Better Equalisation for 78 rpm records.

 

A simple introduction for beginners – like me. 8^)

 

 

 

 

N.B. This is only intended as the merest introduction to the subject, and everything has been greatly simplified. I have probably got some stuff wrong, too! Links to more advanced books and articles are given at the end of the page. 

 

What is equalisation on disc records? What is it for? Do we really need to bother about it when playing our 78 rpm records?

 

Equalisation on a disc record is a deliberate change made in the relative strengths of some, or even all, the audio frequencies being put onto the disc.

 

The purpose of doing this, is to improve the efficiency – and hence the quality – when the recording is played back.

 

The equalisation used on LP discs is a very good example. On an LP, the grooves need to be very close together, so that the record will play for a long time. But if we have heavy bass notes, the cutting stylus will travel widely, and may break into neighbouring grooves. Result: disaster! By contrast, very high notes only require a small excursion by the cutting stylus, but at a high frequency. There’s no risk of cutting into the neighbouring grooves, but it’s difficult to play back those very tiny details because they are so small. To overcome both these problems in one go, it was decided to reduce the magnitude of the bass cut into the master record, and to increase the amount of treble. Then, the bass won’t cause grooves to ‘crash’, and the tiny treble oscillations are much bigger & can be played back easily. Here are some simplified diagrams.

 

graph 1

The pale blue line is the nominal, basic level of all the various frequencies shown on the lower axis, going from lowest to highest. But as we have seen, if we cut low bass at this level, the grooves might crash. So we’ll reduce the bass. Let’s make it begin at -20, and let it return to normal say halfway through the frequency range. We’ll leave the higher frequencies as they were before.

 

graph 2

Now the low bass grooves will be OK – they won’t have such a large lateral excursion. And as for the middle frequencies, well, they aren’t bass notes any more, but middle ones. When we play our LP back, we’ll just increase the bass emphasis in the amplifier and we’ll get back all our nice original bass. But we still need to enhance the higher notes, otherwise the tiny little wiggles of the highest frequencies will be really hard for the playback stylus to trace accurately. So:

 

graph 3

We’ll do more of the same. We progressively increase the strength of the higher notes, so that they will make the cutting stylus move more. Then, the playback stylus will be able to trace them more easily. And when we play back the LP, we’ll design our amplifier so that it decreases the high notes back to where they were before, and increases the bass notes back to where they were before. The red line in the above graph is the recording equalisation we have applied to the original frequencies. To get it back to normal when we play the LP, we’ll apply an equal & opposite playback equalisation:

 

graph 4

The red line is our recording equalisation, and the green line is our playback equalisation. They are equal and opposite, so the bass is ‘pulled up’ and the treble is ‘pushed down’ to where they both originally were – the blue horizontal line. Bingo! It’s a really great idea, and I thought of patenting it. Only, it was worked out around 1950 by the Recording Industry Association of America, and put forth as a good way of making the then-new LP records sound better. The RIAA (for short) hoped it would be adopted world-wide, and it sort-of was, but with all sorts of twiddles & variations – but that’s not important right now.

 

What is important, is that I now understand about disc record ‘equalisation’ – or at least I think so, and that’s why I’ve written this page – to see if I can explain it without tripping myself up. I have deliberately simplified everything; the RIAA EQ (short for equalisation) actually has a flat bit in the middle between 500 Hz and 2.1 KHz, where the frequencies are left unaltered, but we just wanted to illustrate the principle of the thing, and besides, straight lines are really easy to draw in Photoshop! 8^)    So there we have it. You have a recording EQ, and then a playback EQ which is exactly opposite; so you end up with what you started out with, only the integrity – the quality – of the sound is better.

 

But why am I rambling on about LPs, when this page is supposed to be about 78 rpm discs?

 

Well, now. A great many normal 78 enthusiasts (you’ll grant it that we are normal, won’t you?  8^)  ) including me, use modern amplifiers. At least, fairly modern amplifiers. I use NAD 3020s, which although they are old now, still work fine. They’re good, solid ‘lower-mid-fi’ units, very reliable. The only problem with them, as well as many, many other middle of the road amplifiers, is that they were only ever intended to play LPs. The NAD 3020 was introduced around 1980, when 78s had been dead & buried for over 20 years. Therefore, no provision was made for playing them. The NAD & all other such amplifiers, quite naturally, have embodied in them, the RIAA LP playback equalisation – the green line in the graph above. This lifts bass and reduces treble, as already described.

