LP with better dynamic range than digital

[Atmasphere]

When I first encountered this phenomena (about 1985) I have to say I was quite surprised! A friend of mine had the same cartridge and amps. We both had just bought a UHQR LP of the same title. His played noisy on his system, mine was quiet. He brought his LP over and it was quiet on my system too! The difference turned out to be in the phono section of the preamps involved.

When we did our first preamp (the MP-1) I made sure that this was not a problem for it! Over the years I've grown very used to people commenting about my system "Are you playing a CD? Its so quiet! But I just saw you put on an LP. Where is your CD player??"

They were not used to not hearing ticks and pops. FWIW I hardly ever clean my LPs with anything more than a $20 carbon fiber dust brush.

 

[ack]

Over the years I've grown very used to people commenting about my system "Are you playing a CD? Its so quiet! But I just saw you put on an LP. Where is your CD player??"

Another funny coincidence... I put an LP the other day on my platter, volume at 12 o'clock (pretty loud), dropped the needle, switched to the phono input... and no sound. Damn it, what happened now; seconds go by, still no sound. I am getting aggravated at that point; while looking at cable connections, I see nothing wrong; I keep looking, fifteen seconds go by, and then there's the music blasting out. What? So silent lead-in grooves? Well, yes, virtual digital silence; and it fooled me.

 

[Orb]

Orb

What is your position on this debate? Simply trying to get a handle on your posts.

in your opinion Does LP, please take the best example you can has a better dynamic range than CD (take the best example you can/have or have experienced). Real world implementation here .

The problem is this thread is debating many aspects and compounding it is also theory and real world and the broadness of "digital" (some will consider CD, some 16-bit theory, some 24-bit, some DSD, some sampling rates... this complicates any position/context).
A person should not have just one position in this thread because there are strengths and weaknesses depending upon which aspect one discusses, and importantly whether technical theory or real world and context whether discussing the technology-solution as a whole (so ADC-recording-DAW/mixing/editing-mastering) or very specific scenarios.
The best and that means IMO limited number of LP batches/releases is exceptional but CDs can be very good, however it is easier to find consistently good CDs compared to LPs, and hirez done properly (again the best hirez is not as common as one would expect) has benefits over CD in terms of real world applied dynamic range and greater flexibility with more/"better" filter implementations (covered this at length much earlier and in other threads) while also capable of being exceptional as some measured-reviewed albums has showns.
Note hirez has more potential to be fubar in real world when considering what is actually released and this has also been shown in measured-reviewed albums.

Sorry if this is not the answer you were looking for, but if it helps I spend most of my time these days listening to CD rather than LP (best white label batches put me off the more "normal" LP releases and did not want to restrict myself to just the best batch LPs I used to receive) or hirez (I find this can also change my listening habit being less satisfied with CD that is vast majority of my listening due to my broad tastes/tasteless hehe).
That said I appreciate fair few will take a different perspective to me.

Cheers
Orb

 

[thedudeabides]

So I have several unopened Sheffields from my analogue days. If I remember correctly, they were pretty quiet.

Is there still a market for these D2D LP's?

 

[esldude]

The lacquers was part of those debates between those talking theory/potential, same way CDs do NOT make most of their dynamic range but does not stop the debate about how good dynamic range of CD is; again theory/potential.
When you mention recovering clipping/compression (dynamic) this must be applied to real world retail solutions, especially when I read your post.

BTW when you say reduce them a bit how did you do that and at what stage?
From my experience any tool recovering end product is pretty limited, I used Death Magnetic as a simple example (same as I used analogue Bat Out of Hell) to show that the recovery is limited either for a clipped waveform or compressed dynamics.

