RETRO-MEASURE: Pioneer CT-S605 Cassette Recorder/Player (1989). On cassette quality / fidelity. Cassettes are "coming back"? And sentimental associations...

After the usual recent posts on DACs, DSP, speaker measurements, let's discuss something very different this week!

The other month, I was cleaning out some stuff from my basement and came across an old cache of "compact cassette tapes" which I had recorded back in my high school and early university years. I also still have a handful of old pre-recorded tapes that provided the soundtrack of those times in life.

Within this collection were some made in the mid-'80s until about the mid-'90s when affordable CD burners (~$500 I think around 1996) came on the market and I quickly shifted to making "mix-CDs" rather than mix tapes. In fact, I found some almost brand new tapes that I forgot I still had:

Left: Fuji DR-I standard Type I Fe2O3 "Normal Bias". Right: Maxell XL II Type II CrO2 "High Bias". Center: Sony Metal-SR basic level Type IV "Metal Bias" for "Digital Excellence" - back in the day :-).

This brought back memories of going to the local music and electronics stores picking up blank cassettes to archive some of the radio programs I enjoyed. Among the cassettes were "Top 40" countdowns of my youth and the yearly New Years Eve "Top 100" countdowns. Those were the days before MP3 downloads and all-you-can-consumer music streaming of course. Days when one could not plug in the name of a song into YouTube and be treated with full albums or videos. Days when recorded music was actually still a commodity that wasn't rare, but one needed to spend some effort (time and/or money) to find - whether at the local music store, waiting to hear on a radio station, or even trying to sign out at the public library if one were fortunate to have a well-stocked library nearby!

While I still have an old "lo-fi" CD/tape-player box, I figured it was time to at least get up to "par" with a decent cassette player and digitize some of these old cassettes. A number of them containing tracks I haven't heard in decades, some from obscure bands and shows I have not seen or heard from since.

Looking around the local online buy & sell places, I was fortunate to find a "minty", barely used cassette deck available at a phenomenal price! Here's the Pioneer CT-605S:

This baby was "born" back in 1989 and had a list price of around US$350. Inflation adjusted to 2021 dollars, we're looking at something like $700; a ridiculous amount considering how far technology has moved on over the decades! Obviously I was able to buy this for "pennies on the dollar" as they say.

It's a typical 2-head unit, "Made in Japan", Dolby HX Pro with Dolby B and C NR systems. Settings are there for Normal/Chrome/Metal tape bias. Remember that tape bias has to do with changes in the AC signal (frequency and intensity) applied to the tape during recording that would improve signal-to-noise depending on the magnetization properties of the tape formulation (see explanation of hysteresis here and also look up "coercivity"). Remember that depending on the tape type, one should use different recording levels to optimize quality as well. For example, on this deck the manual recommends targeting a maximum of +3dB on the level meter for chrome tapes and for metal tapes, target with peaks up to +5dB optimally. There's also a "Record Bias" adjustment to balance high-frequency content and minimize distortion. Bias adjust seems like an art unto itself and some tape decks have a calibration mode where you turn the bias to balance out high/low frequency signals (typically something like 15kHz and 3kHz tones). For me, I'll just leave this at the balanced midpoint which sounds good and I'm not that picky unless anomalies are obvious.

Just in case I lose it and if anyone is looking for it, here's the Pioneer CT-S605's manual scanned in.

There's also an MPX ("multiplex") Filter which removes the 19kHz stereophonic phase pilot tone when recording from FM radio. While not the highest end cassette player out there obviously, this would certainly be the best unit I've had access to since the mid-1990's when my dad had a 3-head Sony TC-K611S that sounded excellent and supported Dolby S, a step above Dolby C.

