NJAS Member Wall Postings and Comments
←Next meeting: Oct 21, 2018 2:00 PM
Postings by society members do not necessarily reflect the views of the society or its other members. The NJAS does not endorse any products or services.
· Hipster American Songbook
posted Oct 18, 2018
A Cynic's Guide to a Perennial Topic
Slate article to read at your own peril.
Since online journalism consists of mostly opinion pieces written by twentysomethings these days, we now have the insane mashup of the perennial "new american songbook" with hipsterdom. When culture feature article writers run out of ideas, they come up with a "new american songbook" list--essentially songs they would like to see live perpetually in the mainstream, even if they were never really well known to start with. You see articles like this all the time.
What the writers gloss over is why the "American Songbook" exists in the first place. It has to do with desperate record producers, mostly. Say you are a producer and the new record doesn't quite have twelve songs yet--the artist has run dry. Heck it might be an album of covers. Or, you need to save some money on songwriters because your artist can't write their own material. What do you do??
The answer is to do what other producers have done. In fact, they essentially share a list amongst themselves. It's a list that's been honed down to "things shown to have worked in the past". Experimental failures have been eliminated. We are talking about songs that meet certain criteria: widely familiar, easy to sing and arrange, nothing offensive, nothing too artsy. In other words, the songbook is about things that eventually get covered over and over again by multiple artists. These are not songs that get played by the public over and over, or sung at weddings, or used at live events like baseball games, but covered by other artists. They aren't going to be challenging. They aren't going to be obscure. They need to be singable by most artists. They need to allow simple (read: cheap) arrangements. They shouldn't offend anyone. There are reasons artists cover "Misty" and "Moon River" a lot. It helps that the original version is somewhat lost to time, because everyone knows where "Somewhere Over the Rainbow" comes from to the peril of remakes. Songs that eventually date themselves, like "Tea For Two" do fade, though never die.
The problem for the songbook idea these days is the perpetual availability of the originals through streaming and the loss of the concept of an "album". Not only can bands and artists get away with releasing just singles, but unlike in the past, consumers are never far from the original or at least most popular version. Covering a song when there are ten versions on Spotify at everyone's fingertips is just inviting competition.
The Slate article is ridiculous for several reasons: it contains songs from within the past five years--hardly time tested, it contains smash hits that will make the remakes pale by comparison, too many "difficult" songs, songs that are already dated, is too inflected by the choices of music critics and not producers, and the list misses whole classes of songs from other areas such as movies, TV shows, video games, Broadway, anime and other admittedly non-American sources. Latin-American hits are already beating songs sung in English for YouTube view numbers, and Asian pop hits can't be too far behind in the coming years. Gone also are the days that everyone knows the same core songs and therefore provide a ready list of starting material that producers can go to and not risk their necks.
Where are covers coming from these days? Mostly the Eighties and still a lot of stuff from the Seventies; but since streaming has all of this covered, these fillers are going mostly unnoticed unless they are a TV theme song or pick something out of obscurity. That said, the artists doing the most covers are largely Latin, where the idea is more popular and sometimes allows the song to be translated into Spanish. Songs that aren't being touched largely come from earlier decades, particularly the sixties--though songs from the sixties were being covered almost from the moment they were released--traded around by producers like baseball cards--and artists often covered themselves by reworking their past material until they were satisfied with the results. It was the golden age of covers but seems to have exhausted itself.
Of course, artists are probably less than thrilled to have released a new album only to see one of the covers be the biggest hit--their original material immediately fading into obscurity. A sale is a sale, I guess, though it's got to hurt some. Novelty acts like Susan Boyle come and go without leaving a legacy outside of aspirational karaoke circles. Singers these days have to provide the whole entertainment package: singing, dancing, personality, personal drama and sex appeal. They sell books and lines of clothing and make videos, appear in movies, TV shows, commercials, video games and magazines. They constantly market themselves on talk shows, Twitter and getting into "news cycles" with their antics. This all suggests the music alone is not self-sufficient nor self-sustaining. Will this just get worse over time?
To get back to the Slate list specifically, the only song that might fit the bill is "My Heart Will Go On" by Celine Dion. Its main weakness is the strong association with Madame Dion herself, eclipsing potential rivals without a fight, but a producer could throw together a cover in less than a day and it fits the other criteria really well, so it gets my vote. Caveat reader.
· What is an FPGA??
posted Oct 7, 2018
Often used in audio, but rarely explained...
Wikipedia FPGA Article
The Field Programmable Gate Array, or FPGA, is something you hear a lot about in audio circles, either with something digital like DACs, or sometimes with amplifiers. While a "Gate Array" is a series of selectable logic circuits, like AND, OR or NOR gates (basically transistors that handle logic commands), "Field Programmable" means you can alter the behavior of the circuit itself after it is manufactured. This is a lower level of generality than you get with a CPU inside of a computer, since any CPU comes with a fixed instruction set that any software built on top it (in resident memory) would have to obey. FPGAs, at least in their basic form, allow one to create what are essentially custom CPU chips on the fly (or, more likely, a custom microcontroller). This is a less expensive option than fabricating ones own chips, and has the flexibility to be altered later if necessary. One of the main uses for FPGAs therefore is in prototyping. Each FPGA design essentially becomes a CPU with its own I/O instruction set, but no resident memory. However, they rarely get to a level of generality that computer CPU's have, often designed to complete a relatively basic set of operations. They can be very fast and have high I/O.
Modern FPGAs are more of a hybrid that lands between the raw basic designs of old and a modern CPU, as this makes them easier to program. Since the great expense is in the programming, and since the skills involved are highly specialized, making the easier to program lowers the cost of using them. However, the true FPGA will still allow programming at the gate level. A lot FPGA programming solves the same problems over and over again, so these are often built-in ahead of time to save the hassle.
For any manufacturer, the decision to use an FPGA comes down to the cost/benefits over the total sales. If sales are high enough, then custom chip fabrication is warranted. If it lands in the FPGA sweet spot you pick those, and at the low end you would pick a standard microcontroller/GPU/CPU approach. GPUs are used to mine bitcoins, for example, because they have enough parallelism and memory in combination to be better than FPGAs for mining, and they are available at lower cost and greater numbers.
An audio manufacturer can buy FPGAs and use them for their entire product line to spread out the cost. For example, on DACs at various price points, but the higher priced units might use more than one, or make improvements elsewhere in the circuitry to justify the costs.
FPGAs are not priced at the hobbyist level yet, so for personal use the closest relative might be something like an Arduino board. Also, the learning curve for Arduino and things like the Raspberry Pi I/O boards is much lower. However, they aren't good enough to create anything like your own DAC or ADC. In theory, you could hook up stacks of $5 computers on a stick devices to make your own FPGA, but getting them to talk together reliably and powering them would be an enormous challenge. Any audio device that uses on shows the designer was pretty serious about research and development, and they wouldn't be found in mainstream playback devices from say Sony or NAD.