Sound Cards for Broadcast Use

Computer audio sound cards are the norm at nearly all radio stations. I often wonder, am I using the best audio quality sound card?  There are some trade offs on the quality vs. cost curve.  At the expensive end of the curve, one can spend a lot of money for an excellent sound card.  The question is, is it worth it?  The laws of diminishing returns states: No.  High quality reproduction audio can be obtained for a reasonable price.  The one possible exception to that rule would be production studios, especially where music mix downs occur.

I would establish the basic requirement for a professional sound card is balanced audio in and out, either analog, digital or preferably, both.  Almost all sound cards work on PCI buss architecture, some are available with PCMCIA (laptop) or USB.  For permanent installations, an internal PCI buss card is preferred.

Keeping an apples:apples comparison, this comparison it limited to PCI buss, stereo input/output, analog and digital balanced audio units for general use.  Manufactures of these cards often have other units with a higher number of input/output combinations if that is desired.   There are several cards to choose from:

The first and preferred general all around sound card that I use is the Digigram VX222HR series.   This is a mid price range PCI card, running about $525.00 per copy.

Digigram VX222HR professional sound card
Digigram VX222HR professional sound card

These are the cards preferred by BE Audiovault, ENCO and others. I have found them to be easy to install with copious documentation and driver downloads available on line.  The VX series cards are available in 2, 4, 8, or 12 input/output configurations.  The HR suffix stands for “High Resolution,” which indicates 192 KHz sample rate.  This card is capable of generating baseband composite audio, including RDS and subcarriers, with a program like Breakaway Broadcast.

Quick Specs:

  • 2/2 balanced analog and digital AES/EBU I/Os
  • Comprehensive set of drivers: driver for the Digigram SDK, as well as low-latency WDM DirectSound, ASIO, and Wave drivers
  • 32-bit/66 MHz PCI Master mode, PCI and PCI-X compatible interface
  • 24-bit/192 kHz converters
  • LTC input and inter-board Sync
  • Windows 2003 server, 2008 server, Seven, Eight, Vista, XP (32 and 64 bit), ALSA (Linux)
  • Hardware SRC on AES input and separate AES sync input (available on special request)

Next is the Lynx L22-PCI.  This card comes with a rudimentary 16 channel mixer program.  I have found them to be durable and slightly more flexible than the Digigram cards.  They run about $670.00 each.  Again, capable of 192 KHz sample rate on the analog input/outputs.  Like Digigram, Lynx has several other sound cards with multiple input/outputs which are appropriate for broadcast applications.

Lynx L22-PCI professional sound card
Lynx L22-PCI professional sound card

Specifications:

  • 200kHz sample rate / 100kHz analog bandwidth (Supported with all drivers)
  • Two 24-bit balanced analog inputs and outputs
  • +4dBu or -10dBV line levels selectable per channel pair
  • 24-bit AES3 or S/PDIF I/O with full status and subcode support
  • Sample rate conversion on digital input
  • Non-audio digital I/O support for Dolby Digital® and HDCD
  • 32-channel / 32-bit digital mixer with 16 sub outputs
  • Multiple dither algorithms per channel
  • Word, 256 Word, 13.5MHz or 27MHz clock sync
  • Extremely low-jitter tunable sample clock generator
  • Dedicated clock frequency diagnostic hardware
  • Multiple-board audio data routing and sync
  • Two LStream™ ports support 8 additional I/O channels each
  • Compatible with LStream modules for ADAT and AES/EBU standards
  • Zero-wait state, 16-channel, scatter-gather DMA engine
  • Windows 2000/XP/XPx64/Seven/Eight/Vista/Vistax64: MME, ASIO 2.0, WDM, DirectSound, Direct Kernel Streaming and GSIF
  • Macintosh OSX: CoreAudio (10.4)
  • Linux, FreeBSD: OSS
  • RoHS Compliant
  • Optional LStream Expansion Module LS-ADAT: provides sixteen-channel 24-bit ADAT optical I/O (Internal)
  • Optional LStream Expansion Module LS-AES: provides eight-channel 24-bit/96kHz AES/EBU or S/PDIF digital I/O (Internal)

Audio Science makes several different sound cards, which are used in BSI and others in automation systems.  These cards run about $675 each.

Audio Science ASI 5020 professional sound card
Audio Science ASI 5020 professional sound card

Specifications:

  • 6 stereo streams of playback into 2 stereo outputs
  • 4 stereo streams of record from 2 stereo inputs
  • PCM format with sample rates to 192kHz
  • Balanced stereo analog I/O with levels to +24dBu
  • 24bit ADC and DAC with 110dB DNR and 0.0015% THD+N
  • SoundGuard™ transient voltage suppression on all I/O
  • Short length PCI format (6.6 inches/168mm)
  • Up to 4 cards in one system
  • Windows 2000, XP and Linux software drivers available.

There are several other cards and card manufactures which do not use balanced audio.  These cards can be used with caution, but it is not recommended in high RF environments like transmitter sites or studios located at transmitter sites.  Appropriate measures for converting audio from balanced to unbalanced must be observed.

