Knowing when or not to buy vinyl

[Plush]

Originally Posted by ds11

Very cool! Have you digitally archived the master tape? at 96k? It would be great to hear both that and the vinyl A/B'd on the same sound system.

I am working on this LP now and yes I do work at 96 and have done for 17 years. I am making a compilation from many different sources. These will be played out to the Stellavox with ending fades and then the sequence edited on the 1/4."

To be mastered at Abbey Road ( all analog chain--on preview machine) by a classical heavy (just for fun and the tour!) and pressed at Quality Record Pressing in Kansas.

So after this is released, one could do the experiment you asked for above. Adjusting the two formats for the same realistic play back levels, it could be a good demo.

Mostly I'm doing this project because I started on vinyl and decades later want to have a vinyl record (sic) of some of my recording highlights. I'll press and sell about 500 copies.

[Plush]

I salute DC for offering his substantial knowledge in a contributing and polite manner. O to have the denizens of GS take a leaf from his book.

As for proposing that the LP has anything over basically 65 dB dynamic range, it is all a chimera. And the mang from the Golden Clam Machine is going to have to revise his assertion that various analog formats cannot be characterized for illustration purposes (and for comparative purposes) as a listing of bit accuracy.

Indeed the inventors of noise shaped dither do exactly that in the paper I cited. Those inventors are Craven, Gerzon and Stuart.

[ds11]

Sounds like a nice project. Hopefully you'll be able to offer some insight into how it compares with the digital archives.

[Paul Gold]

Originally Posted by Plush

As for proposing that the LP has anything over basically 65 dB dynamic range, it is all a chimera. And the mang from the Golden Clam Machine is going to have to revise his assertion that various analog formats cannot be characterized for illustration purposes (and for comparative purposes) as a listing of bit accuracy.

Comparing for illustrative purposes is fine as long as the parameters of comparison are clearly defined. I understand dynamic range. You wrote the following on page 1.

Recording scientists also know how poor the technical specifications for L P are in comparison to cds. Even the best LP presents a fidelity level of around 12 bit.

Perhaps as a scientist you can enlighten me on the best method to measure "fidelity level"?

[audioaddict]

Originally Posted by wado1942

On another note, I got a hold of Wasting Light on vinyl, digitized it at 24/96 and analyzed it. It is full of audible limiting and digital clipping, showing that it was taken from the same master that was used to make the CD. I'm just glad I didn't actually buy it, because I'd be writing Dave Grohl personally to demand my $30 back.

Oh man, very, very sad.

And you know what...even if the master tape was good, it's now gone.

What is there cannot be undone! LOL.

Thank you for saving the time and effort.

[audioaddict]

Originally Posted by Plush

As for proposing that the LP has anything over basically 65 dB dynamic range, it is all a chimera. And the mang from the Golden Clam Machine is going to have to revise his assertion that various analog formats cannot be characterized for illustration purposes (and for comparative purposes) as a listing of bit accuracy.

A word or two about the dynamic range of LP vs. CD vs. DVD-A/SACD vs. tape, etc...

Many like to focus on the #s, and if you look at it purely from that it appears that redbook CD and the high resolution formats destroy the analog formats.

However, from my listening experience I have yet to see (hear) that. In fact in most cases I find the more dynamic recordings to be either from LP or analog 1/4" 7.5/15 IPS tape (ok, cassette tape you got beat).

Now there can be the argument that the mastering on vinyl is usually better and if both are done "right", the digital will be better.....but even in those tough shootouts, I've never heard digital better. And even if it could be better, nothing along the lines of 20-30 dB better!

Personally, I feel what is missing in these discussions is the fact that in analog playback there is no D/A conversion. So I'd like to propose that the D/A process will take away from dynamic range. And if it isn't that, it must be something else....but what could it be??

If no dynamic range is lost by the #s....then I propose that you cannot compare those dBs across analog/digital domain, the same way you cannot equate 10 pounds to 10 kilograms or 10 miles to 10 kilometers.

