Frequency Response Experiment with Cassettes

[vinylisfinyl]

Since I've been stuck at home these past couple of weeks, I've gotten back into playing with cassettes and have been experimenting with recordings on my Pioneer CT-M66R. Today I got the idea to run a frequency sweep through it just to see what kind of response I'd get.

To run this, I used the Log Frequency Sweep Track 6 and 1.6 kHz tone Track 33 from the Bink Audio Test CD, playing them via my PC connected to my Onkyo TX-8050 stereo through a Schiit Fulla DAC with the Pioneer recording on Tape Out.

I used two tapes for the test: a Type I black-shell Maxell Communicator Instant Start series and a Type II black-shell Maxell UD XL II. The latter was an old mix it tape from a cousin of mine, while the former was brand new from shrink wrap as of couple weeks ago. Both had music recorded on them at least one time prior.

After recording the test tracks, I played back both tapes to my PC using and the Line In port on the motherboard (old Biostar from 2008 with an Intel Core 2 Quad). I recorded the tracks using Audacity at 32 bit, 44.1 kHz and exported both tracks as 24 bit FLACs. I then used Spek to draw spectrographs of the frequency response and foobar2000 to playback the files via my DAC/stereo combo through my Grado SR125 headphones. I also exported the original Bink test tracks as one FLAC file to compare

The results were quite interesting. As expected, both tapes had a noticeable roll-off starting at 14 kHz onward, but I was pleasantly surprised I could still hear tones up to about ~16 kHz, the same as the original track (probably a combination of my hearing limit and my headphones). It appears that the Type II tape has *slightly* more dB above the 14 kHz point, seen as more green around the main line. The Type I tape held up quite well considering it's original market intent.

The spikes one sees in the Type I chart are drop outs, which can be seen in the Audacity picture too. Extra harmonics are seen as the matching and mirrored lines elsewhere from the original audio. If you play the files, you can hear them slightly during the 1.6 kHz tone which sounds more pure in the original file than the records. You can also hear a "swishing" above 14 kHz (seen as verticle traces in the spectrographs) which I think are probably filters or something hitting a resonance. Finally, one can see the tape noise as constant deep blue color in the spectrographs, though I can't hear much if any hiss when I play the files.

Overall, a cool experiment with interesting results. If you wish to do this yourself over this long stay-in, you can get the software and test tracks here:
Bing Audio CD test tracks: https://www.dropbox.com/sh/vhtje97km4m713j/AABEFdcsEKGVQRj4SwOi6LzHa?dl=0
Audacity: https://www.audacityteam.org/
Spek: http://spek.cc/
foobar2000: https://www.foobar2000.org

Comment away!

 

[WernerO]

Cool. But a much simpler method exists. If your sound card supports full duplex operation, i.e. playing and recording concurrently, then http://www.audiotester.de/ contains all you need in one program. There is a free evaluation period, so you can try before you buy. But be warned, there is a learning curve, and for two-head decks this is steeper because you have to synchronise recording and playback carefully by hand.

Also note that cassette deck frequency responses are recorded at -20dB, in order to avoid treble saturation. Additional measurements can be at higher levels, but -20dB is the one for comparison purposes.

Here is an example of a 1985 Maxell UDI in a Nakamichi BX-300, at four different levels, with 0dB corresponding to Dolby level, 200nWb/m ANSI:

And here is a 1987 TDK SA in a Kenwood KX-880HX. This being a two-head deck the measurement starts only from 200Hz, because synchronising at lower frequencies is much tougher.

 

[twiiii]

You should try the same experiment with a Dragon or a later Model 1000 XZL. The performance at -10 db will almost match the - 20 db result the B-300 A properly adjusted B-77 1/4 trk at 3/ 3/4 at -10 referenced to 512 nwb will almost but not quite match the B-300 response at -20 db. using Maxell chrome II tapes. All I ever used my 682 and 582 models for were to dub tapes for the cars, and make tape for friends. The 682 has been retired for over 12 years and the 582 was retired today. I haven't played or dubbed a cassette in years so I figured it was time. Plus I need the space for a new TT. I kept my CD and DVD recorders. But thats not the same thing, is it?

 

[vinylisfinyl]

Cool, I shall have to see about that software! I think my sound card might support simultaneous playback/recording, though it's the build-in base sounds card in my PC's motherboard so will have to experiment. Unfortunately my Pioneer is a two head deck so I would have to do the synchronization you mention. Are the extra lines in the audio test the output vs the input?