 

Q. So what happens when we play back a 78 through such an amplifier?     

 

A. It all depends on what type or make and how old the 78 is. Let’s begin with the earliest 78s.   

 

You’ll be glad to hear that, up to the mid-1920s, there was no EQ on 78s. This was the era of mechanical (so-called ‘acoustic’) recording. It held sway for a very long time. The core date for Berliner’s disc ‘Gramophone’ was 1888; and yet, in late 1924 exactly the same mechanical system was in use. Thirty-six years was a very, very long time for a technology to persist in those days! The answer I like to give, is that it was consistently developed & made to work better and better, so that it could produce pretty good results, and looked set to stay around even longer. As you know, the sound was gathered in via a conical horn, which vibrated a diaphragm at the narrow end of it. A point attached to the diaphragm cut a replica of the vibrations in a rotating disc of wax. Primitive it may have been, but it worked. And, as remarked, it got improved a very great deal. In 1888, you were probably lucky to catch a few words of speech – very early Berliner records of nursery rhymes actually had the words printed out & stuck on the back of the single sided disc. But in a few years, things were already much better. Listen to this short extract of a polka played by Iff’s Orchestra, recorded in Glasgow, Scotland, in 1899. (Iff’s band used to play for Queen Victoria at Balmoral, her Scottish residence.) Right-click on the link & select ‘open in new tab’.

 

iff.mp3

 

The sound is pretty fugitive, but not bad for 116 years ago. 8^)

 

As I said, recording improved very quickly. Here’s a bit of the prelude to Gounod’s opera ‘Faust’, recorded in Paris, ~1912. Right click, new tab.

 

faust.mp3

 

Gosh! I can hardly stand the excitement, even though the musicians were playing 103 years ago.

 

The ironic thing is, that when I made these transfers, which I thought sounded fairly plausible, they were completely wrong. I played the discs into the NAD amp with its LP EQ. So the bass was boosted, and the treble attenuated. They should have been transferred ‘flat’ – that is, like our original diagram, because there is NO EQ on mechanical recordings.

 

flat line

 

What we have ended up with, is a complete Pig’s Ear – see below.

 

graph 5a

Yes. We should have started with a flat line. Then, we could, later, have tweaked the various frequencies to our heart’s content & see what we could come up with. But instead, we were foredoomed. Consider the zone X. These are low frequencies, which are poorly recorded under the mechanical system; it had a very limited frequency range, often stated to be about 150 to 2000 Hz. There really wasn’t much bass response at all. Yet, we have seen fit to boost it by a great amount – the green line, you recall, is the RIAA replay EQ. This has increased the rumble, waffle & all the other gunge to be found down on the lower end of mechanical recordings, especially those of Pathé, who recorded that Gounod ‘Faust’. Their discs were dubbed, pantographically, from very large master cylinder records. A great deal of spurious noise came across during that process, which was not much noticed at the time, on early gramophones. But it certainly shows up well on modern equipment! Perhaps worse still, in zone Y, we have the upper frequencies, such as they are. Though meagre, they give the music its presence, its immediacy, its impact. So what have we done? We have attenuated them! We have thrown the baby out with the bath-water. Grrh!

 

No amount of subsequent tinkering is likely to exactly restore the status quo. Even if we eventually somehow did so, would it not have been vastly better to start with no EQ at all; to begin on a level playing field? Of course it would. To add to the irony, neither of these discs remains in my possession, so I can’t go back and do them again, properly. Heigh ho. Still, learning by one’s mistakes may be inefficient; but it is a very effective way of learning…

 

I think we can conclude emphatically, that to transfer a mechanical recording with the RIAA LP playback EQ is a very bad place to start. We need to be able to play mechanical recordings ‘flat’, that is, without any equalisation. Then and only then, should we begin to try out our own ideas on how we may improve the sound. We urgently need a device to plug into our amplifier which allows us to record 78s without having the undesired LP EQ applied to the .wav file. Oddly enough, the little red & black box shown at the top of this page, will enable us to do so. But before we get into it, I’m afraid we have to have more waffle. Sorry and all that…

 

As we have said, the mechanical recording system had got pretty good by the mid-1920s, and might well have continued longer. Except that, since 1920, a New Thing had come along, which rapidly began to undermine it.