20+ dB is a huge difference.....can anyone truly say they've heard it? Honestly??

As for why analog playback is more appealing, I offer up two other ideas:

a) Analog is pure music, digital is quantized/sampled analog
b) Rolled off highs and tape saturation

So for the first....I feel that's the most fundamental part. If music was inherently digital, maybe it would be king. But since it is not, it only seems fitting that analog should sound more natural. The same "losses" I propose in D/A, might very well just be the musicality.

For the second point, this is something that bothers pro guys but a good # of audiophiles don't care. Many pro guys (from what I can tell) only care about information, specs, and frequency response to 50 MHz (if it were possible, e). Select audiophiles, care about the warmth that comes with rolled off highs and accentuated midbass.

While there are some pro guys that might want to dismiss these viewpoints, keep in mind that most audiophiles that prefer digital do so more because there's less noise and perhaps more detail/spatial separation...and of course some like a brighter sound. :D

I've never heard any of them say the CD was more dynamic, more bass, or smoother sounding. Never, ever, ever.

The worst insult most will give is that the LP sounds the same as CD. Which in my experience I can relate to....some commercial blockbuster release that was probably cut from the commercial CD. They basically do sound the same except the LP is slightly smoother at the top end...maybe a tad bit fuller in the midrange, and that's it. And of course has noise.

The best vinyl & tape I've heard were from mostly analog sources...and the ambience/depth is just on another level that I've yet to hear from a digital source.

Hope that gives some new perspective. Arguing over bit rates is not going to be meaningful, IMHO.

[Paul Gold]

Originally Posted by audioaddict

Arguing over bit rates is not going to be meaningful, IMHO.

I am trying to separate meaningful and valid technical comparisons like dynamic range from meaningless mumbo jumbo like 'fidelity level of 12 bits'. I couldn't find the fidelity level graph in the AES paper.

The long and entirely subjective post above is fine because you make no attempt at being a recording scientist.

[wado1942]

Originally Posted by dcollins

No, it's pulse width modulation. You should read the paper I referenced. The authors actually call it "duty-cycle modulation" which is the same thing, but it's forgivable as they are Canadian.

Here is a picture of the actual waveforms, showing both the PWM waveform and the recovery of the signal after averaging.

DC

The width of the PCM signal can absolutely not vary. The width of each pulse is by necessity, fixed. The dither may cause more pulses to be in an "on" state or an "off" state in a random manner, even several of each in a row of like value, but the width of each pulse is constant.

[Plush]

Well, of course, fidelity level is another way to say what resolution level the
system operates at. In this case we are talking about bits. Perhaps not the best term to use, but that is what I meant.

[Paul Gold]

Originally Posted by Plush

Well, of course, fidelity level is another way to say what resolution level the
system operates at. In this case we are talking about bits. Perhaps not the best term to use, but that is what I meant.

By fidelity level or resolution level do you mean something other than dynamic range?

I understand how a 12 bit system has about the same dynamic range as a record. I understand that resolution refers to bit depth in a digital system.

Resolution sounds like a subjective term as applied to analog audio. What do you mean by resolution in addition to dynamic range?

[audioaddict]

Paul, would you be willing to entertain the idea that good sound whether it's good mixing, mastering, speaker design, amp design, etc. is as much an art as it is a science?

I ask this because if you claim everything was pure science, then we would all have the answers and everything would sound the same (or roughly the same).

For the record I'm a licensed Engineer, so I do have some vested interest in the "recording science" but have learned this is one area where your ears are as important if not more than the theory.

[Paul Gold]

Originally Posted by audioaddict

Paul, would you be willing to entertain the idea that good sound whether it's good mixing, mastering, speaker design, amp design, etc. is as much an art as it is a science?

Yes. Of course. I don't think I've commented at all on the sound of either format or which I prefer. Assuming I prefer one to the other.

