| What is the number 44.1k derived from? [message #15505] |
Sun, 08 August 2004 02:38  |
 |
 trexrox
Messages: 165
Registered: June 2004 |
Active Member |
|
|
I know it has a little to do with frequency bandwidth... but why exactly 44,100 samples per second and not some other number like an even 44k?
Someone asked me this the other day, and you know, I couldn't give them a good answer, so I thought I'd bring it up here.
Will Loftin
Shangri-La Recording
|
|
|
| Re: What is the number 44.1k derived from? [message #15543 is a reply to message #15508 ] |
Sun, 08 August 2004 13:12  |
Gunnar Hellquist
Messages: 206
Registered: May 2004
Location: Stockholm |
Active Member |
|
|
Interesting tidbits. Just a few pointers that may help further.
NTSC is really 29.97 Hz (not 30.00) if I remember correctly, a minor thing really.
Interesting web sites:
http://www.oneoffcd.com/info/historycd.cfm
http://www.mrichter.com/cdr/primer/history.htm
Another interesting one. Dat came into play around 1986 (and I believe they did choose 48kBit simply to make it very difficult to make digital copies of CD-s.
I have no answer though, it probably was a compromise between Sony and Philips on a number of technical issues.
Gunnar Hellquist
unafiliated
|
|
|
| Re: What is the number 44.1k derived from? [message #15544 is a reply to message #15543 ] |
Sun, 08 August 2004 13:49 |
 |
John Klett
Messages: 468
Registered: April 2004
Location: Carmel, NY |
Active Member |
|
|
Mastering studios originally used PCM-1630 (and earlier version converter/encoders) that stored on U-Matic video decks like DTR-4000's that were, pretty much updated BVU-800 type machines that were dedicated to digital audio storage, were black and white only and ran at the black and white rate of 30 Frames per second. These facilities used 60Hz (field rate) black burst generators... If you used a standard U-Matic machine you wanted color or, at the very least, color correction, turned off.
If someone used a 29.97 clock the mastering and encoding from analog would be done a half percent slow so the CD's would play back a little faster... still happens if you are not careful with how you generate sync, what you sync to or how you define your clocks. That's why the stand alone clocks like Aardvark, Lucid and others have the pull up / pull down rates
48KHz was geared to broadcast and runs on the NTSC 29.97 Hz clock and is a number that was compatable with both this and PAL... so in post facilities dealing with digital video machines like the D2 and others all lock to 59.94Hz (field rate) burst. The digital audio on a helical scan machine like a D2 (which has to be obsolete now - I stopped chief teching heavy post facilities in '95) is put on the tapes by the same scanner as the video all in the same scan lines with a couple bursts of stereo audio data and then a big strip of video data. Frame rate has to be 29.97 and those machines would ONLY run at 48KHz.
John Klett / Tech Mecca
http://www.technicalaudio.com
|
|
|
| Re: What is the number 44.1k derived from? [message #47771 is a reply to message #15505 ] |
Sat, 26 February 2005 13:42 |
|
44.1K is from the following formula:
(perfectly good analog)-(high frequency tone)+(dither)X(nyquist serum)+(odd harmonics)X(distorted video clock)=CD
www.tonelux.com
As I look behind me, I see my past getting further away...
|
|
|
|
|
| Re: What is the number 44.1k derived from? [message #48727 is a reply to message #15544 ] |
Wed, 02 March 2005 11:02 |
ssltech
Messages: 4264
Registered: May 2004
Location: Florida (previously UK) |
Platinum Member |
|
|
Half a percent slower? -I think not... the difference between 29.97 and 30.00 is exactly 0.03, which is 0.1%
If you recorded to NTSC at 29.97 you were supposed to pull down 0.1% which gave sample rates of 44.056kHz and 47.952kHz.
Back in the land of PAL, we never had problems with colour or black-n-white timing differences, and it always worried me about these oddball sample rates we used to hear of in the USA.
In fact 48000Hz was the professional standard, 44100 was deliberately intended to be a deomestic "second-best" standard. Saving about 8.5% of bitrate for more playing time was a tempting bonus in CD capacity.
Keith
| MDM (maxdimario) wrote on Fri, 16 November 2007 21:36 |
I have the feeling that I have more experience in my little finger than you do in your whole body about audio electronics..
|
|
|
|
| Re: What is the number 44.1k derived from? [message #48729 is a reply to message #15505 ] |
Wed, 02 March 2005 11:06 |
|
John Klett
Messages: 468
Registered: April 2004
Location: Carmel, NY |
Active Member |
|
|
the frame rate in NTSC color video is 29.97 Hz - NOT 30.01 Hz
The sampling rate for digital audio has to be at least double whatever the highest frequency you want to accurately record and reproduce. That is called the "Nyquist" frequency.
44.1 kHz came about because digital audio was encoded and then stored as video on video tape as white and black bits.
Since there are a number of video standards a sampling rate had to be chosen that would be divide up well on to video formats having different frame rates and scanned lines per frame.
So...
In US we have NTSC. This has a frame rate of 29.97 Hz and there are 525 lines per frame (in two 'fields" per frame at a field rate of 59.94 Hz). There are 35 blanked lines that not not available for recording video information so you end up with 490 lines per frame (245 lines per field) that you can use to record stuff on. If you record 3 samples per line you get you can end up with 44.0559 kHz. (29.97 frames per second x 490 lines per frame x 3 samples per line = 44,055.9 samples per second) - which is not 44.1kHz but we'll get to that.
PAL is the other big format and is based on a frame rate of 25 Hz. There are 625 lines per frame (also in two fields) and 37 lines of blanking so that leaves 588 lines per frame (294 lines per field) to record on. If you record 3 samples per line you end up with 44.1 kHz. (25 frames per second x 588 lines per frame x 3 samples per line = 44,100 samples per second)... ah! that works!
Now... the information was stored as black and white video. On the Sony 3/4" machines used for mastering (BVU-800 and others before we got the official purpose built machines) you had to turn off the color - color correction circuits totally play hell with the black to white transitions - you don't want that.
The standard for black and white TV - before the NTSC (Never The Same Color) color framing rate was rolled out - was 30Hz. Like PAL this is half the electrical line frequency. So... 30 frames per second x 490 lines per frame x 3 samples per line = 44,100 samples per second... WOW!
THIS whole 30Hz vs 29.97 thing has spawned a whole set of options and requirements in the gear we use... and created a lot of confusion with digital audio sample rates and time code with reference to house sync and video frame rates. Among other things mastering studios working with old PCM1600 and PCM1630 based U-Matic video tapes need 60Hz Black Burst generators (in the USA) if they want the machines to run at 44.1 kHz... and so on... all good fun and job security all at the same time... naturally this is not a problem in the land of PAL because the frame rates for color and black and white are the same.
John Klett / Tech Mecca
http://www.technicalaudio.com
|
|
|
|
|
|
|
|
|
|
|
Members
Search
FAQ
Register
Login
Home
and the weird 44059Hz SR. There are "compressed" versions in the Proteus modules- compressed in that the SR is reduced and the loops start sooner so as to save memory space.