And no, recording on CDs or PCs isn't quite the same as tinkering with an old tape deck to see which tape sounds the best.

 

[Late Boomer]

Cool. But a much simpler method exists. If your sound card supports full duplex operation, i.e. playing and recording concurrently, then http://www.audiotester.de/ contains all you need in one program.

Is this your blog? https://audiochrome.blogspot.com/2020/04/cassette-tape-comparative-measurements.html Tons of info.

 

[Late Boomer]

I kept my CD and DVD recorders. But thats not the same thing, is it?

Of course, it is not the same thing. Instead of searching and paying for a high-end deck and then refurbishing it and keeping it in top-notch condition and getting the best tapes, you simply burn a CD and get 20-20K on pretty much anything that can play it, including portable "jogger" players ($20 on eBay and if it works it works, not like "kinda sorta" works with wobble and dropouts and constant fear of chewed tape). And you can copy it with no quality loss. And the quality does not deteriorate with each play. And the tape is not eaten or stretched. And when you switch to files, you lose jewel boxes and paper inserts and booklets. Ah, the shiny soulless virtual world.

 

[twiiii]

But there is something special about analog if done properly or something very wrong with digital if not done correctly. How else can you explain the resurgence of LP's. Most of them were less than stellar, but those that are will give a CD a run for the money any day especially if its a reissue not produced with care or by deaf sound people and producers.

 

[Late Boomer]

But there is something special about analog if done properly or something very wrong with digital if not done correctly. How else can you explain the resurgence of LP's. Most of them were less than stellar, but those that are will give a CD a run for the money any day especially if its a reissue not produced with care or by deaf sound people and producers.

LPs' only usage is for live DJing. And nostalgia. And novelty feel for Generation Z. They are made on obsolete equipment off digital files - what's the point? Just listen to the file. Cassettes became obsolete for hi-fi systems when CDs came out, but made sense as portable media unti portable CD players with buffering appeared in mid-1990s. Out of three major consumer media types (gramophone records, magnetic tape, optical media) only the latter still makes sense: durable, best quality, offers some physical properties like it can be purchased after a life show or can be gifted, the booklet can be read, etc. If you don't care about physical aspect, FLAC/OGG rule, even MP3s with high enough bitrate. I do not trust the cloud, a friend of mine sync'd his collection into Apple's cloud and deleted his local files, then Apple "lost" his sync data, so he lost his collection. I prefer having my stuff on my own media, DRM-free.

 

[twiiii]

Its the recordings that were made in the analog domain only that surprises me. You can have your ADD stuff. Direct to disc recordings that no matter what you do can never be equaled with digital miscarriages of producers and engineers. There is a company that is re releasing all of Sheffields direct to disc recordings and the sound is second rate at best. Sheffield recordings were used with totally tube mics and tube mike pre-amps and specially built mixers. Even some of the Neuman record cutting master lathes had tube electronics. So there is definitely a color signature there. But the tube sound is compromised and effectively lost as are the dynamic peaks and ambient background on the digital reissues. The digital files would be just to large if everything was kept linear. Yes you can copy digital files for ever, but they are compressed files and compromised files. Yes analog isn't perfect, but neither is digital.

 

[WernerO]

Are the extra lines in the audio test the output vs the input?

All is output, at three or four different levels relative to Dolby level.
Plotting the input would only yield flat lines.

See my blog, link above, for larger versions of these graphs.

 

[WernerO]

Even some of the Neuman record cutting master lathes had tube electronics.

That would then be custom jobs. Of course anyone can slap a tube power amp onto a cutting rack.

I have the circuit diagrams of the Neumann SX6* and SX7* series. All transistors and, shock, horror, early solid state opamps.

But this is OT.

 

[perryinva]

Interesting blog, but why on earth use the 1994 version of XLII as a reference? Did you not notice that the vast majority of your tests show bias settings in the negative range? Also, as you already know, the farther from the reference input sensitivity level the tape is, then the farther the bias difference. This does not negate the usefulness of the tests, but using that 1994 tape as the reference exaggerates the apparent differences.

I don’t agree with your statement that one should not record peaks with NR above 0dB because of high end saturation as the dynamic range and relative noise floor difference is not worth it. Peaks do not occur with treble but with bass, and the level one chooses shiuld be dependent on the tape and the deck. A BX-300 can record a good TypeII with NR with +3dB peaks all day and never approach saturation. 3dB is certainly worth the noise floor difference.

Also, you should be aware that audiotester is no longer supported by the author and has issues with many soundcards and newer operating systems.

of course, all IMHO.