Why were the RPM speeds of records chosen

[terrybull]

Another thread sparked this thought. Was there any reason why 78, 45 and 33 (33 and a third I think) chosen. I realise it is all a compromise between playing time and frequency response but why those speeds and perhaps if 75 rpm etc had been used I would be asking the same question but just wondered if there is some reason for the specific reason for the final choice.

[Andrew2]

I have no idea, Terry. I do remember that in the 50's when we had an old wind-up gramophone, some of the older shellac records in our collection were marked '80 rpm', so it seems there may have been no agreed standard at the time those records were made. I must admit to never having thought about it, so maybe some research is required. Have you had a poke around on the web?

[avocollector]

I would guess at those being the lowest speeds at which acceptable frequency reproduction (at the time) could be achieved for realistic sales. However it's just a guess!

Don't forget there was also 16rpm for talking book records and similar.

[G4XWDJim]

And is it just a coincidence that 45 plus 33 equals 78.

Jim

[David G4EBT]

Peter Copeland, of the British Library National Sound Archive, London, gives a comprehensive explanation at ‘Vinylville’ at the link below.

Briefly, he states that records of 33 1/3 rpm were developed in conjunction with films. A 12-inch 78 with Berliner-type grooves could hold between 4 and 5 minutes per side. The first practical sound films produced in the US in the late 1920s had their sound on separate disc records and it was more important for the sound to be continuous. A reel of film might run for 11 minutes, so a rotational speed of about 32 rpm was required to make the sound match the picture.

Much more information here:

http://vinylville.tripod.com/spindoc.html


Here’s a clip from the link below which sounds plausible:

Clip 8-<

"The exact value of 78.26 rpm was not standardized till the 1920s. I believe that it was set according to the electric motor commonly available, which gave a speed of 3600 rpm, and a commonly available gear (46:1), so that it would fall in this range. As can be seen, 3600/46 = 78.26. The site further explains about 45 & 33 RPM. Not the done thing to copy chunks of text from internet sources, but suffice to say that this site also confirms Peter Copeland’s explanation that the development of ‘talkies’ necessitated longer playing time than a 78RPM record was capable of. 78 rpm offered very high quality, but the disc could hold only 4-5 minutes of music. On the other hand, a film reel lasted 11 minutes".

End clip 8-<

More information about motor speeds and ratios etc, and lots of other questions and answers that might spring to mind:

http://www.quora.com/How-were-the-ro...-78-rpm-chosen

Wiki also has lots of stuff:

http://en.wikipedia.org/wiki/Gramophone_record

Hope that's of interest.

[richrussell]

According to Wikipedia, 78rpm was arbitrarily chosen as a standard once record players started using synchronous motors so would have a well regulated fixed speed. Most early records were anywhere between 74rpm and 82rpm and you just altered your gramophone to make it sound right.

33⅓, 45 and 16⅔ were standards produced by different record companies, though no explanation is given for why they chose those speeds. I would guess that 45rpm was chosen as being easy to derive from a 60Hz supply maybe.

[Nicklyons2]

I believe it is total coincidence that 33 + 45 = 78. The speeds were chosen, as mentioned above, as a good compromise between frequency response, needle wear, and playing time. The constituent parts of (what became) EMI did, as separate companies, differ slightly in what the figure was e.g. Columbia Graphophone 80, HMV 78.

The reasons for 33.333 & 45 is entirely due to a Beta vs VHS type war between CBS and RCA in the USA. The rise in the plastics industry around WWII had meant cheaper, lighter and most importantly, unbreakable records, could now be made from vinyl.

RCA simply wanted to replace the 78 with 7" 45s so, for example Dvorak's New World Symphony would appear on 4 sides of 45s instead of 78s. Note here that the original RCA spec for 45s only had the large centre.

CBS looked further and thought that cutting technology had improved to a point where a slightly slower speed would work and that using a bigger disc - 12" an entire symphony could be accomodated on 1 disc. The need for 'singles' being accomodated with 'kiddietunes' size discs running also at 33.333. All CBS discs were to use the same 'std' centre as 78s.

The 'battle' raged for a couple of years where each company would issue its artistes only on thier own formats and with consumers having to buy 2 record players to be able to listen to everything OR to have to plump for your favourite artist, say Frank Sinatra (CBS) and then be able only to hear artistes from that 'stable'.

Peace broke out when it was seen that this was ruining the sales of record players and records and the compromise was that large records/long duration would use CBS' std if CBS would agree to issue 'singles' to RCA's spec.

