Why was sound quality so poor on wax cylinders?

Wax cylinder recordings had a lot of high-frequency noise ‘static’ . I think, even when new. Am I right? If so, why?

Low signal to noise ratio. There was really no way for you to overcome the fact that the horn would transfer everything it “heard” to the stylus cutting the cylinder, apart from speaking as loudly as possible. Think of it like trying to speak to someone in a noisy bar or club.

I’m not sure that’s true. Surely the recording would be made in the quietest possible place to avoid extraneous noise.

I had always assumed that it was that the ‘groove’ on a wax cylinder would deteriorate very quickly. I remember 78rpm records played with a steel needle. When the record and the needle were new, the sound quality was pretty good; once it had been played a few times, and the needles were being re-used because the new ones had run out, it began to sound scratchy. Shellac is a lot harder than the wax on a cylinder too.

I’m with Bob++ on this one. Of course recordings were made in as quiet a place as possible.

Another factor is the noise generated by the needle sliding along the groove, and this depends on the material the groove is cut in. I have no data on wax cylinders but shellac 78s created a lot of noise. Groove noise was much reduced when vinyl microgroove LPs were introduced around 1958.

Also, cutting the groove was not too linear a process so there was plenty of distortion.

And again- the conical horns used for playback then had a poor frequency response.

In the early days there was no good reproduction method of cylinders. A performer played in front of a group of recorders. Often there would be a rubber tube going from one to another carrying the sound. This really affected quality.

The main advantage of discs is easy stamping out of copies from a master.

Also, cylinders grooves were cut vertically- up and down. This type of tracking had a lot of issues which is why discs are cut laterally, side to side. Even on a “silent” section with vertical cut you get noise due to unevenness of the tracking.

I’d be interested in a cite for that. (The rubber tube and multiple recorder bit)

Yes, that’s a disabling problem with cylinders. Also, you can store discs far more compactly.

No. Discs have the two sides of the groove cut differently, to give stereo.

Tracking problems are something else again, and refer to the needle rattling about and not following the groove.

Cite: End of this section.

Stereo? Who is talking about stereo???

I’m not talking about tracking problems in and of themselves. But a consequence of the tracking and the cutting method.

From Wikipedia: https://en.wikipedia.org/wiki/Edison_Disc_Record

"Edison had previously made only phonograph cylinders but decided to add a disc format to the product line because of the increasingly dominant market share of the shellac disc records (later called 78s because of their typical rotational speed in revolutions per minute) made by competitors such as the Victor Talking Machine Company. Victor and most other makers recorded and played sound by a lateral or side-to-side motion of the stylus in the record groove, while in the Edison system the motion was vertical or up-and-down, known as vertical recording, as used for cylinder records. An Edison Disc Phonograph is distinguished by the diaphragm of the reproducer being parallel to the surface of the record. The diaphragm of a reproducer used for playing lateral records is at a right angle to the surface. "

A friend of mine has an old Edison phonograph. He also has an adapter that changes out the Edison up-and-down motion diaphragm for a Victrola-style sideways motion diaphragm. **ftg **is right about the Edison records being cut vertically.

First, the wax itself (actually a metallic soap) isn’t perfectly smooth. It has a structure that imparts noise.

Second, there’s noise from the recording mechanism. There were no electronics involved; there was a big horn leading to a diaphragm with a cutting needle attached, there was a mechanical motor (often powered by a falling weight) driving the cylinder, and there was another mechanism to drive the needle across the cylinder as it turned. All of this mechanical stuff introduced noise (and other types of distortion).

Playback mechanisms had the same problems: a bunch of mechanical stuff that would introduce noise.

Because there was no electrical amplification, the strength of the signal wasn’t high compared to the noise of the recording medium.

There were the problems of making copies of wax cylinders, which have already been discussed. Reproduction was easier when they switched to celluloid, because the material would shrink slightly when it solidified, making it possible to remove it from a mold.

Flat records were introduced by Emile Berliner in 1894. Mass-produced stereo records for home consumers were introduced in 1957. For over 60 years flat records were mono.

Can someone comment on what in the mechanical chain (if any one can be identified among the mess) is most responsible for the “high frequency” noise?

I’m sure finding the identity of frequency in noise (while, red, etc) is wrongheaded, but I might as well learn about that, especially became “high” is mentioned in OP.

