When you were recusing yourself for not being an EE, you were happy to cite
another man's criticism of digital audio filtering. His big paper on how
phase eq is needed in addition to frequency eq appears to ignore the
abundance of minimum phase digital audio plugins which have been
de-emphasizing the RIAA equalization for many years - without significant
error. Ask Glenn Meadows or Bob Ohlsson, if you doubt it.
On 4/11/12 2:10 PM, "Tom Fine" <[log in to unmask]> wrote:
> Hi Andrew:
> Any scientific measurement of bass input vs. output shows that 88.2 works just
> fine for
> analog-digital transfers.
I am paraphrasing Dan Lavry. It is he who insists that higher than 55 kHz
Fs is where inaccuracy creeps in. It doesn't flood in. But the sweet spot
appears to be there. Like the human heart, which requires a chemical dose
between each beat in order to get adequate rest, since, otherwise, it is
perpetually busy, the chips in converters need time to settle between pulses
in order not to introduce timing errors.
> But, as I often say, horses for courses. If you like
> 55kHz sampling rate,
> enjoy. If you do work for anyone who's getting grant funding, you probably
> have to adhere to NARAS
> specifications and use 96kHz.
I agree that 96 kHz has benefits - at least 88.2 does, anyway. DSP sounds
better at higher than necessary sampling rates. If all signal processing
could be done in the analog domain, then 24/48 would be pretty good for the
> The rest of your posting seems to be a slag of PCM in general, that it's
> allegedly not reproducing
> some sort of "automagic" (your word), which is fine for an opinion but I
> choose not to argue over
The automagic is that there is harmonically-enriched saturation, and freebie
compression with analog tape, and a preconditioning of the signal for
speaker wiggles before playback, caused by the plastic reluctance of the
lacquer on the "acetate" disc. In the case of the latter, compatibility
processing needs to be applied which further dumbs down the onslaught of
treble, due to restricting the groove velocity for the playback stylus
radius of curvature and to protect the cutter head from the excessive
currents of strong treble. However, in digital audio, anything goes - no
signal under 1/2 the sample rate is too pathological to encode. This means
that it's too easy for producers to get used to a glaring sound and not hear
it that way until the work is compared to a phonograph.
> Digital denial is fine, but the rest of the world has moved on.
I accept that digital audio is here. But no one has really gone anywhere.
Earth-bound pre-dead skeletons are stuck with both analog and digital, now.
Digital sounds great for speech.. It's just that analog is better at
summing traffic-jammed overtones. If the program doesn't need that kind of
fidelity, or the producer would rather trade some of the musical setback for
some of the good of digital, such as simple portability and immunity from
generational loss when copying, I recommend digital audio. I even
specialize in CD premastering, so I haven't quite turned my back on digital.
However, I don't think it's better than analog, in many ways, and I like
analog better just for being intimately physical in operation - it keeps in
touch with the continuous and irrationally bespeckled line (e.g., precise
location of root2), rather than the integer.
> My opinion
> is that it's ignorant
> to dismiss a technology because of flaws -- real or perceived -- in
> implementing it (or using it,
> for that matter).
As for what is wise, if a technology has flaws, it would be ignorant not to
dismiss it. You'd have to turn a blind eye not to see the havoc done to a
10 kHz square wave after encoding to 24/88.2 and decoding back to analog.
Check this out:
Just the 10k and the first harmonic appear barely to be salvaged.
As for the clicks and pops on vinyl pressings, not only have I just espied
Cat Stevens' Tea for the Tillerman sputtering sibilance and high treble in
the square wave zone of my scope (the 20, 10, 6 µs, and higher settings of
the Tek), the clicks and pops were also visible up there. So, in your
digital audio capture, like mine, there was steep filtering and smoothing
before hand-back. This is not bad-sounding down in fundamental land, but it
makes for a crappy sounding impulse. I find vinyl clicks in the digital
domain horrifying. Whereas on the tt, through all analog, it's easier to
tune out. I wonder if there isn't more to it than just the expectation that
CDs are supposed to sound clean and surface noise-free.
One can use a technology he dismisses as sub-standard, after all. Isn't
that what psychoacoustics is all about? (:
> I can't imagine a frame of reference where a mass duped cassette sounds
> "better" than a CD (assuming
> same master source and competent production of both media). Some seem to
> prefer added hiss and
> distortion as "euphonic," so I can imagine a frame of reference where a
> cassette recorded on a good
> cassette deck from a favorite LP would be preferred to a CD-resolution A-D
> transfer of the same LP,
> although I certainly don't share that frame of reference.