My comments have been purely to separate subjective claims from meaningful technical comparisons. Saying a record has a dynamic range of about 65dB and a 12 bit digital system has a dynamic range of about 65dB is true. Saying a 12 bit system has the same "resolution" is meaningless unless you define resolution. It's subjective mumbo jumbo.

In the real world if I was handed a 12 bit converter I'd have to spend a lot of time being very very careful about low levels. Not so with a disk recording system. High levels are more dangerous. Why?

[Plush]

I'm keeping my commentary on the topic asked about by the OP. I don't really have a lot more to say since anyone reading can understand that I'm pointing out that the LP has some inherent noise. It does not bother me while playing a record although seldom do I play it much louder than 90dB.

For audioaddict who posted an interesting post, this noise and the LP's limited rate of data delivery are the causes of it's low dynamic range in comparison to modern digital.

The comparisons of LP and cd playback mentioned by audioaddict recall the Linn Sondek turntable company's challenge of the early 80's. Put on a LP and a cd, synchronize them the best you can and switch back and forth for a listening comparison. The LP usually won in the area of having the listener deem it more pleasant and less "canned / dessicated" sounding. Of course it didn't help the cd's cause that most of the early compact disc issues were bright and had awful conversion quality.

Today one can do the same test but we again won't hear too much difference probably since modern releases have so severely truncated dynamic range and come from the same master.

I don't any more wade in to discussions having to do with high end hi-fi or with audiophile issues since this is mostly the domain of the listener rather than the producer. However, one can be sure that instead of the d/a offering less dynamic range, it is offering the full dynamic range and everything else that the digital playback has. In fact, it is probably offering too much detail.
(and must be tamed)

In my opinion the comparison is not between cd and LP. (because those are the same program material.) The meaningful comparison (hopefully from analog tape) is between the 2 buss output of the analog console and the same program output from the d/a of the editing work station. Use a speaker switcher to go direct to the monitor system and do a meaningful comparison. It would be a rare case indeed where the d/a playback had more pleasant-ness, conveyed more emotion or was more true to the source.

Maybe the best way to buy vinyl is to do like in the old days where there was a listening booth in the record store. These days vinyl freaks gather at a friend's house to listen. Then hopefully they go out and buy the music sound and the music style they have authorized and endorsed.

[Cheebs Goat]

Just a quick response to the first post:

Go ahead and buy the crap outta' that Black Keys LP. It sounds great from my first hand experience.

[dcollins]

Originally Posted by wado1942

The width of the PCM signal can absolutely not vary. The width of each pulse is by necessity, fixed. The dither may cause more pulses to be in an "on" state or an "off" state in a random manner, even several of each in a row of like value, but the width of each pulse is constant.

I'm not sure you're understanding what is happening here. The addition of d*ther is allowing extra information to be encoded in the LSB that would otherwise be lost. Yes, digital bits have only have two states, but there is now enough energy to toggle the LSB.

Here is another view of the same thing: See the difference between the top and the bottom?

But you should really look at the paper.


DC

[wado1942]

I understand how dither works. You're talking about PWM, where the width of each pulse is varied depending on the signal that's combined with it, equating to varied amounts of output energy depending on the width of each pulse. That's not what happens with dither. The LSB is still always an an on/off state and the width is constant. Yes, the dither keeps them moving where the law of probability allows the embedded signal to average out to more values than what was previously available without dither. When there's no input signal, the LSB averages out to an equal number of on-off states. The input signal causes more pulses to be positive and more can be negative, but that's it, no variation in width. They are two different principles.

[Paul Gold]

Originally Posted by Plush

I don't really have a lot more to say since anyone reading can understand that I'm pointing out that the LP has some inherent noise. It does not bother me while playing a record although seldom do I play it much louder than 90dB.

