The Big Mystery: Analog Tape - why does it sound so good?

My deck has been described as semi-pro. Certainly, it has many 'pro' features. If I execute the recording very well, the results are quite good and that 'execution' is very important to the outcome...

My goal is acquire a true 'pro' deck with 15ips capability.

Bob

Here is my deck...

 

Well, the results you are getting with it are truly stunning!

 

This thread is all about having a *preference* for a certain recording and playback medium. The majority of the posts here seems to prefer analog because of that "warm" and "yummy" sound it delivers. As a self-confessed vinyl fetichist, there is *nothing* wrong with that. Nobody denies that analog masters sound "different" from digital. Maybe this is partly due to the soft clipping effects as the tape goes into saturation and the higher noise floor of analog which can add some spaciousness to the sound. Kinda like an aural exciter. There is no denying that analog can delivery the goods as has been demonstrated by really great recordings.

Since audio is physics, it means that it is measurable. And in the numbers game, digital has the edge. Preference aside, it is a fact that digital produces a more *accurate* reproduction of the signal. Consider this quote from Jeff:

The end result of Jeff's DAD transfer captured the analog "yumminess" that he felt was missing from the "cold" CD source. What does this tell you about the transparency of digital compared to analog which somehow *changed* the sound. In Jeff's opinion, this change was for the better, but the actual fact that it *did* and that the final transfer back to digital *captured* this change does speak for itself.

So why is such a great medium as digital so misunderstood. Consider a statement like this from UncleAl:

The magnetic aligment of domains modulated by an analog control voltage is not any more a different *recording* than a discrete time sampled signal as both of them are artificial representations of an air pressure function. An analog system has the same constraints of discontinuity, within a given bandwidth, as an equivalent digital system. So whether it is a digital or analog recording we are in effect approximating the information given the limits of their resolution.

Every signal that is recorded is nothing more than a long sequence of voltages along a time line. On CD, the voltages appear as a series of numbers, but if you look at the output of a properly-functioning DAC with a high-bandwidth oscilloscope, you will see that it is a continuous wave indistinguishable from the comparable analog signal. The output of a DAC is *not* a series of samples.

Markus, since you just mention digital in one word, I assume you are referring to both Redbook and hi-res formats? If this is correct then it is a rather bold statement to make. I hope you can back this up with solid evidence.

A properly-functioning digital system reproduces *all* of the information within the band-limited spectrum. There are no "holes" in the information. PERIOD. This is not a debatable point. A properly-functioning analog system with the same band-limitations will reproduce exactly as much information. However, based on comments in this thread, there seems to be one caveat, it is not as transparent as digital considering the "yumminess" that seems to be added to the original signal during the recording process. However, to equate this "yumminess" with "more resolution" is, again, pretty bold.

Markus, considering that the noise floor with even just 16 bit digital is lower than it is for analog tape, please provide the evidence to back up your claim.

________
Thom

 

The only objective information I have seen relates to this. Tape hiss, even when very low, adds an element to the sound that the brain often interprets as additional sonic information, more air, more high end perception. But tape is also far from linear in frequency response, unless great care is taken to calibrate the recorder to the tape and to play it back on a properly aligned and set up player. It almost has so many fudge factors that a calm and rational discussion can't get much traction.

Richard.

 

Richard,

Important points when recording on Analog tape - paramount!

Thom,

Regardless of science, measurements, tests and data, I use my ears as the ultimate judge of sound quality. If the sound 'moves' me, I consider the sound quality to be excellent regardless of what domain the music has been recorded, mixed and mastered in. More often than not, that domain is ultimately Analog. For me, music is an evocative emotional experience and the impressions prevail over other considerations. I see CD's as a wonderful playback medium. The best sounding CD's that I own were executed with all the processing in Analog and then transferred to Digital. My view of music and sound is a 'purely' human one....

Bob

 

Bob, I agree with the emotive part. And how we reach that state of emotion is highly subjective.

For me it is the music rather than the delivery medium that I react to. Which means that I get goosebumps by listening to, for example, the ending of Stravinsky's Firebird whether it comes from an AAA or a DDD delivery format, or an audience recorded bootleg tape of a Zappa guitar solo.

In my post I was not knocking anybody's listening preferences. I was simply trying to explain what I feel were misrepresentations of digital technology, which when used correctly, IMO can exhibit wonderful results.

_______
Thom

 

Thom, I read you loud and clear!

 

The whole discussion is ironic in a way. It seems that many fans of analog are trying very hard to "prove" that analog is technically superior to digital ("analog holds more information," "digital sampling creates holes," "the Nyquist theorem can't be true," etc.) to justify their preference. Why is that necessary?

