The Giants of Tape, Part Four

The Giants of Tape, Part Four

Written by J.I. Agnew

This time we’re going to have a look at another Ampex tape machine, an ATR-104 with plenty of extras!

It lives at Airshow Mastering in Boulder, Colorado (not too far from PS Audio headquarters). Its owner, David Glasser, has eclectic tastes, also using a Studer A820 tape machine and Dunlavy SC-V monitoring loudspeakers, in a specially designed room, along with a selection of head stacks and electronics for his Ampex 104.

David does a lot of audio restoration work from historic tapes, as well as mastering for new recordings. One of the most interesting tools at his disposal is the Plangent Processes system. This is a means of correcting tape speed, wow and flutter and other transport issues, which may have been recorded into a particularly valuable tape, by using the bias signal as a reference.

Some technical and historical discussion is appropriate here, to fully appreciate how this approach works.

The Ampex ATR-104 and Studer A820 tape machines at Airshow Mastering. The Ampex ATR-104 and Studer A820 tape machines at Airshow Mastering.


In the very early days of magnetic recording, DC bias was used, which was not ideal for high-fidelity recording, but was necessary to linearize the medium, since one of its inherent limitations is magnetic hysteresis. In plain English, if you try to magnetize an unmagnetized ferromagnetic material, at first it resists magnetization. As the magnetizing field is increased, the material ultimately becomes magnetized. At this point, if the magnetizing field is removed, the material will retain some of its magnetization. To bring it back to an unmagnetized state, a magnetizing field of opposite polarity is required and so on.

If we were to try to directly record an electrical audio signal to tape without bias, the recording head would present an alternating magnetizing field to a ferromagnetic object rolling past it. As the ferromagnetic object would refuse to precisely play along, severe distortion would occur, along with high levels of noise due to several magnetic domains remaining at random orientations, which would generate a random signal (noise) upon playback.

DC bias could overcome these issues to some extent, but with side effects, by pushing the tape to a permanently magnetized state along the B-H curve (which is the curve that indicates how a material will respond to an externally-applied magnetizing field), where the hysteresis effect is reduced.

The major breakthrough in magnetic recording was the advent of AC bias. A high-frequency signal, significantly above the audible range (typically 150 kHz on many 1970s tape machines and 432 kHz on the Ampex ATR-100 Series), is used to magnetically bias the tape. This signal effectively remains recorded on the tape, and is rejected by the playback electronics of a conventional tape machine, allowing only the audio signal to reach the output.

Regardless of the exact choice of bias frequency, this signal is supposed to remain extremely constant, as the oscillator circuits used to generate it were very stable, and their output was tuned to the recording head, very much in the same way that RF circuits are tuned to resonance.

As such, any variation in the frequency of the bias signal would be due to tape transport errors, the effects of which would be detrimental to the audio frequency content, as it would essentially be reproduced at the wrong speed, or with speed instability. However, since tape machine playback electronics typically reject this signal, there is usually no way of knowing what is going on in this regard. Unless, that is, the playback electronics do not reject the bias signal! This is exactly how the Plangent Processes approach works.

The bias signal is captured along with the audio and used as a reference to correct any tape machine errors. If you can make the bias signal stable again, the audio will be made stable along with it.  Sometimes these effects are rather subtle and not immediately obvious to the untrained ear, just by listening to the sound. Variations in the bias signal can be detected by measurement instruments, at magnitudes below threshold of audibility in playback of the sound. There is, of course, a lot more to audio restoration than this, but I shall prefer to let David tell you more about it:

J.I. Agnew: How did you get professionally involved in audio?

David Glasser: My first professional audio job was operating the tape duplication systems of the Boston Symphony Transcription Trust, a radio syndication service, in the days when that meant shipping tapes to subscriber stations weekly. After several years, I also helped record concerts by the Boston Symphony and Pops orchestras at the legendary Symphony Hall.

JIA: What made you choose the ATR-100 for your work?

DG: My first professional tape machine was an Ampex ATR-102, but that particular deck was finicky and the transport was difficult to align. I replaced it with a Sony APR-5000, an excellent but not a very popular machine. Thirty to forty percent of my mastering work was from 1/4-inch and 1/2-inch analog tape, so owning a first-class machine was essential. In 1995, the APR-5000 was replaced by a Studer A820, which I still own and use. A very discerning client was not sold on a master created using the A820, so I borrowed an ATR-100, and that did the trick.

