The sound of magnetic tapes: comparing musical samples from digital versions and magnetic tapes Review

[Jean.Francois]

Hello,
Analog magnetic tape has long been a reference medium used by the biggest bands on machines that are legends today, such as the Studer J37 used by the Beatles or the Studer A80 used by Pink Floyd for their album The Dark Side of the Moon.
Whether stereo or multitrack, with bandwidths from 1/4″ to 2″, tape has been used to produce thousands of albums and masters. I had the chance to see and listen to these marvellous tape recorders, and you can find all the information I obtained during my visit to the DES studios.

Today, I'd like to propose an experiment. We're going to record 10 musical samples from different digital media and in different formats: PCM 24 88.2, DSD64 (SACD), DSD 256, DXD (24-bit 352 kHz)... Several musical styles are represented: Classical, Jazz, Rock, Pop, such as Patricia Barber, Miles Davis, Tsuyoshi Yamamoto Trio, Oscar Peterson, Gnomus, Thomas Schirmann, Anne Bisson, Michael Jackson, Dire Straits, Roxy Music.
For the recording, we use a TEAC A3440 (2 tracks) and a Dolby Model 363 with Dolby SR noise reducer. The tape is a Recording The Masters SM900 and the tape recorder has been calibrated for this tape.

TEAC A3440 and Dolby Model 363 SR
The reference tracks were recorded on SM900 tape at 38cm/s (15ips) with Dolby Model 363 SR, and played back with 24-bit 176.4 kHz digitization.
Before listening, let's start by comparing the bandwidth of the digital version and the tape-recorded version below: the straight line (blue) is the digital version (perfectly linear over the entire spectrum up to 88 kHz), and in white the tape, we see a linear bandwidth rising above 20kHz, with attenuation below 40Hz.

Spectrum bandwidth Tape (white) vs Digital (blue)

For each of the samples below, you can listen (here) the reference version and the version recorded on magnetic tape, as well as the spectrum comparison graph:

• Patricia Barber – This Town
• Miles Davis – So What
• Tsuyoshi Yamamoto Trio – Misty for direct cutting
• Oscar Peterson – You Look Good To Me
• Gnomus – Mussorgsky (Fritz Reiner)
• Thomas Schirmann – Too Yong To Die
• Anne Bisson – Killing Me Softly
• Michael Jackson – The Girl Is Mine (with Paul McCartney)
• Dire Straits – Private investigation
• Roxy Music – Avalon

You'll find full details of how we made this test, with all the samples and measurements, here.

Which do you prefer, the digital version or the tape-recorded version?

Enjoy listening
Jean-François

 

[Deleted member 21219]

It's interesting to play both files at the same time, with a quarter-second delay in one. This allows the ear to compare percussive sounds with practically no delay, although low-frequency sounds with slow rise are essentially obscured. You sacrifice one to gain the other.

This trick sounds "wonky" at first, but the brain soon adapts to it. The results can be ..... interesting.

Jim

 

[fpitas]

I'll note that tape degrades with time. Your experiment is still interesting and valid as far as it goes. But those old tape recordings will have gone downhill with time.

 

[Head_Unit]

Michael Jackson – The Girl Is Mine (with Paul McCartney)
Which do you prefer, the digital version or the tape-recorded version?

I don't want to say, because I'm a lover not a fighter! I am quite certain however that the doggone girl is mine!
Ha ha this is very interesting, I'll have to try it. I'm pretty much a digiphile but some great stuff was done on tape without sounding tape-y (The Doobie Brothers TheCaptain And Me for instance). And the absolute best demo I ever heard at an audio show was tape: The Sugar Hill Gang's "Rapper's Delight" through giant MBL speakers. Where did the pristine sounding tape come from? "Friends in the studio"

 

[Julf]

9 Best Tape Emulation Plug-ins

 

[Salida]

Thanks for running this experiment. The one thing that really stands out is the large amount of modulation noise centered around the 1 kHz test tone. This signal dependent noise definitely contributes to the ‘sound’ of analog tape. We have gotten so used to it listening to recordings mastered on analog. Some prefer it.

 

[Avp1]

Thanks for running this experiment. The one thing that really stands out is the large amount of modulation noise centered around the 1 kHz test tone. This signal dependent noise definitely contributes to the ‘sound’ of analog tape. We have gotten so used to it listening to recordings mastered on analog. Some prefer it.

Surprisingly samples 2 and 4 did not looked like high-res at all despite having high sampling rate. They likely were up-sampled from 44.1kHz source somewhere in the chain.

Overall tape does seem to give better results than CD format, but technically inferior to real high-res. But some listeners simply like added tape distortion like those who prefer tube electronics to solid state.

 

[Julf]

Overall tape does seem to give better results than CD format, but technically inferior to real high-res.

What do you base that statement on? Clearly you believe "real high-res" (whatever that is) is somehow "better" than CD, but is the noise, compression and distortion of tape "better"?

 

[Avp1]

What do you base that statement on? Clearly you believe "real high-res" (whatever that is) is somehow "better" than CD, but is the noise, compression and distortion of tape "better"?

With proper tuning of machine you get:
1. Almost no compression in normal sound volume range with smooth compression above reference level, like one routinely used in microphone preamps in tracking.
2. Distortion is less than you find in tube electronics. Only 2nd and 3rd harmonics are visible, with no high order distortion. Low IMD numbers too.
3. CD format limitation of effective frequency range impacts transients. This is easy to hear if you know how it sounds. Tape has not such limitations. I mentioned MQA, which was supposed to overcome limits of low sampling rate in delivery media.

 

[Julf]

With proper tuning of machine you get:
1. Almost no compression in normal sound volume range with smooth compression above reference level, like one routinely used in microphone preamps in tracking.

While CD gives no compression at all. CD-tape 1-0.

2. Distortion is less than you find in tube electronics. Only 2nd and 3rd harmonics are visible, with no high order distortion. Low IMD numbers too.

Distortion is even less on CD. CD-tape 2-0.

3. CD format limitation of effective frequency range impacts transients.

No, it doesn't.

This is easy to hear if you know how it sounds.

No, it isn't.

Tape has not such limitations.

Yes, it has.

I mentioned MQA, which was supposed to overcome limits of low sampling rate in delivery media.

No, MQA was supposed to compensate for the lost income from MLP (as used by DVD-Audio, now a dead format). The technical claims were all disproved.

 

[GXAlan]

No, MQA was supposed to compensate for the lost income from MLP (as used by DVD-Audio, now a dead format). The technical claims were all disproved.

Dolby TrueHD uses MLP as its compression algorithm which is widely supported by Blu-Ray and UltraHD Blu-Rays

 

[Julf]

Dolby TrueHD uses MLP as its compression algorithm which is widely supported by Blu-Ray and UltraHD Blu-Rays

But the royalty income had been dropping pretty badly by the time they came up with MQA. Anyway, as we know, MQA is dead in the water, because the claims turned out to be bogus.

 

[Sal1950]

Anyway, as we know, MQA is dead in the water, because the claims turned out to be bogus.

We knew that all along...