Further, there are many ethersound systems coming into the product pipeline which convert audio directly to TCP/IP for routing over an ethernet 802.x based network.  These systems are coming down in price and are being looked at more favorably by broadcast groups.  This is the future of broadcast audio.

AM can’t wait (can it?)

Click on picture for original memo (.pdf)

I was digging through some old manuals at the shop today and I found this June 1987 memo from Orban to AM stations titled “AM radio CAN sound almost like FM.”

The main purpose of the memo was to get AM radio stations to implement the NRSC standard for pre-emphasis and high frequency roll off to improve the sound of AM broadcasts on ordinary radios.

I am not sure why the receiver manufactures never designed an IF filter that would be compatible with NRSC, it seems like a fairly simple design.  Instead, what we have is “digital” AM radio (IBOC) which does not work well, creates many more problems with interference that of pre NRSC broadcasting.

If one were to look at the entirety of AM broadcasting history, one would find some striking parallels with what is happening with IBOC today on both AM and FM.

To start, the NAB began petitioning the FCC to allow more AM broadcasting stations, even as it was known that these stations would create interference with existing stations, especially at night.  Still, the NAB persisted and the FCC relented and through the fifties, sixties, seventies, and eighties many more class II and III stations were established on what used to be clear channels (classes I and IA).

Once the AM band was chock full of stuff, they began going to work on the FM band with 80-90 drop ins.

You see, for the NAB, more radio stations means more dues money, greater lobbying power because of the larger size of the industry.   Then came deregulation of ownership limits.  By this time, Big Group Radio was calling the shots and they wanted more.  This led to the great consolidation rush of the late 1990’s from which the radio industry is still reeling.  The consolidation rush led to highly overpriced radio stations being leveraged to the absolute maximum, leading to the recent bankruptcies.

Finally, the NAB’s great push toward adopting IBOC digital radio in the early years of the 00’s.  IBOC was supposed to save the day, greatly improving quality of both AM and FM and bring radio into the 21st century.  Except that the promised technical advances never materialized.  IBOC remains a great expensive boondoggle and I am beginning to think that perhaps we should stop listening to the NAB.

The memo itself is a fascinating thing, which were one could substitute AM with RADIO and come to some of the very same conclusions today regarding analog and IBOC digital radio.  For example, this paragraph on AM stereo:

AM stereo was thought to be an answer (to improve AM), but AM stereo was embraced with the false assumption that having ‘stereo’ automatically meant having ‘high fidelity’.  While AM stereo did provide somewhat better fidelity, it was not comprehensively engineered to get the best fidelity from AM.  It was hoped that the gimmick of having two channels would be enough to save AM.

AM stereo could have been an improvement, had it been properly implemented.  Unfortunately, the underlying problem of bad sounding receivers was never addressed.  About which, the same memo notes:

Receiver manufactures did what they could to reduce listener complaints – – they narrowed the bandwidth (thereby reducing audio fidelity) until the complaints about interference stopped.  Listeners clearly indicated, through their buying habits, a clear preference for lower fidelity over continuous irritating static, buzzes, whistles, and “monkey chatter’ from adjacent stations.  People accepted this situation for a long time – – until the simultaneous advent of improved receiver technology and the FCC’s anti-simulcasting rules created the FM boom of the late 1970’s. (ed note: I remember listening to FM because there were fewer commercials, not better sounding audio)

Then the memo goes on to stress the importance of implementing NRSC standard for AM broadcasting that included the sharp frequency roll off at 10 kHz, noting that receiver manufactures would design “fine new receivers” that would take full advantage of the new standard, but only if broadcasters first showed good faith by widely and promptly implementing it.

As I recall, NRSC-1 was adopted as rule of law by the FCC in 1989, about two years after this memo was written.  One could reasonably expect that receiver manufactures then started producing radios that took advantage of the NRSC pre-emphasis curve with IF filters that did not cut off audio frequencies above 3.5 kHz, but rather rolled them off in a gentle slope until about 7 kHz, more aggressively after that until 10 kHz, where they cut off.

Except they didn’t.

Instead, twenty years later, AM radios universally sound bad, with audio bandwidth of about 3 kHz or so.

I believe that AM receivers could be made with three IF bandwidths, automatically selected based on signal strength.  Within the 5 mv contour, full (10 kHz) audio can be reproduced using a high frequency roll off described above.  In the 1 – 5 mv contour, a 6 kHz bandwidth and less than 1 mv a 3 kHz bandwidth.  The automatic selection could be defeated with a “wide/narrow” IF bandwidth selection switch like the GE superradios have.  Of course, if one where listening to stations transmitting AM IBOC, the “narrow” setting would be the best.

Half of me thinks that the ship has already sailed on AM broadcasting.  The stations on the air will continue to decline until they are no longer able to broadcast due to expensive repairs or replacement, at which time they will be turned off.  The other half thinks that AM radio, as evidenced by the huge public response to WEOK and WALL broadcasting the true oldies channel, can be revived.  With the impending inevitable FM IBOC power increases, translator shoe ins, LPFM, etc; the FM band may become worse than the AM band.  At which point the public will have to decide whether free radio is important to them, or 3G/4G services will become the new method of broadcasting.