Remember that sound quality of compact cassette tape technology is limited by the track width (one of a number of limitations). The tape itself is 0.15" or around 3.8mm only. Each "side" has 1.5mm of tape width to record on, and being stereo channels, we're only looking at 0.6mm of recording width for each channel (0.3mm gap between channels). During record/playback the tape runs under the head at 1.875"/s (4.75 cm/s) only. Compact cassettes therefore are physically arranged as 4-tracks, with 2 channels + 2 sides running in opposite directions on the magnetic tape. Compare these kinds of numbers with reel-to-reel technology typically with 1/4" (6.4mm) 2-track tape spinning at 7.5"/s - 15"/s (19-40cm/s) for high-fidelity playback, sometimes even faster at 30"/s.

Within these limitations, techniques have been developed to improve recording and playback quality to optimize the precious "reel-estate". Dolby HX Pro is one of these advancements described as a "headroom extension" system first created by Bang & Olufson then licensed to Dolby, you can see this indicated in the text on the back:

There are "Control" jacks which I guess are used in a daisy-chained fashion for "Pioneer System Remote Control" devices that allow you to pause and presumably unpause with other devices.

Dolby HX Pro is a recording technique that varies the bias to improve high frequency headroom and dynamic range. The best technical description I have seen is from a 1984 paper from Jensen & Pramanik. HX Pro is generally available from mid-level cassette recorders and up, and does not require playback decoding. Like with this machine, typically the feature is always on whenever a recording is made. In contrast, Dolby B (1968) and C (1980) are noise-reduction "NR" schemes based on pre-emphasis on recording and de-emphasis on playback - read more about this on the Wiki. Generally, if you're playing audio, stick with whichever technology it's recorded with. Calibration does play a significant role which compounds issues if you're using multiple machines, each with potentially slight misalignments (not to mention variations in the physical tape itself!).

As a generalization, for mixed-machine use, the combination of Dolby HX Pro + Dolby B is a good middle-ground which allows for good sounding recordings even if you have a non-Dolby player. However, for best recording and playback, Dolby HX Pro + Dolby C recording should provide the best fidelity if you have a Dolby C player, both of which equivalently aligned and calibrated.

With the various technologies, it has been claimed that high-end cassette tapes/players could achieve 15Hz-22kHz +/-3dB, wow-and-flutter <0.025%, SNR 61dB with metal tapes, up to 75dB with Dolby C, up to 85dB with Dolby S, and even 90dB with dbx noise reduction; subjectively on par with CD quality for most listeners. Of course these are aspirational numbers for the most part and in the real world with analogue technology, unless one had immaculate maintenance and head calibration, it's really hard to achieve much less continue this level of performance in the long term!

As a mechanical device, the first thing that I needed to do when I got this home was to grab some Q-Tips and clean out the erase and record heads. A little bit of 99% isopropyl alcohol will do the trick using those Q-Tips. While at it, clean out any residue from the capstan and pinch roller mechanism. The pinch roller can get gummy/sticky over time so check for that. Remember to give it a few minutes for the solvent to dry, especially if you use lower percentage alcohol before use. Here's a shot of the tape heads and roller mechanism in this unit, a bit of tape residue on the pinch roller:

Some recommend using a head demagnetizer but I think this is controversial and some do not believe cassette tape heads actually magnetize over time (unlike reel-to-reel players), thus one could actually risk causing damage if the demagnetization procedure is not done right. Head azimuth realignment might be needed; I'll leave that to the experts!

Here she is in my sound room hooked up and playing one of my old tapes:

Multi-segmented level meter OK as a gross approximation... Analogue meters would be better and look more awesome. :-)

Hmmm, despite the age of my tapes and the fact that they were recorded with various tape decks over the years, playback actually sounded pretty good. Even the FM radio stuff sounded alright. Certainly of adequate quality to rekindle memories and conveniently converted to digital for easy access and safe-keeping in the archives.

I thought it would be fun to run some test signals through a few cassette tapes with this player to see what kind of results we find. Caveats apply here - I'm using tapes decades old (but minty), the deck is decades old (but in excellent shape), but not calibrated recently. Even if not in absolute optimal state, I think the results can be useful to help us compare measurement characteristics with modern digital... A bit like the measurements I did years ago examining vinyl playback quality.

