organ notes

As a routine practice many classical LPs were emasculated of any low bass content OR cut very softly AND/OR dynamically limited to avoid "problems."  

The other thing that was routinely done was to mono all the bass - the last thing you need on vinyl is a load of LF on one channel only - that plays merry hell with everything in the record-replay chain.

Incidentally, I've just been to listen to a 64ft stop (briefly) - and whilst you can just feel it, it really isn't very 'loud' at all, especially in comparison with the rest of the organ. You wouldn't really miss it if it wasn't there, I think. I'd rather have a 32ft bombarde on a recording any day of the week.

Before coming on-line and reading the latest posts I had just finished some Frequency Analysis experiments. I located some bass in an organ-only section a bit more than two minutes long. This happened to peak at 45.71 Hz, as near as I can tell from the graph. I chose this peak simply because it fell to hand easily when  listening for bass.

This peak stands out from surrounding frequencies by being 35dB higher than the beginning of its low frequency side slope and 46dB higher than its high frequency side end. There are, of course, also peaks for higher frequencies across much of the spectrum, but this peak is the highest on the graph.

Another possibility for the high point of this peak, for anyone trying to relate to a organ's output, is 45.33Hz (FFT Size 16,384, Blackmann-Harris window type). Nothing in-between those two frequencies is available at 16,384.

I marked a point to be the center, selected 0.5 second surrounding it, did a Frequency Analysis scan, and recorded what the graph told me about levels by placing the cursor on the graph at 45.71Hz. I also made a print screen to allow visually  comparing each measurement.

From there I clicked on the ‘zoom out' button, which triples the display size, selected the entire displayed duration, then did the same measurements and print screen. I repeated this three more times.

At each zoom position I also clicked on my center point que mark (no selection), brought up Frequency Analysis, and verified that nothing had changed due to the zoom level when only that point is selected. I then selected the entire on-screen duration and looked at that with Frequency Analysis before doing the scan. All those non- scanned graphs were identical, regardless of the duration covered. This seems to verify  the Help file's comment that the graph shows what is at the cursor position or at the selection center point (prior to scanning).

At each zoom out jump the 45.71Hz peak lowers and broadens. Looking at just the left channel (originally 2.1dB greater amplitude than the right), the decrease in amplitude for each of several zooms is 0.942dB, 3.091dB, 7.714dB, 7.35dB. By the third jump out this peak is no longer the highest on the graph. On the forth jump the peak disappeared.

I can still locate and measure 45.71Hz by placing the cursor. Perhaps it should still be called a peak of sorts. At just that point the upward slope levels out briefly before continuing up to the real local peak at 53.97Hz.

At this place the 45.71Hz peak is more than 19dB below its first measured amplitude. The adjoining 53.97Hz peak is 11.3dB below the 45.71Hz level of the first scan (which is still at the center of this selection). Therefore, in regard to MusicConductor's question, the final graph (which coves no outer rim or inter-track extents) is quite a bit lower overall in the bass region.

In fact, no peak, at any frequency, is as high in the extended selection as in the smaller selections. Those peaks are not being obscured by noise or anything else registering higher; the measurement is obviously telling us something different. That may be something very reasonable, but I am uncertain what it is.

Duration is not included in the Frequency analysis graph nor explicitly mentioned in the help file. However, any given note, such as the one producing this 45.71Hz peak, may be used from time to time in a piece but is unlikely to be used continuously. When it is used, it is unlikely, in many compositions at least, to always be at the same level. Therefore what should a scan show as the amplitude for a given frequency when that frequency appears in several dis-contiguous places in the scanned duration, each at a different levels?

Perhaps the displayed amplitude is some kind of average, but average of what over what?

The other thing that was routinely done was to mono all the bass - the last thing you need on vinyl is a load of LF on one channel only - that plays merry hell with everything in the record-replay chain.

I didn't mention that because I can't think of any classical LPs from any label where I can be certain that was ever done, while a great many pop/rock/Christian LPs I've seen over the years are all done that way.  Didn't they use a summing network below 200 Hz?  Makes a rather markedly different groove pattern to say the least!  Some of the early RCA stereo discs have some very disparate bass levels and out-of-phasiness which no doubt proved a challenge to those early stereo systems.

I'd rather have a 32ft bombarde on a recording any day of the week.

Word!  It doesn't depend on frequency response for its effect since it has no fundamental to speak of.  By the way, I've never heard or seen a true 64...  (only a resultant, which doesn't count).

Perhaps the displayed amplitude is some kind of average, but average of what over what?

I think that's exactly what has happened.  Andy's experiment shows something I've never seen but at times wished for in the analysis window.  This should become a feature request, because a switchable peak/average mode would be helpful but has never been implemented!  I don't know if I've overlooked this phenomenon in the past (possible but not likely) or if you've discovered a bug with that particular version of the software.  Some testing will need to be done!

The other thing that was routinely done was to mono all the bass - the last thing you need on vinyl is a load of LF on one channel only - that plays merry hell with everything in the record-replay chain.

I didn't mention that because I can't think of any classical LPs from any label where I can be certain that was ever done, while a great many pop/rock/Christian LPs I've seen over the years are all done that way.  Didn't they use a summing network below 200 Hz?  Makes a rather markedly different groove pattern to say the least!  Some of the early RCA stereo discs have some very disparate bass levels and out-of-phasiness which no doubt proved a challenge to those early stereo systems.

I think that it depends where you get your records from, to a degree, as to what frequency they sum at. The few that have been released with significant sounding out of balance components generally get  summed at some frequency though, I believe - you'd still percieve the harmonics in the correct position, and this would make the overall effect sound as though it was emanating from the correct place, even though not all of it actually was!

I'd rather have a 32ft bombarde on a recording any day of the week.

Word!  It doesn't depend on frequency response for its effect since it has no fundamental to speak of.  By the way, I've never heard or seen a true 64...  (only a resultant, which doesn't count).

Yes, it's only the speed at which the tongue is vibrating (and producing a 'clack') that constitutes any sort of a fundamental. I'm not sure that there's ever been a true 64ft open*; the majority are stopped 32ft pipes, and they don't speak very fast at all. There are also a few 64ft reeds around, but I've never heard one of those. Rather more common are synthesised 64ft stops (as in Washington Cathedral) - this is a lot cheaper to do, and probably sounds just as effective.

* The Atlantic City 64ft dulzian is a diaphone, and I don't think that this counts!

[Lionel Rogg]

Which is partly why I haven't bothered to convert my collection of Lionel Rogg playing Bach's entire organ output at all, even though it's in pristine condition.

Strange the way things go... completely unexpectedly, and out of the blue, I met the man briefly last night!