It seems clear to me that the qualities of analog that so many here enjoy -- the "air," the sense of "fullness," the "warmth" -- are due to properties that can be considered "flaws" in a strictly technical sense -- surface noise, channel crosstalk, and other euphonic distortions. So what? Why is that such a problem?

In other words...

Why does analog have to beat digital objectively in order for it to be OK to prefer it subjectively?

Ryan

 

Thom - great post!

So, whatever the "it" is, be it tape hiss, harmonic distortion, non-linear frequency response curve (heck, probably all of the above) - the "tape sound" is what I like. And I do agree that transfering it back to digital appears to capture most (if not all) of that "tape added magic".

Thanks again for your thoughtful response!

jeff

 

For me, it isn't. I like them both. Both can sound great, or horrid.

It's just that with digital, garabage in WILL BE garbage out. But *sometimes* with tape, garbage in can yield a nifty-sounding out.

I agree - it doesn't have to. They're just different.

I guess ultimitely, my question is really about using tape as an effect. If one could figure out *exactly* what tape is doing, one could make a black box filter out of it, and add it to your preamp. Does a recording sounds cold or lifeless? Hit the "tapelizer" button. Does it sound great as-is? Turn the effect off.

 

That hits the nail on the head. With digital (not digital *processing,* just digital as a storage medium), what you get out is pretty damn close to what you put in, at least within the bandwidth allowed by the sampling rate. Analog is a much more temperamental and mercurial medium -- that can mean it can give you (from a subjective perspective) something worse, something about the same, or, from time to time, something better than what you put into it.

And that's a very interesting and potentially productive question. I wish I had a comprehensive answer.

My post was aimed more at the sidetracking into objectively bashing digital, which isn't really what your question was all about.

Ryan

 

I am like GortBob, in that I just go by what sound hits me the best, no matter how it comes.

 

Ahh... it's so refreshing to read such a calm and clear-headed "debate" about analog-vs.-digital and not see it get sidetracked with bias and temperment.

Ryan, that's one of the smartest and most concise statements I've ever come across regarding analog-vs.-digital. Coming from a scientific background, I tend to be objectivist about sound issues, but at the same time I know what I hear and I try to keep my mind open to those who hear much better than I do.

I also don't care whether something is digital or analog so long as it "sounds good."

 

The ears are the best way of detecting sound quality, but sometimes the numbers and technical info (waveforms) can be interesting so that you can show others what you have personally heard.

 

Thom,

Thanks, I felt bold when I posted that. Yes, I meant both Redbook and hi-res. No hard evidence, just some experience and decent pair of ears.

If "A properly-functioning digital system reproduces *all* of the information within the band-limited spectrum.", why do we need hi-rez? Isn't Redbook enough, then? I think DSD sounds close to analog, but maybe we need ultra-rez to get to the level of good analog.

Noise floor on analog tape and or on an LP vs. digital isn't the issue here, IMO.

 

1) To make the "band-limited spectrum" bigger. (Though redbook's limits are beyond the practical range of human hearing.)

2) To allow for greater dynamic range. (Though almost no one uses the dynamic range on redbook.)

3) To enable multichannel mixes.

4) To sell everybody something new.

Ryan

 

If you be so kind, let me first take this opportunity to rant a little bit.

The are areas in audio, like bandwidth, dynamic range/resolution, that we understand very well and have done so for nearly 70 years. Improvements in the reproduction of the aforementioned issues are something a lover of *hi-fidelity* should be grateful for. Not for the *subjective* enjoyment a piece of music gives the listener, as this can be achieved via a cheap AM-FM radio given the right circumstances, but rather as an *objective* yardstick of how *accurately* it is reproduced. Obi-Wan might tell Luke to "trust your ears, do not succumb to the numbers and be drawn to the
dark (cold) side." And Obi-Wan has a point there. But aren't you at least a bit curious to discover *why* something works or does *not* work for you? That is why a basic understanding of technical measurements (numbers) as it relates to audio is a *good* thing. Never *mind* that the numbers on a certain piece of equipment or recording does not subjectively* stack up in your favor when, *subjectively*, it moves you. An understanding of why should, IMO, be for the betterment of our hobby.

But I could be wrong.

Now....

Hard to argue with that kind of subjectivity.

Redbook *and* hi-rez (DSD, DVD-A) are *all* based on *digital* technology, hence the collective terminology "digital". Other than that, Ryan does put it short and sweet as opposed to my ramblings. <sigh>

Dynamic range is the ratio between the largest signal and the smallest unambiguously resolvable change. That *is* its defintion.

