I know that FM radio uses 32khz sample rate digital .
I was just wondering if anyone knows is there a prefered format for submiting music ,
The tracks I want to send are originally 24bit 96khz , Im sure its only a press of a button for them to sample rate covert but preferably I want to be able to quality check whats going out on air rather than leaving it to chance.
Are there any other things I need to do considdering bandwidth at 16 khz would already be a few db down from flat @32khz FS?
I had someone suggest sending 16bit 44.1khz is the way its done ,but that makes little or no sense to me , 32khz is an easier multiple of 96khz for sample rate conversion, where 44.1 khz is likely to be more of an approximation and could possibly contain artifacts.
Nowhere in the specification for FM stereo does the number "32khz" apply.
FM stereo, as an analog system, uses a 38kHz double-sideband, (suppressed) subcarrier for L-R (stereo) program material. Somewhat similar to digital sampling bandwidth requirements with regard to the Nyquist sampling frequency, any input material over 19kHz would be aliased back into the main (mono) channel. A synchronized "pilot" sine at 19kHz at -20 is also included as part of the stereo transmission standard, for the receiver to use as a reference to accurately re-create the 38kHz needed to decode the L-R on the subcarrier.
To eliminate aliasing and to protect the integrity of the 19kHz pilot, FM stereo transmission equipment has always used low-pass filters at its left and right channel inputs to restrict their audio bandwidth (Mono FM had no such limitation). When FM stereo was adopted in the US in 1961, the technology to do this was rather primitive, usually using L-C filters; the best they could do was to allow an ultimate bandwidth typically out to about 15kHz.
Even today, you can confidently expect that any material you provide to an FM radio station will ultimately be low-passed to roughly 15kHz bandwidth, but this exact figure is dependent upon the filter design of the stereo encoder used by the radio station.
The upshot is that I wouldn't worry at all about sampling frequency when sending material to a radio station.
MUCH more important, from a creative standpoint:
In the US, 75µsec. pre-emphasis/de-emphasis is used in transmission (in other parts of the world, its a less drastic 50µsec.) The collision of technical, legal and subjective requirements results in modern broadcast audio processors **heavily** dynamically manipulating the upper octaves. Long story short*: you will never see 15kHz at full amplitude at the output of an FM receiver, more like -17dB, with respect to 1kHz reference in the U.S. If you, as a content provider, expect to push dense high frequencies in your music, expect those highs to be somewhat (or even dramatically) muted when played over an FM station.
David Reaves
Recklinghausen, Germany
*The LONG story is that the FCC and other regulating organizations set a limit to the maximum modulation (peak audio level) a station may use, in order to prevent interference with other stations. For FM broadcast, 100%, the legal max, is a +/- 75kHz occupied bandwidth at the station's licensed frequency, which is assigned between 88 and 108 MHz.
FM is inherently slightly noisy, with the noise spectrum having a rising tendency toward higher audio frequencies. So from the beginning, pre- and de- emphasis was implemented as a form of non-dynamic noise reduction. Since pre-emphasis boosts the high audio frequencies disproportionally (with 75µsec. pre-emphasis the boost is +17dB at 15kHz), high frequency program material will reach the 100% legal modulation maximum much sooner than lower frequency material. To prevent over-modulation with the high frequencies, it meant that as long as the signal was unprocessed, typical low frequency program material would never come even close to 100% modulation. This was the state of the art in 1961 and was fine so long as program material, such as classical music, was not overly rich on the high end. Light-duty, broadband peak limiting after pre-emphasis was sometimes used to allow some compensation for overall levels, but too much of that would be very audible, so at that point overall loudness of FM was relatively low.
With the advent of processed pop music in the 1960s, dense high frequencies in program material became much more common, and the FM standard - with its existing processing equipment - was fully unsuited to it. At about the same time, radio station managers decided that to be competitive they wanted more consistency in the loudness, which pre-emphasis seems to prevent.
Enter specialized multi-band FM audio processing, capable of independently pushing down excessive high frequencies while generally leaving low and mid-range material intact. Effectively, the modern processor for FM dynamically reverses the effects of pre-emphasis, and this has become the norm. When fully engaged, the effect is to produce a dense concentration of energy across the audio bandwidth, AFTER pre-emphasis, to modulate the transmission consistently at 100%. Upon de-emphasis in the receiver, this concentration of energy follows the de-emphasis curve, resulting in the reduction of high frequencies, -17dB at 15 kHz.