It's not that tape sounds better. It's that digital reveals too much that
is in-band and conceals too much that is out.
> -- Tom Fine
> ----- Original Message -----
> From: "Andrew Hamilton" <[log in to unmask]>
> To: <[log in to unmask]>
> Sent: Wednesday, April 11, 2012 1:40 PM
> Subject: Re: [ARSCLIST] Recording_78rpm_records
> On 4/11/12 12:14 PM, "Tom Fine" <[log in to unmask]> wrote:
>> To prove Karl's point, and I am sure this experience is not unique, I
>> encountered a stereo LP this
>> morning that had the exact kind of vinyl "pimple" described previously. The
>> pops and ticks
>> transferred just fine at 88.2/24-bit PCM, despite being very sharp rise times
>> and dropoffs. They are
>> definitely not square waves, but they rise quickly between samples. The
>> waveform editor did not show
>> any "ringing" at the top of the rise, and the dropoff was uniform also, so I
>> conclude that PCM
>> sampling captures these as they are produced by the analog equipment.
> I hope it's understood that what I wrote before in this thread was, "...the
> pops and clicks are near square waves, in terms of their rise times..." -
> not that a click or pop = square wave. They propagate suddenly. The clicks
> which propagate more suddenly than 3.6 µs, when sampled at 88.2 kHz Fs, are
> simply not encoded - whence the absence of perturbation. Still, that's a
> nice fast sample rate which, according to Dan Lavry, is technically too fast
> for perfect accuracy of the bass. 55 kHz is the current sampling speed
> limit, according to the designer. 88.2 kHz is good for this test, however,
> since we are talking about high frequency events. Still, the music may be
> best at 48 kHz Fs, in spite of the worse Squarepushing. It was suggested to
> me that program with only a few instruments (chamber music or solos?) would
> be best at 48 kHz, and music more rich in overtones would possibly be better
> at 88.2 kHz.
>> It is
>> also clear that the
>> needle is not leaving the groove because the pops "ride" on top of the music
>> waveform and thus can
>> easily be edited out with the pencil tool. When they occur during pauses in
>> the music, they are
>> clearly "riding" on top of the surface noise and rumble. This can easily be
>> seen by zooming in far
>> enough in the waveform editor. It was reasonable ambient humidity and the
>> vinyl was not carrying a
>> bad static charge, so I didn't encounter any of the saturation-pops to
>> analize, but I have done that
>> previously and stand by my previous description, adding that those, too, are
>> not square waves.
> Here's what I sent to one of the thread responders about the term, square
> wave, as I used it, earlier, in this thread. Perhaps we can get some kind
> of "rise" out of it:
> " The square wave's rise time is what was described as happening in clicks
> and pops in this thread. Not [actual] brief, but perfect, squares. Yes,
> impulse-like. Not pure impulses, of course, which would happen in less than
> a measurable length of time... (instantaneous propagation)
> Square waves - which are sudden - are created by summing many frequencies
> together, over many octaves. Square waves are meant nearly to be forming
> with the summed overtones of a full orchestra hit, according to my
> protected, and infallible, source.
> Digital audio sampling at, say, 96 kHz has radical attenuation above 48 kHz.
> This means that normal LPCM release formats can't "approach" the square wave
> as well as can its analog counterparts. Faster than 48 kHz Fs LPCM, also,
> is less accurate with the low frequency audio content, due to settling time
> requirements being exceeded (according to Lavry, who don't offer rates above
> 96 kHz Fs in their high end professional equipment). Once you go to DSD,
> there's too much hf noise that needs to be filtered for square wave
> bandwidths to be cleanly decoded.
> There's an urban legend going around which might be true that Stan Ricker
> found he had recorded the tape recorder bias signal (over 100k) on a lacquer
> disc. It was very quiet, but not silenced. That would approach DXD
> bandwidth, albeit quietly...
>> Karl is correct that no equipment, digital or analog, can reproduce all
>> waves and impulses,
>> but good ones can handle that information within the bandwidth of human
>> hearing. And, in defense of
>> the CD format, if 44.1/16 were so limited in the ability to capture, process
>> and reproduce square
>> waves and impulses, then almost all synthesizers would sound like total crap
>> on CD, and
>> "electronica" as a music genre would have never risen up in the CD era.