Cheers
Orb

Well if someone uses a medium badly it isn't the medium's fault. You can produce clipped CD's or CD's limited so hard it looks about the same. If one wishes to criticize the digital medium because it clips drastically (which it does) and praise analog because it can be over-driven with less ill effect, then you can say clipping is rare and not much of a problem with digital (which it isn't) at the initial recording phase. Digital recording only rarely results in clipping. And if not clipped it is clean. Now ridiculous processing that creates those problems later on is a different issue, and again not one necessarily endemic to the digital audio process.

As for how I fix clipping, I zoom in and redraw the flat topped bits. I simply round them and lower them a bit manually. That is a distortion, but it does away with the awful sound of something hard clipped. That is why I described it as tedious. If you have a recorded song with 500 instances of clipping then this isn't practical. Now I am referring to files from a live recording and not recordings purchased as a consumer. I don't spend time going in and fixing CDs or purchased music.

 

[Don Hills]

Speaking about hyperbole ....

If I've told you once, I've told you a thousand times: Don't exaggerate.

 

[Atmasphere]

...snip...
That said I appreciate fair few will take a different perspective to me.

Cheers
Orb

Actually, +1 on that!! I agree with the entire post.

Well if someone uses a medium badly it isn't the medium's fault. You can produce clipped CD's or CD's limited so hard it looks about the same. If one wishes to criticize the digital medium because it clips drastically (which it does) and praise analog because it can be over-driven with less ill effect, then you can say clipping is rare and not much of a problem with digital (which it isn't) at the initial recording phase. Digital recording only rarely results in clipping. And if not clipped it is clean. Now ridiculous processing that creates those problems later on is a different issue, and again not one necessarily endemic to the digital audio process.

Bingo. +1 on that too. Wish this site had a button for that.

 

[Soundproof]

(not to mention the higher pitch frequencies the LP can reach) and I am sure we can verify by measuring SPL.

Please elaborate as to what "higher pitch frequencies the LP can reach."

 

[ack]

Please elaborate as to what "higher pitch frequencies the LP can reach."

Simply put, the high notes of the violin in the LP extend much higher in frequency than the CD; in the same sections, the violin in the CD sounds softer and doesn't reach as high in the high frequencies. On top of that, the violin has more power in the midrange from the LP than the CD; both of these, make the violin more real in the LP.

 

[thedudeabides]

I hasten to ask but ask I must.

What reliable, objective, measurable data do you have to support this premise that is not recording / mixing / mastering / manufacturing / individual system dependent?

 

[ack]

I hasten to ask but ask I must.

What reliable, objective, measurable data do you have to support this premise that is not recording / mixing / mastering / manufacturing / individual system dependent?

What does this even mean. It's the same recording, mastering should be at its best from FIM/LIM and from the master tape; ditto for XRCD manufacturing; mixing? System dependent??!??! Show me data that any of the above IS a factor. I picked these recordings as an example exactly because all relevant parameters should be optimal - I don't think this is hard to tell.

 

[thedudeabides]

Oh well.

Nuff said.

 

[Soundproof]

Simply put, the high notes of the violin in the LP extend much higher in frequency than the CD; in the same sections, the violin in the CD sounds softer and doesn't reach as high in the high frequencies. On top of that, the violin has more power in the midrange from the LP than the CD; both of these, make the violin more real in the LP.

There's an oft repeated canard about the higher frequencies of LPs, that results from the use of a guide-track for 4-channel playback from LPs. That guide track was laid down with half speed mastering (enabling the cutting needle to score the higher frequencies.) But that guide track has nothing to do with the music signal, it's information for the 4-channel decoder, and placed far above human hearing range.
But this has been used as evidence that LPs carry high-frequency information.

You'd be surprised if you had a look at the frequency spectrum of any LP in your collection. The best, late generation Neumann cutters, using half-speed mastering, managed to put down high frequency components of a music signal up to 25kHz, but most LPs have very little above 16kHz; and checking treasured Mercury recordings for their frequency spectrum should be a revelation.