MEASUREMENTS:

Before we start with what I'm going to do here, remember that there is a whole literature about the measurement parameters of the magnetic properties of tapes, their performance characteristics including acronyms I have never used in this blog... Stuff like Maximum Output Level (MOL), Saturation Output Level (SOL), Maximum Twin Tone Level (MTL), Relative Tape Sensitivity (S), Print-Through (P), etc. on top of physical characteristics important for the recording/reading device like exact tape thickness, coating materials, tensile strength, etc. My interest as an audiophile here is mainly to show whether I can find differences between different settings like Dolby B and C, and develop an idea of cassette characteristics relative to what I've experienced with high-fidelity audio. So no... I won't be bothering with stuff like MOL and SOL. I see that the Audiochrome Blog has done some really great work on this! Check there.

Here's what the playback/record chain looks like for these measurements:

To create test signals: Raspberry Pi 3B+ "Touch" --> USB --> SMSL M100 Mk II DAC --> shielded 3' RCA --> Pioneer CT-S605 cassette deck --> cassette tape of choice

To playback signals + testing: Pioneer CT-S605 + cassette tape of choice --> shielded 3' RCA --> RME ADI-2 Pro FS ADC --> USB --> measurement computer

For ease, I had the little SMSL M100 Mk II DAC set up already from previous testing and we know the output quality is good (absolutely better than analogue cassette tape fidelity). We'll use the line level (up to 2Vrms) DAC output to record the test signals to the various cassettes with the appropriate settings. For Normal and CrO2 tapes, I will not record with output level past +3dB. However, I could go up to +5dB maximum with Metal tapes.

First, let's have a look at the cassette output playing 1kHz and 10kHz stereo sine waves sent to a digital oscilloscope like I normally do with DACs. The signal was recorded to a Sony Metal-SR tape at a strong +5dB level, Dolby C with HX Pro:

Good looking sine waves. Both 1kHz and 10kHz have about the same output level so I'm expecting at least a reasonably flat frequency response from 1-10kHz with this metal tape. The two channels are quite well balanced in amplitude although there's some fluctuation depending on tape quality. The right channel was slightly louder as we can see best with the 10kHz signal. No phase issues, but on occasion during realtime playback, I can see transient anomalies likely due to tape irregularities.

With a 1kHz signal at +3dB record level on the multisegmental meter, let's have a quick peek at the THD+N FFT:

Even if we don't go much further with the testing, I think that graph above already gives us a good indicator of the distortion magnitude found in cassette tape performance. Notice the THD of 0.2% and THD+N of 5.7%; noise level tends to be quite high with cassettes from 20Hz-20kHz. While I can improve those numbers a bit by optimizing the recording level, clearly the performance here is much inferior than digital. Notice from the FFT that very commonly, the 3rd harmonic predominates in cassette tape output. In my testing with this Pioneer, the 2nd harmonic increases with further increase in recording level as well.

Another observation in the FFT worth pointing out is the wide "base" to the 1kHz tone. This is similar to and in fact much worse than any decent DAC with random jitter! That wide base is a reflection of frequency instability around 1kHz.

Speaking of frequency instability then, can we see this in the time-domain? Sure... Let's look at the wow-and-flutter characteristics. Unfortunately, the Dr. Feickert PlatterSpeed app is no longer available on the app stores (Apple or Google). However, for us Android users, we can still download the version 2.1.1 APK here. Although I can no longer calculate DIN values like I used to as part of the in-app purchase (hey, can I get my money back!?), we can still look at the pitch stability graphically:

The SMSL DAC graph had a couple of blips at the start and end... The tablet is picking up sound from my hand moving near while recording the 3150Hz playback; otherwise tonal accuracy is spot on!