> CD alias-dampening is not by any means a deal-breaker. It's a question of
> high fidelity, only.
>> same can be said for
>> highyl-synthesized genres like hip-hop, 80's pop-rock and post-disco dance
>> music, not to mention
>> movie soundtracks. Yes, there is more "there" there with higher-resolution
>> digital, but there's
>> a lot of "there" with the Red Book resolution and it can jump out at you
>> better without tape hiss or
>> LP surface noise and rumble competing with it.
> Too much of the wrong, "there," and not enough of the automagic
> enhancements, as well as the ability "to approach" the square wave with
> derived integrity.
>> Given that, by the time CDs
>> took hold as a mass
>> medium, the most popular mass medium for music distribution was fast-duped
>> cassettes, the Compact
>> Disc was a great leap forward
> In certain measurable ways, LPCM is like an elegant advanced alien audio
> technology. In others, it is like Big Brother's electronic harassment for
> crowd dispersal. Depends on where you place the probe. I agree that most
> CDs sound clean (in an antiseptic and glassy or white noise, sort of a way)
> . But I listened to those old cassettes so many times the rust was
> practically rubbed off. You can have most of the CD's I've bought over the
> last 30 years. What shrill junk. (: Rhetorically offered, only - I might
> need them for critical reference, still., (l;
>> (cassettes outsold LPs soon after the Walkman
>> Revolution took hold;
>> CDs didn't outsell cassettes until the early 1990s -- see numerous references
>> and industry data
>> available online).
> OK, but sales are one thing and good sound may be something else. People
> bought those Blue Amerols, since that's what was up, and it was once "new."
> Ringtones are cooler, nowadays, than CDs, aren't they. The Target® depends
> on the shopper.
>> -- Tom Fine
>> ----- Original Message -----
>> From: "Karl Fitzke" <[log in to unmask]>
>> To: <[log in to unmask]>
>> Sent: Wednesday, April 11, 2012 11:20 AM
>> Subject: Re: [ARSCLIST] Recording_78rpm_records
>>> Dale's question (or I suspect, point made in the form of a question) is
>>> on the mark.
>>> All electronic systems, analog or digital, have their limits. The ability
>>> record, reproduce,
>>> process, square waves or short impulses for example, depends on a system's
>>> high frequency response
>>> and dynamic range. There is NO technology out there that can record a pure
>>> square wave or
>>> infinitely short impulse. Some systems are of course better than others at
>>> recording audio band
>>> square waves, impulses, and complex waveforms with important high frequency
>>> content. But whether
>>> they are digital or analog is not the determining factor.
>>> It's a mistake to close our minds to digital audio, and its most prominently
>>> employed sampling
>>> methodology, Linear Pulse Code Modulation (LPCM), just because we might be
>>> dissatisfied with CD
>>> quality digital audio (16-bit 44.1 kHz sample rate LPCM).
>>> On 4/11/12 7:04 AM, D. Allen wrote:
>>>> Dale Francis
>>>> On Apr 11, 2012, at 2:05, "Andrew S. Hamilton"<[log in to unmask]> wrote:
>>>>> It's that the pops and clicks are near square waves, in terms of their
>>>>> rise times, and digital
>>>>> can only capture sines. So, it changes the attack and harmonics of the
>>>>> surface noise and makes
>>>>> it ugly, whereas the same noises on the vinyl are easy to ignore and might
>>>>> even be exciting>?<
>>>>> Incidentally, he said this is also why digital fails to deliver,
>>>>> even though it does
>>>>> great telephony, since the combined harmonic overtones of a full orchestra
>>>>> during a fortissimo
>>>>> would approach a square wave on an oscilloscope, but the CD can't make
>>>>> happen. So, for
>>>>> all its accuracy and quietude, its just a stomp box (LPCM).
>>>> Can either a phono cartridge or a speaker can accurately encounter a square
>>>> wave without side
>>>> issues ... needle leaving groove ...
>>> Karl Fitzke
>>> Audio Engineer
>>> Macaulay Library
>>> Cornell Lab of Ornithology
>>> 159 Sapsucker Woods Road
>>> Ithaca, NY 14850
>>> [log in to unmask]
>>> Our Mission:
>>> To interpret and conserve the Earth's biological diversity through research,
>>> education, and
>>> citizen science focused on birds.
Serif Sound ♬ CD Premastering
➣ Dingbat Lacquer Sound Disc
2836 Spring Grove Avenue