Oh no, not again! I simply can’t keep up with all these converters. Just when I thought I could settle down with my Apogee AD-8000 (PAR, 11/98, p. 18) and withdraw from converter lust for a while, Apogee did it again and produced another winner. I had even convinced myself that all this talk about 24-bit recording at 88.2 kHz and 96 kHz sample rates was mainly advertising. I was wrong! At the higher sample rates, the 88.2/96 kHz Apogee PSX-100 sounds considerably better than its previous 44.1/48 kHz converters – of course I’ve got to have one.
Product PointsApplications: Recording; mastering.
Key Features: Two independent converters in one unit: a stereo 24-bit/96kHz analog to digital converter, and a stereo 24-bit/96kHz digital to analog converter; bit-splitting processor for recording stereo 88.2kHz or 96kHz 24-bit files across all eight tracks of an MDM; format converter
Contact: Apogee Electronics at 310-915-1000.
+ Smooth sound
+ Relatively low price
+ Flexible I/O possibilities
The Score: If you own an MDM and do a lot of 2-track recording, this is the box for you.
In a stylish purple 1U rack unit, Apogee has crammed a stereo 24-bit/96 kHz analog-to-digital converter, a stereo 24-bit/96 kHz digital-to-analog converter; a bit-splitting processor for recording stereo 88.2 kHz or 96 kHz 24-bit files across all eight tracks of an MDM, as well as a format converter for translating between AES/EBU, S/PDIF, ADAT, and TDIF I/O.
Its A/D and D/A sections can be operated and clocked independently, or cross-connected in several ways. Typical Apogee features, such as two low-jitter clocks, soft limit circuitry and UV22 processing functions, are all present. One can record at any of four sample rates: 44.1 kHz, 48 kHz, 88.2 kHz and 96 kHz; and output either full 24-bit resolution or use UV22 to reduce the word lengths to 20 or 16 bits.
Metering either the A/D output or the D/A input, and monitoring either the A/D output or your recorder’s return signal through its D/A converter, is possible. You can use the Apogee Bit-Splitting (ABS) protocol to record the unit’s high sample rate output to 16-bit machines via ADAT, TDIF and AES I/O connections. Double-wide AES digital outputs permit recording non-bit-split signals onto four tracks of a true 24-bit recorder, such as a Nagra D.
A single AES/EBU connector permits external digital metering of its output, as well as providing an easy way to get a44.1 kHz or 48 kHz digital output from an 88.2 kHz or 96 kHz source.
There are three basic operational modes possible with the PSX-100. The most common one is confidence monitor mode, in which the unit’s A/D and D/A sections act independently, taking advantage of the unit’s dual low-jitter clocks. If you have your analog source, MDM recorder and analog monitoring equipment properly connected, the press of a button connects the recorder’s return signal to the input of the D/A converter. A recorder with confidence monitoring is necessary or you’ll hear what we used to call “EE” mode during recording.
Analog monitor mode routes the A/D input to all digital outputs (including that decimated auxiliary out) at the press of a button. This mode saves tying up the mixdown deck simply to hear the sound of the Apogee unit itself (A/D conversion, soft limit, UV22, etc.).
The third mode is digital copy mode, in which the selected digital input is routed to all digital outputs (including aux), as well as to the D/A input. This mode is useful for doing multiple format conversions.
There are 12 push buttons in all, including the power switch. Depending upon whether they are pushed quickly or held in for a couple of seconds, different things happen. The results are shown on some of the 35 LEDs on the front panel. Once you get the hang of how to push the buttons properly, things go OK. Until you get the hang of it, however, there may be unexpected results.
For example, to enter the world of 88.2 kHz or 96 kHz recording, press and hold the A/D sync switch. If this is correctly done, the x2Fs LED lights up. Don’t even think of trying to figure any of this out without reading the manual!
The rear panel is easier to understand: XLRs for analog I/O, dual XLRs for high sample rate digital I/O, and single XLR for the auxiliary digital signal, which outputs every other sample from a high sample rate recording. There are also a pair of RCA S/PDIF I/O connectors, a pair of ADAT lightpipe ODI connectors and a pair of BNCs for word clock I/O. The rest of the rear panel is completed with a typical TDIF I/O connector, a 10-gang DIP switch and an IEC power connector. The DIP switches control such setup functions as: whether the TDIF clock is WC input or LRCK, whether pin 2 or pin 3 is hot on the A/D input, configuration of the bit-splitting mode on the digital XLR connectors, etc., etc.
I listened to the PSX-100 in my studio and on a remote session. I did everything from air checks of my favorite Sunday night disco oldies show, to solo voice, guitar, oboe and piano in my studio, to a choir recording down the hill at the local Congregational church. All my recordings were also done simultaneously to my 24-bit, 44.1 kHz AD-8000 converter. After each test, I came to the same conclusion: while the AD-8000 sounds great, the PSX-100’s character (at 88.2 kHz or 96 kHz) is several rungs higher in the absolute scale of faithfulness to the source.
For the church choir recording, I had been asked to provide a custom CD-R of two songs written for my town’s 225-year celebration. These were to be played on the Fourth of July – believe it or not – from four horn speakers on the top of the church’s bell tower! Now that modern electronic carillons comprise basically a computer and a CD changer to play digital recordings of real bells, it didn’t take much for a gray-haired church lady to figure out they can play other kinds of CDs as well.