The noise floor sounds bright in wax cylinders not due to the source but due to the limitations of reproduction. Note that Edison was cut up and down and Victor was side to side so the perceived noise will differ.

The brightness comes from fall-off in volume in upper and lower frequencies is quite large and mechanical players enhanced this effect. The human ear also has a peak around 2–4 kHz which will sound louder which enhances this effect before mixes started to compensate for it.

For reference:

2Khz sine wave

4Khz sine wave

Even with the much later LP’s with electrical amplification that claim 20 Hz to 20,000 Hz reproduction have huge fall offs on each side and also have to limit dynamic range to avoid skipping, which is part of the reason many records sound better BTW because they cannot be as compressed as digital recordings are today.

While each part will color this the noise floor, or the sum of all noise sources is bright because of the limited reproduction bandwidth and ear sensitiveness. So it is really not related to the 1⁄f pink noise or red f^2 noise but really approximates the form of white noise that the device can reproduce.

Note the frequency range that is reproduced here.


On a semi-sort-of related note, if you read about the equalization method used with electronic methods starting c1940* you can learn a bit about the limits of mechanical recording methods.

It’s sort of astonishing that people actually enjoyed listening to these things back then given all the noise problems.

  • Which was still years before HiFi and stereo.

That sounds wholly contradictory to me. In fact you have to be very careful with recording levels on vinyl, controlling cutter velocity, cutter acceleration, the amount of out-of-phase amplitude etc. There is no such need with CDs; they can be and used to be recorded with no compression at all. You can record DC if you want to include the barometric pressure.

Then came the loudness wars:

You can compress CDs but there is absolutely no incentive to do so. With vinyl there is.

Here is a location where you can find nearly empirical data to support my claims.


There is no technical reason for this on CD’s outside of fashion trends. One common reason is that bumping up the overall volume is a cheap and dirty way to “remaster” CD’s and have another bump in sales.

Here is an example.

And yes it is counter intuitive that vinyl versions would have more DR as the medium should have less but typically vinyl buyers are more concerned with audio quality and with a CD you can compress the average DR to 5db and it will be loud while doing the same on vinyl will just make it quieter.

If you look at that earlier loudness-war site and order by year you will see just how little dynamic range is on typical releases today. To be fair some of that is due to the current fad of ducking and side chaining too.
If you look when CD’s were new the mixes were similar

And even bands which have critical control and require the dynamic range are still mixed a bit hotter.

It is well known when doing A/B testing where quality is more critical like sound production as in instruments or guitar amps that the louder of two will sound “better” in the short term. Over extended exposure that initial “loud is better” wears off for individuals who prefer or place value other aspects but it is pretty universal for a bit of time. MP3’s or other streaming formats also sometimes suffered because people wanted louder headphones and their volume was restricted for health concerns or battery life etc…

I want to be clear I am not saying it is objectively better, but I do prefer more dynamic range and have to dig through old CDs and cannot buy from online stores as they often will not offer the original mix for sale. I am actually envious of those who can overlook the effects of the loudness wars.

I messed up the link for the more recent releases and it is sorted by low DR but not how many pages you have to go through from 2018 before the average DR is a reasonable level.

Note that standards like the ITU-R BS.1770 are starting to help here a bit as DAW meters start to become available and regulations are put in place but there is still a tendency for streaming providers to one up each other to sound “better” in simple short term head to head comparisons.

Here is a link that will explain that complexity.


By asking about “high” I pretty much excluded discussion of Wow, Flutter or Wander – which are also present when listening to wax cylinders – and we landed about where I was interested :slight_smile:

As a Vinyl Collector/DJ/Pre-Digital Olde Farte, I’ve always had the opposite question: HOW can we get so much fidelity out of grooves in a plastic plate? How can just shoving a needle back and forth and up and down capture such subtle nuances of sound?

That was my first thought; glad to see that there was some basis to it.

This is simply answered. You can’t. Wiggling a needle in a groove is never going to be as good as 16-bit digital PCM. It is only usable because our audio perception is very forgiving.

What I dislike most is the flutter induced by the needle scrubbing back & forth as the arm nods up and down. This is usually around 4 Hz, close to where we are most sensitive to flutter.

I’d be very interested to hear more about this.