For this test, I recorded a 3150Hz tone from the DAC onto an inexpensive Fuji DR-I tape (with Dolby B encoding). As you can see, pitch stability of the cassette over around 40 seconds will fluctuate a bit. The fluctuation is on a more rapid scale than the much slower rotating 33.3rpm pattern one sees with LPs. For comparison, see the frequency pattern here for my Technics SL-1200M3D. In any event, the raw deviation from a perfect 3150Hz target with the cassette is about +/-0.3% which is about the same as the Technics turntable that has a very good direct drive mechanism. Too bad I can't calculate the DIN value.

Finally, let's run some RightMark tests with the cassette player using different cassette tapes and let's play with the various Dolby NR settings! The recording levels were kept at a peak +3dB for consistency (even though we could push this further with the metal tape).

I. Fuji DR-I (Normal Bias) Tape (bought 1995):

Here's an "almost unused" Fuji DR Type I that had a previously-untouched Side B. Let's record the RightMark test signal on this without NR, with Dolby B and with Dolby C. Then play back and see what kind of results on the metrics.

As you can see from the numbers above, clearly the tape playback is much inferior to the <$100 SMSL M100 Mk II DAC! The frequency response numbers look dire. This is because the tape playback is clearly not able to achieve a full flat 20Hz to 20kHz spectrum which is what the software is expecting. We can see more on the graphs:

Interesting that the Dolby NR settings resulted in diminished frequency response with the "Normal" tape, in particular the Dolby C setting which stuck out (perhaps this is related to "saturation" effect like with CrO2 touched on below?). Nonetheless, the noise reduction is working with lowered noise floor.

II. Sony UX Type II (CrO2) Tape (bought ~1990?):

Next, let's check out this Sony UX Type II tape which is not as "fresh" as the Fuji above but can be representative of how a CrO2 tape performs I think.

You can see that compared to the Normal tape above, this one improves the dynamic range by about +5dB across the board with or without Dolby. And here are the graphs:

Curiously, we see that CrO2 tape has worse frequency response compared to the Fuji "Normal" cassette above. We are looking at about -6dB by 13kHz. I suspect this is due to "treble saturation" which is shown to be worse with Type II tapes when recording at higher levels. Remember that in reality, music rarely has strong "0dBFS"-type high frequencies.

To confirm that the CrO2 tape indeed can manage reasonable frequency response, let's drop the peak recording amplitude down to -3dB on the cassette peak meter and have another look:

Yup. Indeed, best not to overload the recording level for high frequencies otherwise they won't "stick". With the record level lowered by 6dB, we can see the frequency in fact extends out to 16kHz which is pretty good. One of the many nonlinear properties to be aware of with the compact cassette magnetic medium.

[Addendum: I was reminded by my audio engineer friend that cassette deck frequency response is measured at -20dB due to this saturation effect. Good to know!]

III. Sony Metal-SR Type IV Tape (bought 1989):

Finally, I have a "new old stock", unused Sony Metal-SR that was opened years ago but I never recorded anything on it. Results summary:

Nice to see that the metal tape resulted in yet better performance numerically. And so they should, these certainly cost significantly more than CrO2 tapes!

Interesting to see that this tape benefitted very significantly with the use of Dolby B and C! I like that the frequency response was good to near-20kHz. Remember that since frequency response is an extremely important component in high-fidelity reproduction, the fact that this is so good with the metal tape goes a long way towards the perception of subjectively "high-fidelity" sound quality, so long as parameters like noise level and distortion remain adequately unobtrusive.

In Conclusion...

Over the years, I have not see many measurement of cassette decks although the Audiochrome blog has a number of tape measurements. The Nakamichi Dragon has been considered by many as the "holy grail" of cassette tape recorders/players - here are some measurements on ASR (hmmm, phase issues? questionable test tape?) and in Sound & Vision. 