That convinced me that having several playback options was the way to go. The ATR has four playback electronics options: [getting the audio signal] direct from the stock audio cards, [using the] stock input and output, [using a] transformerless input/output, and [utilizing the] Plangent Processes electronics.

The Ampex ATR-104 with a selection of electronics.
The Ampex ATR-104 with a selection of electronics.


I purchased an ATR-104 in poor condition and had Mike Spitz of ATR Services, Inc. totally restore it and add several modifications. That’s been my preferred machine for 20 years.

JIA: What are the challenges in restoration work?

DG: There are several challenges. The key is [achieving] proper playback. Many older master [tapes] do not have alignment tones, so setting the playback EQ is trial and error, and adjusting [tape head] azimuth can be challenging. Sometimes the track format of the tape is unknown: a magnetic viewer, a variety of playback head stacks, and a selection of MRL (Magnetic Reference Laboratory) test tapes are needed. If the tapes are from the era of “sticky-shed [the deterioration of the binders, or adhesive, that hold the magnetic particles onto the backing – Ed.] we can bake them in a laboratory incubator. With many older acetate tapes, the splices dry out and fail. Just rewinding a reel and fixing all the splices as you go can take a half-hour or more.

A serious collection of head blocks for the Ampex ATR-104, at Airshow Mastering. A serious collection of head blocks for the Ampex ATR-104, at Airshow Mastering.

For projects that have the budget, we will use Plangent Processes, which removes wow and flutter and other speed and tape transport-related problems from analog masters. When transferring, we use Plangent Processes playback electronics and high-bandwidth tape heads. The electronics have facilities for identifying and isolating the bias signal on the tape, which is used as a speed reference for processing the resulting digital audio files, which are sent toPlangent’s headquarters in Nantucket, Massachusetts.

JIA: How would you compare the ATR-100 and the Studer A820 transports?

DG: Both the ATR and A820 transports are very precise and stable, but the A820 is gentler on tapes, and easier to edit on. A future project here will be mating the A820 transport to different playback electronics – probably the Plangent Processes preamp, as the stock A820 electronics are not as musical to my ears as the ATR or Plangent.

JIA: Are there many new recordings coming in on tape, or is it mainly restoration work that keeps your tape machines rolling?

DG: I have a handful of clients who prefer to mix to tape, though it’s a small percentage. Most of our tape sessions are legacy reissues or archive projects.

The Picoscope. The Picoscope.


JIA: Listening to all of the old recordings that go through your hands, and then the new ones, are there any striking differences in sound, approach, or musical performances, between then and now?

DG: I’d say the biggest difference between older analog recordings and newer recordings – analog or digital – is [that] it sounds to my ears that engineers back in the day delivered mixes that were closer to the desired finished product. This was likely because when cutting lacquers [in] real time from tape, there is not as much opportunity to address mix issues as there is in a DAW (computer-based digital audio workstation).

JIA: How important is your monitoring environment for your work?

DG: Monitoring is crucial! Speakers plus room. I’ve been using the same Dunlavy SC-V loudspeakers for the past 23 years, in a control room designed by Sam Berkow of SIA Acoustics. Monitoring should be the starting point in any studio design. Even in a setup for just doing transfers, monitoring is crucial.

The Ampex ATR-104 from the side, with the mastering console visible in the background. The Ampex ATR-104 from the side, with the mastering console visible in the background.


JIA: Do you have a personal favorite tape format (in terms of tape width, speed and equalization)?

DG: Favorite tape format – hard to say, but most of the classic albums that sound so good, and [that] we all love were recorded on 1/4-inch tape at 15 ips (inches per second). That said, 1/2-inch at 30 ips with Flux Magnetics heads has a lot of depth, and I sometimes use that when clients want to bounce their digital mixes to tape. Some of the most realistic-sounding recordings I’ve heard are 1/2-inch 3-track, 15 ips on Scotch 111 tape.

In the next episode, we will move on to a different beast of a tape machine!

Header image of Airshow Mastering and other images courtesy of David Glasser.




Back to Copper home page

1 of 2