Frequency content in LPs (vinyl)

Here I investigated the output from my RIAA stage when playing music and also a swept sinustone from a Cardas test record (at 45 RPM, covering > 30 kHz). Cartridge is a Shure V15VxMR.

First I show a sonogram from a Cowboy Junkies Sweet Jane (mastered with DAT originally, sampling frequency unknown) recorded at 44.1 kHz (16 bit).

The x axis shows frequency content (linear scale) and the y axis the time played. The broad pattern in the center is an sss-sound with peaking (yellow) around 5 kHz, and the sound appears to go further above 22 kHz and is clearly over background (black).

The next diagram shows another song (an old, used record with Suzanne Vega, original mastering not known) recorded at 48 kHz.

Most energy stops somewhere around 20 kHz, but still there are traces that appear to go further above 24 kHz.

So how about overtones generated by the cartridge?

Here is a swep sinustone from Cardas test record, player at 45 RPM and recorded at 48 kHz (the record is supposedly mastered at –14 dB, however, around 10 kHz or so the sweep appears to stop and start over at a lower level. Why, I have no clue. The information on the record cover says nothing about this). It was mastered using a Neuman VMS 66 lathe with a helium cooled Neumann SX-74 cutter head. The pressing however, is in my ears not very high quality and the record is warped.

A lot of overtones can be observed. Also note the trace that goes the “other way”. Higher up in frequency it looks like this:

The response goes up to 24 kHz then there is a distorsion in the A/D that makes the signal “bounce back”. It appears to diminish around 27-33 kHz. The wiggly appearance may be due to the warped record I used.

How about the levels of the overtones? I measured the following using a 1.7 kHz fundamental tone (same sweep):

1^st, 1.7 khz - 18 dB

2^nd, 3.4 - 67 dB

3^rd, 5.0 - 76.3

4^th, 6.7 - 79.4

5^th, 8.4 - 85.7

6^th, 10 - 83.6

7^th, 11.8 - 84.3

So the 3.4 kHz overtone was around 50 dB lower than the fundamental.

Measuring at 9 kHz fundamental tone I found

9 kHz - 17 dB

18 kHz - 52 dB

Thus the overtone was 35 dB lower than the fundamental. The output level at 20 kHz on the Cardas record was – 29 dB (about 10 dB lower than the lower frequencies) i.e. 23 dB higher than the overtone at 18 khz generated from the 9 kHz fundamental.

So what is the conclusion? Well, vinyl can reproduce high frequency tones and there are signals in music albums that are higher than 22 kHz and which are present above background level. The nature of these signals is however not known. If there was a record produced from a digital mastering at 44.1 kHz I could aswer the question whether the traces are due to overtones created by the cartridge or actual music information. The dynamic capacity above 10 kHz is however much limited, especially in the presence of distorsion harmonics from lower frequencies.

Thats all,

Thomas