I think it's hard to say too much about the sound quality of cassettes simply because it depends on the media, the recording technique, and of course the tape player itself. As a generalization, I find cassettes have a more "rounded" sonic quality than modern digital and generally not as "harsh" or "fatiguing". Most likely this is a reflection of limited deep low-frequency and high-frequency extension as well as lower resolution so nuances are more "impressionistic" than detailed. Potentially unpleasant sonic edges are smoothed over. This could sound "euphonic" with certain recordings even though fidelity is limited. While you don't get the scratches and pops of LP playback, tape hiss can be an issue and poor quality tape may have random drop-outs in signal, crackling, and warbling sounds.

With the metal tape and its extended frequency response, CD dubs sound very good. Even though it would be hard to properly do a blind test for me, the sonic difference between a 16/44.1 CD and high quality cassette tape would still be audible based on very significant resolution differences. The CD would be more "precise", and of course if you turn up the volume, noise floor would be vastly better with improved low-level details. Having said this, the old cassette recordings of Little River Band's Monsoon, and the Ghost soundtrack on a couple of my metal tapes sounded very good on the main system despite the age!

A few things I've learned from these measurements:

1. Cassette tapes themselves make a huge difference as you might already expect. For the best sounding dubs and CD copies, the metal tape performs well. CrO2 tapes are good as well of course but it was interesting seeing the "treble saturation" effect with the test signal recorded at high level.

2. Dolby C used with CrO2 and metal tapes on this deck consistently surpassed Dolby B in quality. The performance difference was not as impressive with the Fuji DR-I "Normal" ferric tape. As such, I think with Type I tapes, I'll just stick with Dolby B for compatibility.

3. The measured differences here are audible. So often with digital equipment, people question "do the measurements make any difference when listening?" I can tell you that when it comes to these cassette tapes, the answer is a definite yes! A few seconds with headphones will easily differentiate sound from a normal vs. CrO2 tape, whether Dolby B or C is in use, or if the recording was made at an adequate level. The fact of the matter is that there is a "ceiling effect" once you hit a "good enough" level of performance. Cassette playback quality, even when excellent, has clearly not surpassed the point of actual "transparency" yet (based on the limits of human hearing). For example, with the high noise floor of cassettes, a few dBs difference between various tapes may quite easily be distracting whereas with high-resolution digital, even many dBs of differences between DACs with noise levels down below -110dBFS would be inaudible.

4. Time domain inaccuracies are there but different from vinyl and much worse than digital. We don't see the slow sinusoidal variability as in turntable measurements (these are quite easily audible). Instead, pitch stability on this unit was much more random. Again, I truly do not believe audiophiles should be worried about the microscopic jitter in digital playback compared to the orders of magnitude anomalies with analogue gear.

Well, it has been fun checking out the old cassette technology. Time to stroll down memory lane and rip a few more cassettes over to the computer, archive the music, and put the Pioneer player into its new home on the audio rack for very occasional duties.

--------------------

In the tape world, whether it's the azimuth angle for head alignment, cleaning tape heads, choosing the right cassette to use, whether to apply noise reduction, even lubrication issues for old tapes after all these years... In the vinyl world whether it's the subtle VTA adjustments, measuring tracking force, applying anti-skate, selecting cartridges and phono pre-amps... An "analogue audiophile" can certainly say with some confidence that "Everything matters!".

However, when we look at those graphs comparing cassette playback performance with the <US$100 SMSL M100 MkII DAC, I think we should at least consider whether the idea that "everything matters" actually is true these days when many devices have achieved such levels of high-fidelity that perhaps there is absolutely no need to obsess over every little thing. 

While some audiophiles seem to have trouble with the CD claim of "Pure, Perfect Sound - Forever", when we look at the performance of consumer analogue audio such as cassette tapes here or vinyl previously, it's quite clear that even the venerable CD standard from the early 1980's is miles ahead in terms of fidelity (when the recording and mastering of music are done well of course). And that's why digital is the dominant technology in music reproduction. It doesn't take a "Golden Ear" audiophile to recognize that digital sound quality is clearly superior, and IMO only those who are highly biased or delusional would insist that something like vinyl or even cassettes has better fidelity!

