Musical Pitch

We have a recording at the wrong speed, so what should the pitch be? Given that the abandoned 2015 release had significant pitch issues, we decided to approach the problem by consulting various authorities and looking deeper in the archive. There were four complications arising from the recording’s music content, namely:

1. The singer has an extraordinarily good sense of pitch but also, in the best jazz tradition, she could ‘move’ notes very deliberately and slowly around the target pitch with extraordinary skill. This can occasionally distract forensic listening but most definitely can confuse automated pitch measurement.
2. The 1 kHz reference tone was found to be not precisely at that frequency and was not used as a reference.
3. The piano tuning used a wide ‘stretch’. The accompaniment in several songs was in the upper region where the pitch was measurably different by several cents.
4.  The added reverberation used a stairwell and this brought a tendency for the reverb to decay in pitch (quetzal effect).¹

The mean pitch was measured using Capstan and Sonic Visualiser as well as by direct listening comparison to synthesised notes from Pianoteq. To estimate pitch we selected sections of piano solo in the middle of the keyboard.

Source	Pitch (Hz) for A=
1986 Odin CD	442.55
‘Valentine’ analogue tape	443.10
2005 transfer of Outtakes	442.08
2015 remaster	446.66
2016 digital direct	420.62
2016 analogue capture	413.40
2017 MQA final 2	442.00

1 This last point was very surprising to us, but as we listened more and more closely and also measured the pitch we discovered a non-linearity that makes the whole recording extremely sensitive to being played back at the right speed. It was known (she refers to it in the outtakes) that Radka sang while listening to the mix. Once we observe the effect we can hear that she is ‘singing to the room’ and that the highly nuanced musical communication between her and the pianist take the reverb into account.

2 Morten Lindberg was able to interview one of the piano tuners responsible for the instruments at NRK who confirmed they did all pianos in the house to 442 at that time. The piano tuner at Grieghalle in Bergen was never found, however, Thron Irby, the Norwegian importer of Steinway at that time, confirms A=442 as the most probable pitch and gave a confident range within 442 +/-1 Hz, particularly since this was the piano belonging to the orchestra where tuning was customarily at 442. Accordingly we decided to translate the playback speed in this edition to result in a piano tuning of exactly 442.0 Hz for all tracks including ‘My Funny Valentine’. The original CD releases were fast by approximately the ratio of 44100/44056, presumably because a PAL PCM-F1 was used for the transfer. This means that up to now, no CDs have been at the correct speed.

Fairytales: page 4