I am done beating you up Plush. I would like to point out that the noise on an LP is very different than the noise floor made by dither of any variety, or tape for that matter. That is one of the reasons why drawing sweeping conclusions based on simple measurements is misleading. Both the noise components and the distortion mechanisms are completely different. This is where the sound differences lie.

For those paying attention I am actually advocating listening subjectively and using measurements to explain what you are hearing. Of course measuring a moving target like a record is not easy.

[Cheebs Goat]

Man, this is so much simpler than all of this mess.

I love listening to records. Nothing matters beyond that. Even if it is pure expectation bias, it still doesn't matter. I love listening to records.

[audioaddict]

Originally Posted by Cheebs Goat

Man, this is so much simpler than all of this mess.

I love listening to records. Nothing matters beyond that. Even if it is pure expectation bias, it still doesn't matter. I love listening to records.

Haha, good point.

Something else that I think has been missed in this discussion is the actual playback systems themselves.

Analog playback systems have more potential for improvement than digital. For digital you have the DAC and transport, and that's it. For vinyl you have the phono amp and the table comprising the motor, tonearm, cartridge, needle. More working parts that can be improved.

Good tape decks for the most part need very little or nothing to upgrade as they can beat out the best table/phono setups for not a lot of money.

So while it's interesting (and fun) to debate the best possible setups for each, the reality is that in the real world on most systems, pound for pound analog has a huge advantage. At our recent display at Capital Audio Fest, in one of our rooms we had a $1500 reel to reel deck, a $7000 table/phono setup, and a $70,000 DAC/CD transport combo (the DAC was borrowed from a customer as we had none in stock....you have to own everything you sell, no financing!).

And for the most part, the sound quality was about inverse to the cost. That Audio Note DAC 5 Signature is an amazing DAC....so good that you can almost forget about analog playback, because it's very "analog sounding". But then you hear a nice record or tape and you're reminded why real analog will always be here.

Under $3-4k, the battle is pretty close but the further up you go, the more analog separates itself from digital. A $10-15k vinyl setup is a lot more realistic to most hardcore audiophiles than a $90k digital setup.

[UnderTow]

Originally Posted by wado1942

I understand how dither works. You're talking about PWM, where the width of each pulse is varied depending on the signal that's combined with it, equating to varied amounts of output energy depending on the width of each pulse. That's not what happens with dither. The LSB is still always an an on/off state and the width is constant.

Dave didn't say the LSB is pulse-width modulated, he said the signal is pulse-width modulated. There was a strategically placed coma in his sentence.

I haven't read the paper yet so that is as far as I will comment on that but anyone interested in whether it is possible to encode signals below the LSB in PCM audio might want read the info and check out the example in post 4 of this thread:

Digital Audio and Sampling Rates


Alistair

[12th & Vine]

Originally Posted by Cheebs Goat

Man, this is so much simpler than all of this mess.

I love listening to records. Nothing matters beyond that. Even if it is pure expectation bias, it still doesn't matter. I love listening to records.

Yes ... and isn't it crazy that the subtleties of S/N and how many bits it would take etc somehow don't ever cross your mind once that stylus hits the groove of your fav rekkids?

[RiF]

LP covers smell way better than CD jewel cases.

[Thor]

Tape noise (i.e. the carrier's noise floor) is the analogue equivalent of digitally added dither to encode information below the LSB.

afaik.

Thor

Originally Posted by UnderTow

Dave didn't say the LSB is pulse-width modulated, he said the signal is pulse-width modulated. There was a strategically placed coma in his sentence.

I haven't read the paper yet so that is as far as I will comment on that but anyone interested in whether it is possible to encode signals below the LSB in PCM audio might want read the info and check out the example in post 4 of this thread:

Digital Audio and Sampling Rates - Gearslutz.com


Alistair

[wado1942]

Originally Posted by UnderTow

Dave didn't say the LSB is pulse-width modulated, he said the signal is pulse-width modulated. There was a strategically placed coma in his sentence.