 

[ack]

There's an oft repeated canard about the higher frequencies of LPs, that results from the use of a guide-track for 4-channel playback from LPs. That guide track was laid down with half speed mastering (enabling the cutting needle to score the higher frequencies.) But that guide track has nothing to do with the music signal, it's information for the 4-channel decoder, and placed far above human hearing range.
But this has been used as evidence that LPs carry high-frequency information.

You'd be surprised if you had a look at the frequency spectrum of any LP in your collection. The best, late generation Neumann cutters, using half-speed mastering, managed to put down high frequency components of a music signal up to 25kHz, but most LPs have very little above 16kHz; and checking treasured Mercury recordings for their frequency spectrum should be a revelation.

I am not sure how to interpret this... this particular LP (and all in my 600-LP collection) are two-channel, not 4 - I assume you are talking about quadraphonic LPs (for which a Shibata stylus would be best)??? Or are you referring to the master recording being 4-channel, and if it is, I am not clear what bearing that would have on the LP (allegedly it being 2-channel). I can't look at my collection's high frequency spectrum, and all I can tell you is that this violin coming out of this LP is more real than the CD, especially in the high frequencies; my reference in this case is a neighbor's treasured Strad - see http://www.whatsbestforum.com/showthread.php?10407-Borromeo-Quarter and bottom of http://www.borromeoquartet.org/artist.php?view=bio&bid=900

Thanks

 

[Atmasphere]

Please elaborate as to what "higher pitch frequencies the LP can reach."

Any cutterhead can do 30KHz no worries, and any magnetic cartridge and cheap phono equalizer can easily pass the signal. At least any I have seen. I'm pretty sure this was covered earlier in this thread.

The reason you don't see bandwidth much past 16-20KHz has to do with the devices used to make the recording, not the LP process itself. Tape, mixers, microphones- all impose bandwidth limits except in rare instances.

Our cutter system is a Westerex 3D which is an older technology compared to most of the Neumanns. But 30KHz is no problem. This is the type of cutterhead used to cut most of the RCA Living Stereo LPs from the golden age of stereo. So that bandwidth has been around since the dawn of the stereo LP.

 

[Orb]

Any cutterhead can do 30KHz no worries, and any magnetic cartridge and cheap phono equalizer can easily pass the signal. At least any I have seen. I'm pretty sure this was covered earlier in this thread.

The reason you don't see bandwidth much past 16-20KHz has to do with the devices used to make the recording, not the LP process itself. Tape, mixers, microphones- all impose bandwidth limits except in rare instances.

Our cutter system is a Westerex 3D which is an older technology compared to most of the Neumanns. But 30KHz is no problem. This is the type of cutterhead used to cut most of the RCA Living Stereo LPs from the golden age of stereo. So that bandwidth has been around since the dawn of the stereo LP.

From what I understand the challenge is high energy lowest musical frequency range, especially if it also involves panning one side to the other with regards to it as a stereo image.
I think this is one of the challenges faced by some mastering studios when dealing with modern dance/pop genre musicians who are used to only recording-listening with digital.

Cheers
Orb

 

[Atmasphere]

If you mean bass, it is true one has to be careful during the mastering process. If the bass is panning from side to side that is not so bad. But if the bass is out of phase, as can easily happen if several tracks not recorded at the same time are involved, then you can have real trouble keeping the stylus in the groove. We take it on a case by case basis to avoid processing, but the worst case is that the bass will have to go to mono for a few microseconds until the worst is past. Bass is really omni-directional below a certain frequency (somewhat room dependent) so this is not a huge problem. But is is for sure one of the limitations of the LP that digital lacks entirely!

Thing is, if the recording engineer is paying attention, he will work to make sure those out of phase events are kept to a minimum in a pop recording- you wind up with a better sounding result, regardless of the format.