 

The New Thing was called radio. Radio – or wireless – had been around in practicable form since the 1890s. Like sound reproduction, it got better very quickly. During the catastrophe of the Great War (1914-1918), radio had become quite capable of transmitting speech, and indeed music, instead of just Morse code, as was originally the case. Radio broadcasting as public entertainment, started in the U.S.A. in 1920. (It was later over here in the U.K. – end of 1922, start of 1923.) To cut a long story short, radio worked totally by electricity, and pretty soon, radio sets multiplied & waxed exceedingly in American homes, even though it was an expensive game. New technologies usually are! Listening in on headphones (women hated them, of course, as they still do) soon gave way to loudspeakers. And it was noticed, as the horns on the loudspeakers got bigger, and the drive units got better, that the sound of speech and music on the radio was often superior to that to be got from most gramophones.

 

One would have thought that the main record companies in the U.S.A., would have been preparing to address the inherent obsolescence of their mechanical recording system. I’m sure they were; but the great breakthrough came from an unexpected direction. Western Electric, a subsidiary of American Telephone and Telegraph, had been quietly working on an electrical system of disc recording for several years. It was, of course, not done as an act of philanthropy. It was done to make money. 8^)  

 

And it was during the development of this Western Electric system, that it first became obvious that recording true bass notes would require an excessive groove swing, as we discussed above. Therefore, they decided to reduce the bass on recording, and put it back on replay. This, you are already familiar with. As for the treble notes, they decided to leave them as they were. So the Western Electric equalisation came into existence. It was rather like this:

 

graph6

The Western Electric system (WE for short) was indeed vastly superior to the best mechanical systems, and was eagerly snapped up by Victor and Columbia, and came in in the Spring of 1925, and remained in use for quite a few years – in this country, at least until 1933. Without going through a lot more of my waffle, it’s clear that we mustn’t transfer a Western Electric recording using the RIAA LP EQ on our modern amplifiers. The bass notes will come out vaguely OK, but the treble will be greatly reduced. So our projected ‘red & black box’ will need a setting to put back the bass, but leave the treble flat.

 

WE didn’t sell their system & its equipment. They licenced it, just like computer software sellers do today. What’s that? Did you think you had ­purchased some software for your PC and that you owned it? Great Heavens, no! You only have purchased a licence to use it. Still, you don’t have to pay an additional fee every time you use the software. But the major record companies in the later 1920s did. They had to pay WE a half-penny for every copy of every side they put out using WE. For years the money came rolling it to WE. But in October 1929 came the Wall Street Crash. Money got very tight; and in the U.K., Columbia thought to invent a new electric recording system of their own, outside the WE patents, so they could save money. Columbia’s brilliant young scientist Alan Blumlein came up with such a system. It has an EQ that rather resembled the WE in that it reduced the bass, and also left the treble flat. However, it varied in that the frequency – the ‘turn-over frequency’ – at which the bass cut was applied, was slightly different. Therefore our red & black box will have to give us a choice of turn-over frequency. Worse still, since all these systems were patented, they were regarded as Trade Secrets; so exact figures were sometimes made deliberately vague. The turnover frequency might be 200 Hz. Or perhaps 250, or 300 Hz. Figures of 350, 400 and 500 Hz have also been suggested and were apparently used, though at different times – the Blumlein system remained in use for a long time.

 

Do we despair? Not at all. A single capacitor and a single resistor, costing perhaps 10 or 15 pence is often all we need to establish a turn-over frequency. We could even have a 12-way rotary switch that would enable us to have a ‘flat’ setting for mechanical recordings, and eleven different turnover frequencies, at the expense of three or four quid. Yet for our prototype, we settled for just three choices: 200 Hz, 400 Hz and 500 Hz. This was because we copied a circuit to be found on-line. We can’t reproduce the circuit, because the site forbids it. However, I don’t think the copyright can extend to prevent someone from building a circuit that has been published on-line? Anyway, I did knock up a rough prototype, which worked OK.

 

This used two small and cheap op-amps, and I am not very much at home among that sort of stuff. We made it up on Veroboard, with which we have always had a love-hate relationship. It took three attempts before we got it right. 8^) Even then, it wasn’t really right, because we misinterpreted the circuit diagram, which showed only one channel; the other channel was built around the other half of each IC. So we used four ICs instead of two.  <8^) . But it did work!

fair protoype

 

There no less that eighteen points at which connections need to go to the board, and there are two outboard switches which carry only capacitors of standard value. The first switch copes with the low frequencies, the other with the high. So there is a plethora of wires in the finished box, especially as we included a gain control. The op-amps turned out to be very perky with our Shure SC35 cartridges, and we also have the odd M-44, which has a much higher output & would probably overload the pre-amp. In any case, a gain control is useful anyway. N.B. This was also wrong; we put the gain control before the pre-amp board – it should go after, on the line output level.

