Secrets Q & A
- Written by John E. Johnson, Jr.
- Published on 12 August 2008
Part 4: The RIAA Curve
When commercial phonograph recordings were being produced in the early twentieth century - on wax cylinders at that time - they realized that the grooves for low frequencies were so wide, they took up a lot of space and reduced the recording time. And, as the frequency response technology improved, they started worrying about the hum and hiss from the electronics, as well as surface noise from initial plastics (which were not very good).
Then they got a bright idea. They realized that all of these problems could be addressed by using "Pre-emphasis" in the recording, which means altering the recording level of different regions of the audible frequency band, i.e., EQ.
However, no standards for doing this had been set up, so various recording companies added the EQ in whatever way they felt was best for them. Of course, this meant that playing back a recording could produce all kinds of tonalities, because the playback mechanism had its own tonality, engraved in stone.
Amazingly, it was not until 1954 that such a standard was realized, called the RIAA curve (Recording Industry Association of America). It took into account the need for increasing the recording time by reducing the recorded level of deep frequencies, and reducing surface noise and hiss by emphasizing the recording level of high frequencies, and then applying the inverse RIAA curve during playback. They also decided to manipulate the region between 1 kHz and 10 kHz to bring forward the sibilance of the music and voices.
So, there was an RIAA pre-emphasic curve and an RIAA de-emphasis curve.
Here is the pre-emphasis curve which is applied in the electronics during the process of cutting the master disc. Setting 1 kHz to 0 dB, you can see that at 10 Hz, the pre-emphasis is basically a 20 dB attenuation, and at 20 kHz, there is a 20 dB boost. Bottom line: 40 dB variance between the EQ for low frequencies vs. high frequencies.
You can see that it is obviously not a straight line. It took into account everyone's wishes to reduce this or emphasize that. The curve actually goes out to 50 kHz, which says something about the wide bandwidth that LPs have in comparison to CDs.
So, when you play the LP at home, the de-emphasis curve is built into the phono stage in your preamplifier or receiver. Here is what it looks like:
It is the opposite of the pre-emphasis curve, and the idea is to make the final frequency response flat, but get rid of hum, hiss, surface noise, and extend the recording time that will fit on the LP disc.
It works well, but like all "standards" engineers were always thinking of ways to improve it. Something called the IEC curve arose in the 1970's, and the de-emphasis curve is shown below.
The biggest difference seems to be in the 50 Hz region.
The IEC curve did not take off, and maybe that is fortunate, because then what would we do with all the other recordings we had that were EQ'd with the RIAA curve?
Application of RIAA to the Phono Stage
Applying the RIAA curve in a phono stage is not as easy as you might imagine (on the other hand, maybe you imagine it is hard). It's not just a crossover point. It is a continued scale of EQ all the way through the audible band.
There are basically two ways to do this. One is the passive method, using a variety of filters. The other way is the active method, which uses op amps and feedback in different amounts throughout the audible band.
The purists will say that the active method is not good, because it uses all those op amps that have voltage limitations and negative feedback which causes lots of distortion.
But, with either method, there is one big issue, namely, phase shift. Remember the discussion of phase shift problems that you have at the crossover point in speaker crossover networks. Well, since the response in the RIAA curve is basically one big alteration throughout, you have phase shift all over the place. It is defined by a complex set of mathematical equations.
Now, when the RIAA pre-emphasis curve is applied at recording, and then the de-emphasis curve is applied at playback, the phase shift should cancel out and you end up with a proper phase relationship throughout the audible spectrum. Right? Well, that would be fine if all the recording systems applied the RIAA pre-emphasis curve perfectly, and all our phono stages applied the RIAA de-emphasis curve perfectly.
Guess what? Nothing is perfect out there.
So, we end up with LPs and phono stages that deliver a sound with varying amounts of phase shift in all areas of the audible spectrum, and I feel that this is one of the defining characteristics of the analog LP sound. It delivers a soundstage that is much different than what you would hear from a CD where such EQ curves are not applied.
And, it's very appealing.