Talk:Kansas City standard
From Wikipedia, the free encyclopedia
I presume "asynchronous" here means "in asynchronous start-stop serial format".
-
- Yes -- Egil 11:34 Mar 10, 2003 (UTC)
As I recall, most of the implementations generated the output in software as a square wave, which was then filtered before reaching the tape. Similarly, AFAIR the decoding was done using a hardware zero-crossing detector with hysteresis, and then post-processed in software to recover the data... Please let me know if I'm wrong about this.
Contents |
[edit] Lack of attribution problem???
The text of this article sounds so much like http://www.alphalink.com.au/~edsa/#kcs, I'm inclined to think it came directly from there, with just enough changes to avoid being a copyvio. RoySmith 14:00, 20 July 2005 (UTC)
[edit] 1200 baud variation
I don't know about Acorn Computer's 1200 baud version but Processor Technology had the 1200 version going in mid 1976. Lee Felsenstein was an author of the standard and a designer at Processor Technology.
I don't know what to do with UFE file format stuff. (Swtpc6800 23 February 2006)
It may not be appropriate to refer to Processor Technology's 1200 baud format as Kansas City Standard as it used different tones (600Hz and 1200Hz). PT did however record their tapes in both formats - 300 baud KCS on one side and 1200 baud CUTS on the other.
Add to the confusion is that PT appear to use the acronym CUTS for two different things: 'Computer Users Tape Standard' and 'Computer Users Tape System' - the former referring to their 1200 baud format; and the latter to their cassette interface unit (which supported both KCS and CUTS formats).
http://www.sol20.org/articles.html has a number of Processor Technology newsletters and adverts.
It is interesting to note that only in the UK was the term CUTS used synonymously with 300 baud KCS - most notably by ACORN and the Compukit UK101 makers. (Ed. 4 November 2006)
I wrote some DSP code to read BBC-B (ACORN) tapes via a soundcard... And I'm certain that it was 1200 baud modulated C/W onto the 1200Hz carrier. The UART strobes were kept in sync with the carrier generator. During playback a phase locked loop was used to maintain the carrier sync while reading a zero. ISTR that earlier machines used a different format, which might have been KCS. The format used by the BBC-B is described in after market editions of the Users guide, and the tape interface ASIC was described in the Advanced Reference Manual. (I'll get you references for this next time I raid my loft!) 194.129.249.240 16:37, 2 September 2007 (UTC)
[edit] BYTE's Audio Cassette Standards Symposium
Manfred and Virgina Peschke - BYTE Magazine, Feb 1976, Pages 72 and 73
-- SWTPC6800 03:24, 9 August 2007 (UTC)
[edit] 1200/2400Hz: What an odd choice!
Never being aware of the technicalities of this standard before, I am quite surprised that 1200/2400 Hz tones are used. While generating them is obviously extremely simple, this choice suffers from a very obvious drawback, in that distorted playback will tend to cause the 1200Hz to be decoded as 2400Hz - the second harmonic. Distinguishing the tones then becomes very difficult and error prone. No doubt square waves were also used in the encoder part, followed by a crude analogue filter. This makes the problem potentially even worse. Is there any evidence to support this poor performance prediction? A better choice would have been to use harmonically unrelated tones as in selective calling systems, though this would have added complexity. A compromise might have been to use 1200/1800Hz in a 1:1.5 cycle ratio, as in FFSK signalling systems. These are very noise and distortion proof, yet easy to generate. Graham 13:02, 8 March 2006 (UTC)
My theory is that the participants in the symposium were not signal engineers, but ranged from computer "enthusiasts" to people who work with digital computers, so they did not have the experience or knowledge to understand that issue. -Luciuskwok 04:08, 29 May 2006 (UTC)
[edit] also used by sinclair?
at one, more geekier time, i did a little bit of audio analysis on a bit of spectrum-sounding noise included in a certain piece of cult pop music (reminiscing on memories of that very computer)... though i can't remember the frequencies involved, the oscilloscope-style waveform certainly looked a lot like what i'd expect KCS to. (even managed to decode what it said, thru bits into ascii bytes). can't see it attributed here - was it KCS that sinclair used (for the zx80, 81, spectrum, etc)?
certainly the loading times would suggest such - even a BASIC listing of a couple hundred lines, or a 16k commercial program took a couple of minutes, a 48k one would allow you time to make a cup of tea, and a 128k time enough to cook lunch and eat it too... plus the infamous R: TAPE LOADING ERROR whenever the rather fragile codec suffered a dropout (or harmonic distortion, etc) somewhere along the line.
[edit] Atari variation?
I know Atari 800 (and later, esp. 65 XL/XE series, I own a few) used AFSK to encode data on audio tapes. The data bit was lead by one start bit, and one stop bit, making it a total od 10 bits per byte. The frequencies are a little different, though - mark tone at 5327 Hz, space tone at 3995 Hz (or close). The effective standard baud rate was 600 (but POKEY could accept it even if it went as high as 820, sometimes more). The data was written in series of bytes separated by carrier tone.
I don't know what kind of 'standard' that is, though. The Atari machine, when instructed to save data to tape, did record its data in one format only, true - but original games came pre-recorder from the manufacturer, and sometimes the recording technique was somewhat different (for example, considerably longer data segments).
Perhaps I am missing an article that already describes this? If not, should this be included? —Preceding unsigned comment added by 85.221.137.248 (talk) 11:32, 6 May 2008 (UTC)
