiter
Collins 26-U compressor limiter

A great online source for tube information is: Electron Tube Data Sheets.

If I have some time this winter, it may be a fun project to fool around with.

Category 5e cable in AES/EBU applications

I love wire. I know, what a geeky thing to say, but it is true. For no reason that I can explain, I have always been fascinated with wire, cables, and electricity.

Category 5e and up cabling is amazing stuff.  Designed for computer applications, it can fulfill a wide variety of rolls in radio and television stations, mostly because of its high bandwidth capacity.  Category 5e cabling has a 100 MHz bandwidth, category 6 bandwidth is 250 MHz, with 6a (augmented) being 500 MHz. AES/EBU audio uses ATM and requires from 4-26 MHz bandwidth, depending on the sample rate (highest sample rate is 200 KHz).  Category 5e cable has a minimum common mode balance of -60 dB, which makes it nearly impervious to RF, electrical noise, mutual interference and other noise issues.

Further, each pair in a category 5e or 6 cable has a different twist rate, to reduce cross coupling between pairs.

Here is a chart of electrical characteristics for Cat 5e, Cat6, and Belden AES/EBU cable:

Category 5e Category 6 AES/EBU*
Impedance (ohms) 100 +/- 10% 100 +/- 10% 110 +/- 20%
Bandwidth (MHz) 100 250 52
DC resistance 1K Ft (ohms) 27.12 27.42 23.68
Capacitance per Ft (pF) 15 15 12
Velocity factor (%) 72 72 76
Common Mode Balance (dB) 60 63 30
NEXT -44.3 -53.4
Configuration UTP UTP SPT
Wire Solid Solid Stranded
Gauge 24 23 24

* Belden 1800F

Specifications above are for Belden cabling, but are typical for high quality category cabling available from other sources as well.

Although the AES/EBU cable specifications call for 110 ohm impedance cable, that specification is pretty loose, calling for +/- 20%, which means 88 to 132 ohm cable will work well.  Category 5e and 6 cable is 100 ohm impedance, +/- 10%, which translates to 90 to 110 ohms, nominal.

Category 5 and 6 cabling can also be used for analog audio, RS232 and RS485 applications.  One area of caution, however, is for T-1 or fractional T-1 services.  On the DS side (between the smart jack and the CSU), T-1 type service runs 3 volts peak to peak.  That is much higher than AES/EBU or ethernet, which run 1 volt peak to peak.  As a result, cables in this type application should be 22 gauge or higher to reduce emissions from the cable.

Shielded category cable is available in Cat5 and 5e.  The shielding acts to reduce emissions from the cable in low noise environments.  It can also act to reduce RF fields around the cable pairs, so long as the proper cable terminations for shielded cable (RJ-45 or more properly 8P2C connectors) are used and installed correctly.  The shield must be connected to a ground on at least one end.  I know a facility that has all shielded Cat5 cable, but they used standard RJ-45 connectors, so both ends of the shield are floating, which completely defeats the purpose of the shield.

Cat5e 25 pair cable25 pair category 5e cable is available for trunk cabling between studios and the technical operations center.  For one studio project, I purchased pre-made cables with RJ-21 connectors on both ends.  Those connectors were then plugged into KRONE LSA-PLUS blocks.  Cable, connectors and blocks were all 100 ohm impedance, category 5 equipment.  Since we did not have to strip any insulation or punch down any wires, we pulled and terminated the studio to rack room trunk cables for five air studios and three production rooms in one morning.  This greatly sped up the studio build out process.

The studios and TOC use SAS 32KD (Sierra Audio Systems) audio router and Rubicon SL consoles, so most of the audio is AES/EBU.  There are, however, several analog audio sources that are included in this system, things like telephone caller audio, off air monitors, satellite feeds and remote broadcast sources.

This facility is located about 1 mile away from a 5 KW AM station on 850 KHz.  Several concerned people commented on the possibility of RFI on the cabling.   In the five years since that project was completed, there have been zero issues with the cabling or the audio quality.

One thing to consider in these installations is the length of the cabling and the sample rate being used across the network.  The capacitance per foot is the deciding quality in cable lengths.  This is because capacitance, which is the ability to store an electrical charge, will begin to distort the signal (turn it into a saw tooth waveform) in the cable if certain lengths are exceeded.  A good way to calculate maximum cable runs is thus:

Most professional AES/EBU devices sample 24 bits per channel, if the sample rate is 48 KHz, the 24 bits x 48,000 Hz = 1,152,000 bits per second per channel.  For stereo, as most applications will be, that is doubled to 2,304,000 bits per second, or 2.3 Mbps.  There is some overhead in an AES/EBU signal, so, for arguments sake, we will call it 4 MHz.

In this facility, the sample rate is locked at 48 kHz by a master clock.  The longest cable length is 145 feet, which adds (15 pF x 145 Ft) up to 2,175 pF capacitance.  From the chart above, we know that Cat5e has a resistance of 27.42 ohms per 1000 feet, or 0.02742 ohms per foot.  That works out to be 145 feet x 0.02742 ohms = 3.9759 ohms.

To calculate the capacitive reactance, the following formula is used:

Xc= -1/(2π FC)

Where Xc is the capacitive reactance, F is the frequency in Hz and C is the capacitance in Farads.

Therefore Xc = -1 / (2 x 3.1415 x 4,000,000 x 0.000000002) = -19.89 ohms.

The characteristic impedance of Cat5e and Cat6 is 100 ohms.  The DC resistance is 3.97 ohms and the capacitive reactance is -19.89 ohms, make the circuit impedance of a properly terminated cable 145 foot cable 84.08 ohms.

The design formula for a low pass filter is thus:

fc = 1/(2πRC)

Where fc is the cutoff frequency, R is the resistance and C is the capacitance.

Therefore, fc= 1/(2 x 3.1415 x 3.9759 ohms x 0.000000002 farads) = 20,014,958 Hz or 20 MHz.

Generally speaking, one should try to keep the capacitance below 2500 pF in a 10 Mbps circuit.  Belden datatwist 1212 cable has a 4.0 dB insertion loss and a 23.0 dB return loss per 100 meters (328 feet) at 4 Mhz.

145 feet is well within the limits of this cable for AES/EBU applications.

Further, all cable circuits need to be properly terminated to reduce return loss.  Using common impedance wiring blocks, connectors and terminations help keep return loss to a minimum.  Stranded wire works better in applications where cabling may move.  There are Cat5e and Cat6 stranded cables available.

As data transfer rates approach that of RF, ethernet, digital audio, and RF are going to seem more and more similar.  1000 Base T (1GBT) and 10000 Base T (10 GBT) networks are coming.