I set up a session like I used to 20 years ago – live to two-track – using a single Manley reference stereo C24 mic, feeding the Crane Song Flamingo mic preamp and my Amek RNCL dual compressor/limiter. That sweet analog output entered the Apogee PSX-100 and was recorded across all eight tracks of my TASCAM DA-38. Since the final product would be a CD at 44.1 kHz and I wanted to test that aux digital output, I used 88.2 kHz. I did six takes of the childrens’ choir and four takes of the adult choir and then schlepped all the gear back to my studio.
I ran the 24-bit 44.1 kHz output coming from that aux connector through my Meridian 618 dither and noise-shaping processor, applied shape A to it, and then recorded it to my 16-bit Dyaxis II workstation at 44.1 kHz – an electronic pathway I’ve been using for years. The Apogee manual says not to use that digital output for anything serious, because the aux connector’s preceding circuitry drops every other sample – and doesn’t filter – but I wanted to hear how bad it sounded.
You see, in theory, one is supposed to filter the signal below the Nyquist frequency before decimation (dropping those samples), so one doesn’t get aliasing distortion. In this case, therefore, anything from 22.05 kHz to 44.1 kHz would alias. This means 24.1 kHz maps to 20 kHz, 43.1 kHz to 1 kHz and 44.1 kHz to 0 Hz (DC). Everything in between 22.05 kHz and 44.1 kHz aliases in like manner.
Notice, however, that harmonic relationships are not preserved when the alias distortions are created. I had hoped not to notice the stuff that aliases below 1 kHz, since human ears aren’t as sensitive to that range. But it would be there, at least in theory.
I went back to the tape, played those 10 master takes into analog, and used the Apogee AD-8000 to make a standard 44.1 kHz 24-bit digital bitstream out of them. To keep as many factors as I could invariant, I ran that signal through my Meridian processor (with the same flavor of dither I had used on the auxiliary digital source), and recorded that version of the 10 choir takes onto my Dyaxis as well.
Lo and behold, by the time my sweet and warm analog signal got to 16-bit, 44.1 kHz, I could hear very little difference between its two completely different ways of getting there! Even while still in the original 24-bit resolution, the sounds were still pretty close. Sure, Apogee warns the user not to use that output for recording, but truthfully, it doesn’t sound bad at all! Certainly nothing like the horrible boxy cardboardy sound one gets when truncating and slamming a 24-bit 44.1 kHz signal into a 16-bit DAT recorder without dither.
Most engineers have learned to recognize that sound, but whichever 50% of the church choirs’ samples came through the PSX-100’s aux digital connector – aliasing distortion included – the result sounded fine to me.
So what exactly was the difference in sound between the AD-8000 (at 44.1 or 48 kHz) and the new PSX-100 (at 88.2 or 96kHz), when each is used as directed? Well, there was a more relaxed quality to the signal, a softer, more open sound. There go those audiophile words again, but I don’t know a better way to put it. By softer I don’t mean lacking highs; indeed, the highs go out another octave.
It was interesting to me that, even when doing an air check of KISS-FM, using my vintage McIntosh MR65B tuner (which doesn’t go past 15 kHz), the relaxed sound was still quite apparent. This despite all the FM multiband processing that I’m sure had gone on top of the 16-bit disco CDs playing over the air.
In comparison to the PSX-100, my AD-8000 had a slightly more closed-in, constrained and etched sound which, by itself, was quite nice; indeed, it’s one of the smoothest 24-bit/44.1kHz converters around. But when I switched between the AD-8000 (at 44.1kHz) and the PSX-100 (at 88.2 kHz), there was no question about which one sounded more open. When used at the two lower sample rates, however, the PSX-100 and the AD-8000 sounded almost identical, as did their implementations of the soft limit and UV22 circuits.
Of course, that still leaves the nagging question, “What about those of us who own expensive AD-8000s, and stocks of AMBUS I/O cards?” Apogee has a long history of providing upgrades. I wouldn’t be surprised if a lot of the sound of the PSX-100 will end up in my AD-8000 one of these days, since so much of its digital circuitry is firmware based.
I have one gripe, and it isn’t really Apogee’s fault – it comes from the laws of physics themselves. It does seem rather inefficient to have to dedicate an entire eight-track MDM recorder just to store two tracks of 24/96 stereo. To continue my analogy from the previous paragraph, that means we’d have to use four of them just to store eight tracks. And as for 24 tracks – as they used to say in my college mathematics texts, “the computation is left to the reader.”
When I complained online that there ought to be some way to shoehorn those bits into fewer tracks, I was told on the best authority (Julian Dunn of Nanophon, UK), that whether one uses Apogee ABS, Rane Paqrat, or Prism MRX protocols, bits are bits, and one can only fit so many into a 16-bit track at a particular sample rate.
Here I am again, teetering on the edge of a brave new world. But I’m so taken by the sound quality of this converter technology, that I’ll just have to learn to adapt. Expect my next few articles to detail the trials and tribulations I’ll no doubt go through attempting to transfer to hard disk and actually edit my 24/96 recordings. I’ll also be writing about any problems I’ll have trying to down-convert the edited sound files (presumably through software, which conventional engineering wisdom asserts as superior) to 16-bit/44.1 kHz ones. Maybe I really should have just sat this round out. Naw – once I’ve heard better sound, I can never go back.