 

[ack]

I am just getting back to this, to round up the following points made a while ago:

ack: "at near infinitely small degrees of fineness" yes, but still discrete. I don't see anyone presenting any proof that "infinitely small" is the same as non-discrete
esldude: I said near infinite though the main point is the degree of fineness is well, well below the LSB
Orb: Yes so it is actually infinite precision rather than infinite amplitude resolution (two different things), and that is mostly theoretical rather than real world IMO
esldude: As for proof what would you accept as proof that levels of signal can be portrayed at differences below one LSB?
ack: I am looking for mathematical proof

esldude:

http://www.robertwannamaker.com/writings/rw_phd.pdf

You can try this. I have only read a few select sections related to this topic. The conclusion may be enough or not.

http://www.users.qwest.net/~volt42/cadenzarecording/DitherExplained.pdf

Shorter version with some illustrations though no real math.

So I read Dr. Wannamaker's thesis mathematically proving the effectiveness of dither, and we have to assume it's been scientifically vetted; in fact, both links/articles are about dither (and I have read the latter many times before), and they are both in agreement (where they overlap). The high-level take-aways from the dissertation are as follows:

1) First and foremost, once more, "we" prove the effectiveness of dither in lowering quantization noise. Moreover, nowhere is it claimed or mathematically proven that digital is non-discrete or pure analog. Therefore, in our prior discussion points quoted above, all claims (from me, eslddude and Orb) are true

2) The following quotes from the dissertation summarize some things very well (emphases are mine):

Undithered quantization can produce audibly deleterious distortion and noise modulation in audio signals, indicating that the mean and variance of the quantization error signal are signal dependent. It will be seen that the use of dither can eliminate such input dependences, yielding an audibly preferable error signal which is perceptually equivalent to a signal-independent random noise.

Analogue-to-digital conversion is customarily decomposed into two separate processes: time sampling of the input analogue waveform and amplitude quantization of the signal values in order that the samples may be represented by binary words of a prescribed length. The order of these two processes is immaterial in theory, although in practice quantization is usually second. The sampling operation incurs no loss of information as long as the input is bandlimited in accordance with the Sampling Theorem [1], but the approximating nature of the quantization operation always results in signal degradation. ... An optimal (re)quantizer is one which minimizes the deleterious effects of the aforementioned signal degradation by converting the signal-dependent artifacts into benign signal-independent ones as far as possible.

Quantization and requantization possess similar “staircase” transfer characteristics ... The step size, ?, is commonly referred to as a least significant bit (LSB), since a change in input signal level of one step width corresponds to a change in the LSB of binary-coded output. Quantization or requantization introduces into the digital data stream an error signal, q, which is simply the difference between the output of the quantizer ... It has a maximum magnitude of 0.5 LSB and is periodic in w with a period of 1 LSB.

The objective of dithering is to control the statistical properties of the total error and its relationship to the system input. In undithered systems, we know that the error is a deterministic function of the input. If the input is simple and/or comparable in magnitude to the quantization step size, the total error signal is strongly input-dependent and audible as gross distortion. We shall see that use of dither with proper statistical properties can render the total error signal audibly equivalent to a steady noise floor.

And my favorite part of it, regarding the _subjective_ audibility of at least one dither algorithm [ack: for clarification, the dissertation researches "subtractive dithering" - SD - and "non-subtractive dithering" - NSD - systems]:

In addition, the corresponding total error signals (output minus input) were used in listening tests in order to check for any audible dependences on the input. Using undithered quantizers resulted in clearly audible distortion
and noise modulation in the output and error signals. A subtractively dithered quantizing system using iid 1RPDF dither was found to eliminate all audible input dependences in the error signal, which was confirmed to be audibly equivalent to
a steady white noise. A non-subtractively dithered quantizing system using the same dither eliminated all distortion, but the residual noise level was found to vary audibly in an input-dependent fashion. When 2RPDF [ack: 2RPDF ~= TPDF or Triangular Probability Density Function, which, as Orb has said, is more desirable and commonly used] dither was employed, no instance was found in which the error was audibly distinguishable from a steady white noise entirely unrelated with the input, although the level of this noise was, of course, somewhat higher than that observed in the subtractively-dithered system. Admittedly, these tests were informal, and there remains a need for formal psychoacoustic tests of this sort involving many participants under carefully controlled conditions.