[Radio Wrangler]

3600 and 1800 rpm are the speeds of 2-pole and 4-pole synchronous motors running from 60Hz

Here in the 50Hz world the speeds are 1500 and 3000 rpm.

I wonder if 78rpm happens to have realisable gear ratios to both?

David

[richrussell]

Originally Posted by Radio Wrangler

I wonder if 78rpm happens to have realisable gear ratios to both?

Yes indeed, for 60Hz a 46:1 ratio gives 78.26rpm, for 50Hz a 77:2 ratio gives 77.92rpm. That's certainly close enough.

[Nuvistor]

Just to add " confusion" Pathe had the 90 rpm hill and dale, start at the inner tack format.
Frank

[David G4EBT]

Originally Posted by richrussell

33⅓, 45 and 16⅔ were standards produced by different record companies, though no explanation is given for why they chose those speeds. I would guess that 45rpm was chosen as being easy to derive from a 60Hz supply maybe.

Peter Copeland, of the British Library National Sound Archive, London, at the link I gave in my earlier post, states:

Quote:

"The 45 rpm speed was the only one to be decided by a precise optimisation procedure (by RCA Victor in 1948). Calculus was used to show that the optimum use of a disc record of constant rotational speed occurs when the innermost recorded diameter is half the outermost recorded diameter. That's why a 7-inch single has a label 3 1/2 inches in diameter. Given the CBS vinyl groove dimensions and certain assumptions about the bandwidth and tolerable distortion, a speed of 45 rpm comes out of the formula".

Unquote.

[avocollector]

Fascinating Wikipedia article - plus it makes you wonder how much influence the physical limitations of the 78's had on popular music. For example most 'pop' songs are 3 mins or under in length which is the same limit the early 78's had.

I was also intrigued to find frequency response did not seem to be a consideration for deciding speed; apparently even 16rpm was used in early times for Jazz records and only later used for recorded speech. Instead motor speed seems to have been the dominant factor.

[avocollector]

Wonder why with the advent of talking films nothing similar (ie photo recorded audiotrack on celluloid) was created for sound alone. I would envisage something like a tape recorder with the tape going past a photoelectric read head? Would have thought that would have been a viable and cheap system even back in the 40's. Anyone know of any systems like that?

[Lucien Nunes]

In the early days of sound film there was such a thing, the sepopt (Separate optical) track. It was recorded by a device known as a 'sound camera' usually on academy 35mm. The problem with it as a stand-alone audio recording format was the cost - silver halide photochemistry has never been cheap and the cost of processing a roll of cine film was only tolerable given the large revenue a roll could generate. Fidelity was inherently limited by light valve and photocell performance and variability in photoprocessing results.

[emeritus]

As was noted in a recent thread, in the 1930's Radio Normandy used to record its English programmes on 35mm optical recorders in the UK for playback in France. I recall one of my (rather old) school physics books that must have been written in the 1940's before the LP record was invented, speculating on the possibility of film-based optical playback .

I heard on a radio programme a few years ago that in the 1940's, President Roosevelt ( I think), being tired of frequent mis-quotes in the press of what happened at briefing meetings, asked RCA if they could make him a sound recorder, which they did using an optical recorder which recorded on 16mm film. The recorder was apparently switched on at the start of a meeting, and then ran continuously until the film ran out. It thereby captured some rather frank comments made by the President after the meetings had ended, forgetting that the recorder was still running. It seems that the existence of the recordings only came to light following a chance remark by a researcher who had gone to the Library of Congress for a copy of Nixon's Watergate tapes [remember "expletive deleted" ?]. In jest he had asked the librarian for the Roosevelt tapes while he was about it, and was astonished to find that such things did in fact exist.

[Nicklyons2]

Originally Posted by avocollector

Anyone know of any systems like that?

Philips Miller system. Used extensively by the BBC, amongst others, in the pre WWII era. Basically it consisted of a 'tape' of clear celluloid coated with a layer of black gelatine. A sort of blunt wedge cutter attached to a moving armature driven by the audio cut into the gelatine layer, the depth dependant on the audio. With a wedge shaped cutter the area of the cut is greater for greater depth - therefore the cutter produced a 'variable area' optical track which could be replayed with the conventional cinema optical pickups of the time. The system had the advantage that it required no prcessing (unlike conventional comopt) and, as with magnetic tape. the recording could be rewound and replayed straight away.

