Let’s Talk Noise Reduction

There’s more to the story of analog tape than what I’ve revealed in recent posts. I’ve written a lot about the failings of analog tape when it comes to its inherent noise floor, which hangs about 60-72 dB below the signal. That’s the figure for a first generation tape recording done on a professional machine using quarter inch tape running at 15 inches per second in stereo. This tape format was used for tens of thousands of masters made during the heyday of analog tape in the 60s and 70s.

My first professional multitrack machine was a 3M 56 16-track 2-inch machine. It was huge, heavy, and ran like a truck. It was the first multitrack machine to have all of the electronics located in the belly of the machine and the meters above the transport. The 2-inch format allowed for the same track widths that were available to the 2-track quarter inch format mentioned above. In the land of analog tape, the wider the tracks the more signal can be recorded and thus the noise floor diminishes. That’s why some esoteric machines run only 2-tracks on a 2-inch tape…I think it’s ridiculous but there are custom machines out there that are built to stretch the limits.

Then along came the ubiquitous 24-track multitrack machine. This format gave producers, engineers, and artists 8 more tracks to play with…but the fidelity of each track suffered because they got thinner. The difference is not huge but it is noticeable. The level of hiss on an analog tape is manageable on the master but grows with each transfer…by about 6 dB. You can mute tracks or use gates during mixdown but a copy is a copy and the increasing hiss level can’t be denied. The final master delivered to the guy cutting the lacquer can be as much as 3 or 4 generations down from the original tracks. That’s one very compelling reason why digital rules…you don’t have any generation loss.

How do you improve the dynamic range of analog tape? If you’re Ray Dolby, you design and build equipment that compresses source signals prior to recording them and then expands them back to their original dynamics on playback. He introduced a 4-band noise reduction method called Dolby A for professionals back in 1966. His approach allowed the mixdown machines another 10-15 dB of noise reduction depending on the frequency in question. His invention was a major step forward for professional studios using analog tape.

But the improvement came with a downside. The equipment required very careful calibration to ensure that all of the compressing and expanding of the signals were matched at every frequency and level. The process was challenging but not impossible. I can remember aligning countless machines during my stint at Mama Jo’s studio in North Hollywood.

The use of noise reduction systems like Dolby A and its competitor dbx were widespread in professional studios back in the 70s and 80s. They were matched with the Dolby B consumer format used on cassette tapes to bring maximize high frequency response.

It wasn’t until 1986 that Dolby introduced its SR noise reduction system for multitrack machines. Called Spectral Recording, Dolby SR was a very aggressive scheme that used a very large number of filters that responded in realtime to the source signals in order to maximize the level of the signal and minimize the noise of the tape. The complexity of Dolby SR was many times greater than Dolby A. It was used in professional studios but was very expensive and cumbersome. By the mid 80s digital systems were gaining a foothold in studios.

I can remember listening to an undecoded SR tape. The sound was almost unintelligible…not merely dull or muffled like undecoded Dolby A. To think that every bit of the source signal is restored without distortion is a stretch. Dolby SR worked but wasn’t embraced by audiophile labels.