When I was purchasing cassettes back in the '80s mainly, a number of commercial releases were of poor quality and many of them would have imperfections and brief drop-outs that were annoying. As a kid, I didn't make enough from the newspaper routes and part-time work to purchase LPs at the time but already was quite dissatisfied with the sound quality of tapes. By the late-'80s, I was glad to hear the digital revolution - always on the lookout for sales and saving up pennies to buy CDs instead.

I hear rumblings of a "come back" in the compact cassette as perhaps part of a resurgence in old-skool analogue. Hey, there's even a company called "Cassette Comeback" I think based in the UK selling blank tapes primarily. Looking at the online prices of the cassettes tested recently, I see the Fuji DR-I goes for £2.99 each, the Sony UX Type II for £14.99 and the Sony Metal-SR at a whopping £24.99 each! I guess like the vinyl "comeback", it ain't going to be cheap...

You can still buy new pre-recorded tapes these days like Lana Del Rey's NFR!, Madonna's Madame X, Lady Gaga's Chromatica, Taylor Swift's Lover, or The Strokes' The New Abnormal presumably appealing to the younger generations. I honestly don't know how many people buy these (at around $20-30 per tape) and I suspect they're more seen as fan keepsakes in a collection rather than listened to regularly. Seriously, they'll sound better streamed and you won't wear out the tapes.

Remember, like vinyl, there's a limited lifespan for the number of playbacks and one should be careful to prevent tapes getting eaten up by players - a risk each time you run those thin magnetic strands. In time magnetic properties can degrade. I've seen quotes that cassette tapes have lifespan of around 30 years when cared for, and here's some excellent tips for all kinds of magnetic media. Also, effects like "print-through" can be easily heard after years of storage (like a pre-echo effect a second or two early).

Anyhow, since nothing really lasts forever, there's no better time to have a listen to some really old cassettes now that I have a "new" player :-).

Epilogue:

As I type this, I'm listening to some cassette rips of mixed tapes and very old cassettes of foreign origin - some old pre-recorded tapes from Hong Kong and Singapore my parents had from the early '80s. One of the tapes was recorded with dbx and it sounded impressive even after all these years.

While I would not consider cassette tapes as a viable format to convey high-fidelity in 2021, there were times I was genuinely impressed by what I heard... Remember that as humans, there is such a thing as "good enough" quality for subjective pleasure.

Listening to these, I am transported to another time, another place. That's a wonderful feature of how music weaves into our lives. How it evokes not just emotions, but along with it, those "state dependent" memories one associates with the feelings and ineffable qualia of those times in life. When I listen to that Enigma MCMXC a.D. tape in the photograph above, the sonic imperfections unique to this tape and heard probably hundreds of times through my old Aiwa cassette player decades ago reminded me of countless late-night listening sessions during medical school almost 3 decades back. 

The power of those associations I believe are linked to the fidelity of the sound itself - specifically the limited fidelity of the tape. For me, some of those memories and feelings are not as strongly evoked with a pristine digital version of the music, but the coloration of the "cassette sound" and its imperfections (I suspect too the "warmth" of vinyl, tubes, high noise level, channel imbalance, crosstalk, etc. of older technology) bring us to that place in the past that seems more personally genuine. However, it's important not to conflate this intimate subjective personal experience with the actual (limited) playback fidelity of the medium or technology.

------------------------

BTW guys, for those looking for a deal on Apple AirPods Pros, I see it's 20% off on Amazon in Canada currently and similarly in US. Not sure if they're planning to reduce prices, but Apple products on sale like this doesn't seem that common. They're not cheap by any means, and whether you're an Apple fan or not, the headphones are actually quite nice with better look, better sound, sweat/water resistant, fits more securely in the ear (for me at least when out running) and has active sound cancellation compared to the original AirPods.

Hope you're enjoying the music, dear audiophiles...

Addendum: (March 11, 2021)
Shortly after the publication of this article, I see that Lou Ottens, the "father" of the cassette tape has passed at 94. RIP.