I'm afraid that doesn't cover it because the signal itself is a pulse code. The pulse code is of discrete, constant intervals which cannot be modulated in width. A PWM signal yields an output of pulses that vary in width, which is an impossibility with a digital signal. I understand what he's trying to say, but his definitions are based on a conflict of terms, which doesn't work in a world of logic (PCM land).

[dcollins]

Originally Posted by wado1942

I'm afraid that doesn't cover it because the signal itself is a pulse code. The pulse code is of discrete, constant intervals which cannot be modulated in width. A PWM signal yields an output of pulses that vary in width, which is an impossibility with a digital signal. I understand what he's trying to say, but his definitions are based on a conflict of terms, which doesn't work in a world of logic (PCM land).

Well, I'm afraid this is also wrong… For our purposes, when we talk about digital it's related to the signal having two states, high and low. One and Zero. There is no requirement that the pulse-widths be equal.

In fact, there is a system called DSD that works just like this; it's a one-bit PCM system where the information is carried in the width of the square wave. It exists in only two states, high and low, but the length (width) of the cycle determines how the signal is reproduced.

Here is a graphical representation:

DC

[wado1942]

That's still not PWM. That's Delta Modulation. The current sample is compared to the previous sample and if it's higher, it's assigned an "on" state and if it's lower, it's assigned an "off" state. Just like with the LSB of a PCM signal, (DM relies on dither to do its job right as well) you can have several pulses in a row that are "on" or "off" and anywhere in between, but each pulse is of constant width. In PWM, each cycle always has a positive and a negative pulse, but each pulse is varied in width, with the positive being wider than the negative or vice versa. On a side note, PWM does not use dither. I agree that the principle is somewhat similar, but there's a reason they differentiate between these terms.

[dotl]

I won't go into math theories...but just want to say this:

Digital recording and reproduction is closer to the source (and there's no q about it, please) but
vinyl/tape sounds 'better' cause it's NOT. Ironic, I know!
There's a reason way movies look different than news channel. Impression of crated realism doesn't have to fully follow principles of known realism to be recognized as realistic. Our mind works as parallel data integrating system operating by interconnecting interactive elements of memory, intellect and emotions. Yes, there's a math that's real but there's also perception that doesn't follow it strictly.

The question is just if ''vinyl can reproduce certain material in satisfactory resolution or can't''. Cause one thing is Miles Davis and another is Ryoji Ikeda or Mahler in crescendo. Every aspect and characteristic of a sound counts. But if vinyl can handle the source satisfactory, i say go for it! Then I would chose it before CD especially for its other advantages in over-all experience.

One more thing...People often recognize vinyl (or tape) as more 'dynamic'.
Compression (probably not being the right word, but i hope you know what i mean) of the material by accumulated noise/saturation just gives impression of increased dynamics to unexperienced listener at moderate listening volume levels. We can find proofs for that all over this forum...I mean gearslutz in general.

[UnderTow]

Originally Posted by dotl

There's a reason way movies look different than news channel.

Sure. Better (digital) cameras. Better lighting. Colour grading. And in general much more time to set everything up on set and edit afterwards and give much more attention to detail during the whole process.

One more thing...People often recognize vinyl (or tape) as more 'dynamic'.

Often because it is! Not because CDs can't be as dynamic but because they suffer from the loudness wars...

Alistair

[dotl]

Originally Posted by UnderTow

Sure. Better (digital) cameras. Better lighting. Colour grading. And in general much more time to set everything up on set and edit afterwards and give much more attention to detail during the whole process.

Yes, to look 'better' in wanted way but not realistic. Realistic can feel wrong...like an incest.

Originally Posted by UnderTow

Often because it is! Not because CDs can't be as dynamic but because they suffer from the loudness wars...

You believe you understand what you think I said, but I'm not sure... There was CDs before loudness war too.

[totalgear]

I understand I am raining on your parade but I find it sad that some are still defending vinyls in 2012.