 

pre amp open top

 

Of course, when the lid is on, it’s a case of ‘out of sight, out of mind’. 8^)   It requires a dual 15 Volt supply. The three twisted wires coming in through the hole in the back bring in the power. This was temporary; later two phono sockets were fitted.   

 

The main thing was that it worked; and we are currently trying out various different settings. The low end switch gives us a flat response, then turnovers at 200, 400 and 500 Hz. The high end switch gives flat, or 1.6, 2.1, 3.4 and 5.6 KHz turnovers. So far we are concentrating on mechanical, Western Electric and Blumlein recordings.

 

We will not leave you without short examples of the device in action.

 

Had you gone into the tea-room of Bobby’s Department store in Folkestone, Kent, in 1934, you would have been regaled by the sounds of Bobby’s Folkestone Orchestra – a string quintet plus piano. Here is 35 seconds of them from a private HMV disc, played with the wrong (RIAA) EQ. Right-click & open in new tab. (You really need to be listening to the samples in quite high quality to hear the differences.)

 

Wrong RIAA EQ

 

Now we have it with its intended Blumlein EQ.

 

Correct Blumlein EQ

 

Naturally, the enhanced treble response has brought out the much-dreaded ‘English Crackle’ * ! But no matter; it makes it much easier for the de-clicking & de-crackling software to ‘see’ it & get rid of it. So we end up with:

 

Correct Blumlein EQ, de-clicked & de-crackled.

 

No digital de-noising on any of these samples – I hate it!

 

Encouraged by this success, we obtained the proper printed circuit boards (superb quality) from Elliott Sound Products in Australia. They have scores of very useful projects on their site : http://sound.whsites.net/index2.html . This pre-amp is their Project No. 91; they are created by Rod Elliott, who is a brilliant and widely experienced electronic designer. This resulted in our Mk. II pre-amp:

 

 

 

Note how small and beautiful the proper PCB is! We used a bigger box to accommodate the rotary switches & capacitors. This one has more settings. You can of course easily add or change different capacitors for different turnovers &c. We added a little innovation of our own – the scruffy little board on the right contains a single 2N3904 transistor, configured as a ‘unity gain’ stage. The red switch on the right side of the box enables this to be switched in or out of one channel. When switched in, it causes a 180° phase shift on that channel; and when you put the two channels into mono. (with the switch on the right front), you may play vertically-cut records with a much cleaner sound than just using ordinary stereo. Oh, by the way: I made this before I realized that gain controls go after the pre-amp board, not before it; so the twin gang pot on the left has been disconnected for the time being.

 

Finally, be warned that messing about with phono pre-amps with 78 rpm EQs is dangerously addictive; because since September 2016, we have made two more of them – both prototypes – but this time using valves. (Vacuum tubes.) Each has its own nearby web page:

 

6SL7 & 6SN7 prototype: valvepreamp1.htm

 

ECC83 – ECC82 – ECC82 prototytpe: valvepreamp2.htm

 

If you are still reading this page, congratulations! This is nearly the end of it. And of course, there is no obligation whatever to play 78s or anything else with any specific EQ, or any at all. I lived the past many decades playing records ‘wrongly’ and they always sounded fine to me. Until now. 8^)

 

Postscript: Quite often, PC software includes a Reverse RIAA curve setting; or, you could construct your own and save it as a preset. Of course, I tried it a number of times; but found the results rather shrill. I understand this is because moving magnet cartridges – which is what most of us use – have an upper frequency response which goes some little way towards being a Reverse RIAA curve. If this is indeed the case, it would explain why imposing a full Reverse RIAA sounds to me a bit tinny, because partial compensation has already taken place. But in any case, it is surely deplorable to play a 78 with the wrong EQ and then to employ another curve to get back to where we should have been in the first place? The music will have been forced through two extra and unnecessary processes. It should always be our aim to get to the finished product with the least number of transductions, for the want of a better word.

 

 

Don’t forget to check out Elliott Sound Products! http://sound.whsites.net/index2.html

 

 

And if you want to learn lots more to do with all aspects of transfer, conservation & restoration of audio media, the late Peter Copeland’s 1990 manual is what you need – it’s a free downloadable pdf at the British Library. It’s 330 pages can be daunting, but it’s all there. May I suggest that you go down the index until you come to a topic that interests you, and start with that?

 

http://www.bl.uk/britishlibrary/~/media/subjects%20images/sound/analoguesoundrestoration.pdf

 

 

 

 

 

 

 

 

 

 

 

Page finished 23rd March 2015.

Reformatted 17th December 2015.

Modified & updated 30th November 2016.