[Nickthedentist]

Didn't we have a fairly comprehensive thread on this not long ago? Can anyone find it?

Nick.

[avocollector]

Cheers for that info people - I did look in the search but unsure what terms to use.

[kalee20]

Originally Posted by avocollector

Wonder why with the advent of talking films nothing similar (ie photo recorded audiotrack on celluloid) was created for sound alone... I would have thought that would have been a viable and cheap system even back in the 40's.

I do believe that Alan Blumlein's first experiments with stereophonic sound recording used optical methods on film. The only other viable method would have been disc recording. Perhaps he thought the sound quality was better, or perhaps it was just easier to have two separate tracks for the two channels, synchronised to within a few hundred microseconds, than would have been easily possible with disc.

[Synchrodyne]

Originally Posted by avocollector

Wonder why with the advent of talking films nothing similar (ie photo recorded audiotrack on celluloid) was created for sound alone. I would envisage something like a tape recorder with the tape going past a photoelectric read head? Would have thought that would have been a viable and cheap system even back in the 40's. Anyone know of any systems like that?

Perhaps in the 1940s and 1950s realization of an optical recording system that was suitable for domestic playback-only use and could have competed with the gramophone disc and tape in both convenience and quality would have been a difficult proposition.

Standards for optical audio recording would not necessarily have had to have any correlation with standards for optical tracks for movies, but the latter provide a starting point for some back-of-the-envelope analysis. In the 1970s, Super 8 mm film optical sound was reckoned to have a realizable upper frequency capability of around 5 kHz. Apparently this represented the shortest wavelength that could be achieved on prints. It was related to the linear speed of the film, which was 3.6 in/s at 24 frame/s. A wider soundtrack, as would likely be used for audio only, might provide a better signal-to-noise ratio (S/N) but it would not help to extend the frequency response. I don’t have any numbers for Super 8 S/N, but 16 mm was typically quoted as 40 dB for new prints, so Super 8 would likely have been somewhat worse.

A 7 inch reel was probably around the size that would have been seen as suitable for domestic use. Such a reel contained 400 ft of Super 8 film, giving a running time of a bit over 20 minutes at 3.6 in/s. Apparently there were available in the 1970s Bell & Howell cassettes that carried 400 ft of Super 8 film, and for example these were used in the Nordmende CCS telecine unit. So one may see that even in the 1970s it would have been a difficult proposition to develop a workable domestic format with an upper frequency limit increased by say one-and-a-half octaves. A combination of higher linear film speed and better printing optics (to allow shorter wavelengths) would have helped, but the higher linear speed would have required thinner film to keep the reel size within the desired boundary. Then add in some form of noise reduction, say Dolby B. Perhaps it was doable with heroic efforts, but the result would probably not have been all that competitive with the LP in terms of audio quality, or with the compact cassette in terms of convenience. Looking back from there to the 1940s and 1950s, the quality relativity with the LP would likely have been worse, and whilst the compact cassette was not then a comparative factor, how convenient the optical record would have been is open to doubt.

Returning to disc record rotational speeds, I did find some useful information on the 16⅔ rev/min speed in Popular Electronics 1957 August, p.63ff, available at: http://www.americanradiohistory.com/...cs-1978-08.pdf.

“Compatible” 16⅔ rev/min discs, that is those that conformed to the microgroove standard and could be played with a 0.001 inch stylus, were said to have an upper frequency limit of around 9 kHz. Whether that was ever extended during the production life of 16⅔ rev/min discs is unknown. It is tempting to think that given that the LP was extended to 45 kHz for CD-4 recordings, then 20 kHz might have been reachable for the 16⅔ rev/min disc. But I am not sure that the extrapolation is valid. In the CD-4 case, the subcarrier was recorded at something like 19 dB down relative to the main channel. Being an FM-PM subcarrier, it probably had some immunity to amplitude disturbances occasioned by harmonic distortions from the main channel that reached into the subcarrier band. And as special styli were required to track the very short wavelengths in the subcarrier band (20 kHz and above) on CD-4 recordings, then one may surmise that the same would be required to track frequencies above about 10 kHz on a 16⅔ Hz recording. So perhaps the original 9 kHz upper limit was not far off the best that could reasonably be expected. And it was evidently seen as adequate for the talking book application. In that regard, I guess that as some kind of threshold, one may invoke the 7 kHz number that is modern thinking on the bandwidth required for voice teleconference circuits.

Cheers,