If you have two systems with equal bandwidth, where normalized maximum levels are the same, and the only difference is the lowest fundamental limit due to noise or quantization and dither, then no matter how the signal is represented, we can conclude that the system with the *lower* NOISE FLOOR has the *higher* resolution. And since the systems are normalized for gain and maximum level and bandwidth, we can also conclude that resolution is directly proportional to dynamic range. So, the wider the dynamic range of the system, the higher the resolution. Hence, two systems with identical dynamic range and bandwidth have identical resolution. It makes no difference what the encoding methodology is. Resolution, i.e. the amount of information stored, is a function of dynamic range and *not* the storage methodology.

Now since you were so bold as to include both Redbook and hi-rez digital in your comparison with analog tape and vinyl, let us look at those nerdy numbers:

SOTA digital: A minimum of 96dB (93dB with essential dither) up to over 140dB dynamic range and ruler flat response from DC to 96kHz. Not bad.

Since you claim we need "ultra-rez to get to the level of good analog", and by that I assume you mean better than what I outlined above, show me *any* SOTA analog audio storage medium/reproducer that betters this.





Oh, and Markus, even if you cannot find one? Don't worry and be happy. Why? Well, here's Steve:

And the one that I live by:

Yeah!

________
Thom

 

A very articulate post, Thom. I'm jealous. I'm not going to argue this anymore, but I still don't agree with you. You are putting me on the spot and I don't like it. I feel like I'm on trial here. I'm going to keep my opinions to myself from now on.

 

Nevermind...

 

Gentlemen,

We're gone way off track, and are on the edge of a "objective/subjective" debate.

Please read our policy on this:

http://www.stevehoffman.tv/forums/showthread.php?s=&threadid=11234

See how this alienates people? Ybe doesn't even feel he can speak his mind without being put on trial.

Lets move on, please. If you'd like to continue this discussion in private, please feel free to do so.

 

Well, I can say that in the car, I get more enjoyment out of playing a tape than a CD, even though I know they sound different.

 

I've read many times now where many pros and magazine reviewers I trust have said that the biggest reason that formats with higher sampling rates improve sound quality is not really because of the increased sampling rate, but instead, that the DAC used to reconstruct the 1's & 0's into an analog waveform can use much gentler filters. Here's a summation of what I have read:

The Compact Disc format uses what's called a "brickwall" filter to get rid of the PCM format's inherent noise. This filter is placed at @21kHz, right next door to the audible frequencies. This filter has to act quickly to keep the ultrasonic noise from entering those lower frequencies, which it does. But a ripple effect occurs from the noise slamming to a stop into this aggressive filter wall (figuratively speaking). This "ripple" bounces back into those lower frequencies, causing subtle but audible problems with cymbals, bells & other delicate high frequencies. And grain and other sonic nasties are born!

Very high quality DACs--as in pro-level equipment or high quality consumer components--have very sophisticated filters to minimise or prevent audible problems as decribed above. But they are quite expensive. And adding to this expense is the fact that many of these still use the multibit conversion system, the best type of digital-to-analog converter for PCM. Multibit types use 16 (or whatever the word length needed) precision resistors to provide the neccesary voltage steps needed to create the analog waveform; these resistors have to have very accurate chemical tolerances; additionally, they are physically trimmed by a laser as a final quality assurance. This all adds up to big $$$. This is why I personally think some professionals report little to no improvement using the 96/24 or 192/24 formats.

But consumer level equipment is a different story. Here's the short version of why higher sampling rates sound better through my and other people's "affordable" hi-res players:

The filters are much different. For example, the filter for 96/24 is set at @45kHz, quite far away from the audible frequencies. Because of this they can be much gentler & in turn cause less signal ripple (i.e. distortion). And as a side benefit, since they are gentler and further away, they don't need to be very sophisticated >>> they're cheap to build. And unless I'm remembering this wrong, us consumers get 1-bit converters for the digital-to-analog PCM conversion but the better filters make up for the lower quality of this type of converter (1-bit DACs work at very high speeds which generates more noise but they are much cheaper to build compared to the multibit types).

Conjecture: the extended frequency response of hi-res digital might also contribute additional natural harmonics to improve the audible sounds, but the experts can't seem to agree on this idea.

And: those extra bits (20 or 24 compared to 16) also contribute to better sound because these extra bits give the converter and other associated circuits "elbow room" if any mistakes are made when handling the signal. This is especially true during additional digital processing during the recording process in the studio (compression, special effects, EQ, etc).

The end result is that consumers get a cleaner and less futzed-with analog signal. And with up to four times as many samples (192kHz rate) we get bonus snapshots of the original sound--I've never heard yet this causes any problems.

[T]