Hope this helps everyone, and thanks to esldude and Orb for contributing. Now back to dynamic range...

 

[Orb]

Ack nice going.
Just to add if interested; (as one the quotes you use show) I used the term in theory because real world has to deal with quantization and still just below lsb error as the real world cannot create the ideal transfer function; this is one area the term stair/step/case can be applied to digital but context being technical function concept and not what one would see/hear (stair-step that may be seen/"heard" in digital would relate to alias imaging and classically seen with NOS/no alias filter so very specific instances that is not applicable to the majority of audio digital listeners).
If digging deeper on what you found can do search on digital with transfer function/quantization/ADC - go beyond wiki if it really interests you and helps to show where coming from with infinite precision in theory (and closer in some scientific scenarios than used in audio) but not in typical real world.
Key emphasis differentiating between the ideal transfer function and practical-real world transfer function with dither.

Cheers
Orb

 

[ack]

This is a good place to read about dynamic range in digital, and also a lesson on digital. And how digital can meet and exceed analog in SNR, and of course that includes LP. As if we have yet to see any proof that LP has more "real" dynamic range (that would be non distorted audio sinewaves or minimally distorted) than digital.

First, the Benchmark article says nothing specific about LP; only that "It can be shown mathematically that a properly-dithered digital system has the same resolution as an analog system having the same signal to noise ratio".

More importantly, there are misleading claims such as this:

Our ears have an amazing ability to hear sounds that are as much as 30 dB lower in amplitude than the noise around us. If we are listening to a properly-dithered 16-bit system, it is possible to hear musical tones that are 30 dB lower
than the noise (or 30+93=123 dB below full scale).

Notice the fallacy: they claim hearing 30dB into the noise "around us", then they add that to CD's practical noise floor (according to them) of -93dB; as if the "noise around us" is as low as -93dB. Nope.

Therefore, I request credible references to: a) research confirming us hearing 30dB into the noise AT ANY noise threshold (clearly impossible, thus obviously a silly over-simplification in the claim); b) mathematical proof that a 16-bit system can encode signals below its theoretical best noise threshold of -96dB (or even -93dB) so that we can possibly hear into that noise threshold. Either way, 123dB claim is just baseless. Or if you think the theoretical best noise floor in 16-bit systems is below -96dB, let's prove that (see below).

Related to question (b), the article claims this:

Noise shaping is now almost always used to master 16-bit CD’s...CD’s are restricted to a 44.1 kHz sample rate and therefore they have a very small attic in which to hide junk. Noise must be moved into the rather narrow region between 18 kHz and 22 kHz. CD’s have a 4 kHz band in which to hide noise. The bottom line is that a noise shaped 16-bit CD system can rival the performance of a 44.1 kHz 20-bit system that lacks noise-shaping. Properly dithered and noise-shaped CD recordings have the ability to audibly reproduce tones that are in excess of 140 dB below full scale. Because of the noise shaping, these 16-bit recordings can sound like they have a 120 dB SNR. I Thus, no one had yet shown proof that CD has higher dynamic range than LP either.

Three problems: 1) the underlined is unsubstantiated - can we prove it mathematically? 2) by pushing noise up in the 18-22kHz range with noise-shaping (e.g. UV22), they conveniently ignore that region as part of the audible range over which dynamic range is calculated - flat-out wrong; the audible range now becomes the *conveniently-squeezed* audible range 3) they make it sound like a free lunch, but higher noise in that region will inevitably affect the high frequencies in the music - and the highs in RBCD is something we have always complained about.

So far, no one has proven that CD's dynamic range is higher than the LP's either.