|
The PNW Section's October 2008 meeting featured James (JJ) Johnston presenting a general overview of how portable music players work, such as the iPod or Zune.
Opening remarks and announcements came from PNW Chair Steve Turnidge. Attendees introduced themselves and were asked to mentioned some audio thing they've seen of interest recently.
Attendance was approximately 15 members & 9 non-members. Opus 4 Studios of Bothell WA hosted the event.
James (JJ) Johnston, of Neural Audio (formerly Microsoft & Bell Labs) discussed the elements that all portable music players need to work, without discussing proprietary brand specifics.
- A player needs a storage medium, a control system, playback and level controls, a headphone drive system and a power scheme, all while remaining small and running as long as possible on a charge.
- While storage space was once the big problem, now power and battery life is. A CPU, amplifiers, hard drives, memory access - all take power. JJ feels that HDs are now outmoded for this use.
- The digital processing must find the files, read the data, convert to PCM, do any processing (EQ, compression, volume in some cases) and then do D to A conversion. Thus, these are powerful little computers with an operating system, file system, decoders, signal processors and so on. Processors nowadays are powerful enough to actually be asleep much of time to conserve power.
- All lossless coders are about the same efficiency, so they are all good. FLAC and Monkey's Audio are good examples and free.
- AAC and WMA Pro10 can be very good lossy codecs. AAC was originated by JJ and Anibal Ferreira at ATT Bell Labs in the 1990s. Some discussion of perceptual coders followed, such as bit rate vs. quality. There are some messy legal issues intertwined with many lossy codecs, so progress is a problem. In general, the more efficient the encoding, the more complex (and CPU power hungry) the encoder and decoder.
- JJ admonished the group to never use more than one layer of perceptual coding (ie: donít convert from one to another), or it will sound awful.
- Signal processing uses CPU power - hence the rudimentary DSP available, such as a few fixed EQ profiles. He would like to see a smart volume system that senses the ambient noise.
- Level control is often a digital PCM operation, which can lead to a poor S/N ratio at low levels. It is better to set the output as high as possible without clipping if using the device as a line-out.
- The user interface uses a lot of power, thus it tries to sleep as much as possible.
- The power amp for the earphones usually delivers several milliwatts - sadly making people deaf, and the EU mandates an SPL limit. Many amp designs usually sacrifice some audio quality for power economy, and the output capacitors required are undersized, affecting low-frequency response. Running into a higher impedance load, such as a line input, helps.
A break was held, and door prizes awarded:
- Andy Hall - Whirlwind cap (AES Convention swag)
- Ken Meyer - Lectrosonics shirt
- Dr. Mike Matesky, Ed Gruse - Fluke VoltAlert (from Rick Rodriguez/Fluke Corp)
- Janey Wallick, Jon Harris - Opus 4 Studios travel cup (from Opus 4 Studios)
- Dan Mortensen - Universal Audio DVD (AES Convention swag)
- Kevin Shank, Bob Smith, Gary Gesellchen, Steve Turnidge - audio CDs (from Opus 4)
- Rob Baum - Bag End disc case (AES Convention swag)
- Tom Monohan - AES 125th convention exhibitor directory (AES Convention swag)
- Rob Riggs - Cogswell College bottle opener (AES Convention swag)
- Mark Rogers, Frank Laico - lanyards (AES Convention swag)
JJ's post-break bonus session was a recap of his participation at the recent AES Convention, a panel recalling the
"worst mistakes in audio." (133.6kb PPT)
JJ's nominations for the ten worst mistakes in audio were all accompanied by spirited discussions:
10 - Applying tube design to transistors.
9 - Calling "magnitude response" "frequency response", without looking at phase.
8 - Citing a Signal to Noise ratio without showing the signal and noise spectra. "A-Weighting" (aka A-message weighting) is another non-useful measurement for audio.
7 - Making an audio demonstration in a paper, workshop, or tutorial session at any conference, anywhere, due to the poor sound systems.
6 -Saying, "We'll never understand....." or "No one understands how the ear works"
5 - Pan pots - without explaining how/when to use it, how it works, how the ear hears it, not having an interchannel time delay, and worrying about mono compatibility today.
4 - Pointing out that slightly louder is perceived as better in blind testing. Loudness wars.
3 - Forgetting about the analog part of digital equipment.
2 - Trying to apply analog processing to digital systems.
1 - Letting people try to use as low a bit rate as possible for music on a perceptual coder.
His closing comments:
- New things come along. Resistance is futile, you will be assimilated.
- You donít want something new every time you have a new problem!
- Enthusiasm for the new does not excuse excess.
Reported by Gary Louie, PNW Section Secretary
|
