Dynamic Range in Mahler VI

Leonard Bernstein recorded two complete cycles of the Mahler Symphonies. The first was made for Columbia Masterworks with the New York Philharmonic during the 1960s. These were analog recordings on magnetic tape, issued on vinyl LPs. Almost all of the second cycle was recorded for Deutsche Grammophon between 1985 and 1988 with three different orchestras, all of which Mahler himself had conducted frequently. This cycle was recorded digitally and issued on CDs.

It is instructive to compare the recorded quality of the two cycles. Because of the limited dynamic range of the medium, the dynamics of the earlier cycle were sometimes compromised when the tapes were transferred to LP. However, when Sony (the company that bought the Columbia Masterworks label) reissued the analog recordings in 2001 it was evident that the dynamic range of many of the original recordings had been recorded on the original master tapes in a form that matches the real acoustic sound levels remarkably closely. A good example is given by the Finale of the Sixth Symphony.

This finale is long and its three major climaxes are crowned with blows of a large wooden hammer. (In Bernstein’s interpretation, the third hammer blow is softer than the other two.) When we look at a graph of volume plotted against time for the New York recording, which was the last in the series to be made, we can see that the first two hammer blows are for very brief moments the loudest sounds in the finale. Although the orchestra plays loudly at other times, and certainly does not sound weak, it never reaches these measured levels. Indeed, the graph follows what one would expect from a study of the score. (The upper trace represents the left hand channel.)

Levels in Maher VI Finale : Columbia/Sony 1967

New York Philharmonic / Bernstein, May 1967

One would expect that the use of digital recording technology with its inherently wider dynamic range would succeed at least as well in preserving the original acoustic levels. But this was not the case, as can be seen from the equivalent graph of the later DG recording. (The overall level of the CD has been reduced by 4dB to make the comparison more obvious.)

Levels in Maher VI Finale : DG 1988

Vienna Philharmonic Orchestra / Bernstein, Sep 1988

One can hardly see where the hammer blows occur on the second recording. One can, however, notice that all the peaks are constrained to the same maximum level, with none of the natural excursions of the earlier recording. The music is severely compressed and, although not actually sounding distorted in the usual meaning of the word, the recorded levels do not accurately follow the acoustic levels of the live performance.

One possible reason for the narrower dynamic range is that the performance was recorded during a concert and that the engineers were perhaps frightened of overloading the digital medium and used an electronic limiter; but overload could have been avoided by other means, and it is a great pity that the later, and more advanced, technology compromised the recording of an important musical performance in a way that can never be undone.


The above examples are expanded from David Pickett’s essay in the book Perspectives on Gustav Mahler, edited by Jeremy Barham and published by Ashgate Press, 2005.

Rev. 12 Dec 2018