Sonata and Temo

One of the vals tandas included in my book, Tango 500, opens with a song called Sonata sung by its composer Augstín Magaldi. The tanda continues with Sin rumbo fijo (with Ángel Vargas) and ends with the wonderful Temo (with Mario Corrales). The orchestra is Orquesta Típica Victor under the direction of Federico Scorticati – or is it?

Well, as far as Sonata is concerned, no, it probably isn’t. Tango.info attributes it to Orquesta Típica Victor, giving the recording date as 3 June 1937, but the song is missing from other OTV discographies. The original shellac disc label says solo con orquesta, but doesn’t name the ensemble. The A-side of the same disc was also a Magaldi recording, but that’s with guitar accompaniment.

OTV was a recording orchestra without fixed personnel. It had a studio session on 26 May and another on 8 July, each producing two recordings, but nothing else was recorded on 3 June. The orchestra usually recorded with ten to twelve players, but the ensemble from 3 June sounds smaller. It seems likely that a handful of players were hired for the session, drawn from the pool of OTV regulars. If anyone has further information, I’d be interested, but it’s a nice song, anyway. Here’s an excerpt:

Tuning

Many tango recordings are published at the wrong pitch/speed. All three of the songs in this tanda require retuning to return them to concert pitch. I analysed the tuning of my transfer of Sonata using the nnls-chroma tuning plugin with the Audacity sound editor and found A = 432.8 Hz (and I am assuming that A = 435 would still have been in use at the time of recording). With some rounding of the values, the published pitch is 9¢ (0.5%) flat, a small (but audible) variance. Correcting it shaves one second off the song duration, and most people wouldn’t notice it – the variance is only just above the threshold for hearing a pitch difference.

I also checked Sin rumbo fijo and found A = 446.9. That is a much more significant 47¢ (2.6%) sharp. The variance is nearly half a semitone, so my transfer (Euro Records Coleccion 78 rpm) is almost exactly in the middle of any key you would find on a piano tuned to A = 435. Correcting the error adds four seconds to the transfer but, more significantly, has a very clear effect on the pace of the song and of the timbre of the singer’s voice. It sounds much better.

Here are two 15s excerpts, before and after:

Lastly, I checked Temo (also Euro Records Coleccion 78 rpm), although I already knew that something was very wrong. In the opening phrase of the introduction, the violin vibrato is obviously too fast, and when the singer enters we hear a rather feeble tenor with an over-fast vibrato, rather than the baritone, Mario Corrales (better known as Mario Pomar from his 1950s recordings with Di Sarli). The measured pitch was A = 445.6 Hz (42¢ fast), almost the same as for Sin rumbo fijo, but correcting it left the transfer still sounding odd and Corrales’ voice barely recognisable. There’s no harm in trying alternatives, so rather than retuning down to the nearest key, I corrected by 142¢ or 7.8% (another complete semitone), and listened again.

Straight away, the violin tone sounds right, and at the singer’s entry, I recognise Corrales’ voice. The pace of the song is very different, too, of course. Here are three 45s excerpts, the original pitch, corrected by 42¢ and then by 142¢.

One final issue with Temo is that by 1940 (the recording date) most orchestras in BsAs had retuned to modern concert pitch, A = 440 Hz. I’ve never seen any source commenting on the changeover date, orchestra by orchestra, so there remains a question over whether OTV were still recording at A = 435, or had already changed. The pitch difference is 20¢ (a fifth of a semitone) – quite audible, but not dramatic. If I’m playing Temo in a tanda with earlier valses, tuning to A = 435 means no jarring clash of relative pitch, but the required speed variance is reduced slightly at the higher pitch.

I don’t know whether I’m right. It seems extraordinary that such a massive pitch error could creep into a published transfer (and as far as I can see, all the alternatives fall into line*), and yet the timbre of Corrales’ voice sounds right at the slower speed. At the slower pace, lots of little details in the violin parts like vibrato and bowing articulations sound so much more natural. The tempo is still 204 bpm, so by no means a ‘slow’ vals. The original rattles along at 220 bpm, which is quick.

[Edit: *The tango-dj.at database lists one version (sourced from a private collector) timed at 3:07, compared with the 2:54 of the Euro Records transfer or 2:55 for CdT, and while I haven’t heard it, I have to assume that it, too, is tuned down a whole semitone.]