LP with better dynamic range than digital

[Orb]

Yes indeed, don't hold your breath! But of course, sound preference is not always about the best replication, for example tape typically has almost no second harmonic distortion but mostly third order. And folks like that, the ear apparently finds "detail" in that. Its that trick of pleasing the ear, and that certainly comes back to preference.

Keith Howard did an investigation into different type of harmonic distortion (created a model-stimulation that could be applied to real music and used various patterns), his preference came out to be no distortion.
I think more study really is required with trained listeners before it can be taken for granted whether the statement you or others give such as euphoric distortions being preferred compared to a clean content.
Yes I appreciate there are some historical papers, but IMO a modern study is required utilising more advanced tools even to what Keith Howard did and importantly like he did applied to real world chosen music.

Cheers
Orb

 

[Orb]

I know it seems complex, perhaps it is. However, record IMD tones into a digital recorder at half full scale, and record into a cutter at half full scale for lp and then cut a record we can playback on our gear, and when you play back each and look at the results, the lp version will have more distortion. So, at some level, it has to be accepted that there is more distortion. Why it adds up to something we like is buried across a lot of threads we have here on wbf, atleast why we THINK it is so... your idea of a good study into this would be welcome and I think would finally allow digital folks to "duplicate" what is best about the LP sound for many.

Sure, but the problem is Keith Howard did some complex and better testing with various harmonic patterns, and in every case his conclusion was that distortion was always a negative.
He is not the only one and there has been a few comparable studies, but his approach could be applied to any music.
Unfortunately a lot of conclusions relating to distortion/harmonic patterns is based on a lot of circumstantial approaches; whether we are talking about sensitivity to harmonic distortion or preference to type of harmonic distortion patterns when applied to music.
We also hear a similar debate-case with tubes all the time; and yet it does not hold up when one looks into it in further detail and how this can be contradicted by the different performance between said top tube equipment.
Same is also said about white noise being audibly benign...... But then what about those (and I am one) who absolutely hates white noise when listening to music.

It is not possible to at present IMO to generalise about distortion/noise/etc and their real world influence on preference until studies are done with music that also is controlled (so has without and with distortion/noise variables and at various threshold levels).
Sorry if it seems I am being a bit picky on this subject, just I think it is a leap too far at this stage to say categorically the influence when applied to extended (not necessarily all at once) critical listening to music.
I can accept it is a position that is divided between members and that it is probably an agree to disagree proposition
And comes back to the dreaded trained listeners for studies-investigations

Thanks
Orb

 

[microstrip]

IMHO we should separate "perceived dynamic range" from "measured dynamic range." Several good solid state DACs I have owned sound much less dynamic than my Metronome CD setup, and measure much better in the classic dynamic range parameters. And sorry, my objective in debates is perceived dynamics - if the measurement does not map it, we need another way of measuring! BTW, IMHO most of the time we are debating the implementation of the hardware use to play the media, not the intrinsic properties of the media itself.

I believe that digital playback still has a lot to evolve - it will be a question of time until it reaches some perceptual aspects still reserved to top analog. Also IMHO (and I know many people disagree with me) the systems of most people owning top analog systems are not optimized for digital playback.

 

[Robh3606]

This is an interesting thread. As far as I know one of the most dynamic recordings ever pressed was the Telarc 1812. It had no where near 120dB of dynamic range more in the 60's and it was very difficult to find an arm and cartridge combo that could make it through without mistracking. Not even going into the demands if put on the rest of your system.

Rob

 

[Atmasphere]

Mike

I understand your sentiment, however those quotes do not tell why LP has more dynamic range than CD which I know is hard-limited to 22.1 KHz ... I find the 120 dB of LP figure rather surprising ... Just some references would be enough.

That's easy!

Digital audio relies on a linear means of encoding. So what is happening is that as the signal level is decreased, bits are slowly turned off to reflect that smaller number of the resulting analog voltage.

By the time you get to about -45 db, half the bits of the 16 bit word are turned off. The result is that there is less resolution (this is why CDs are not particularly good a low level detail).

This loss of information is quite audible! So to get around the problem during the recording process the signal is run as high as possible. This takes advantage of the greater number of bits available. The problem is, if you go over 0VU, all the bits are now ON and you can't go any higher. The result can ruin a recording. So it is common to use some form of limiting, either during the recording, or (if the producer has a greater sense of what good quality is) the recording was done at a lower level and the limiting or compression occurred in the mastering process.

Now there are several things that happen in the CD mastering process. Each track in the raw digital file(s) has to be normalized- this involves scanning each track for the highest level in the track and calculating a ratio such that the highest level in the track now corresponds to 0VU and each digital word in the file is recalculated according to the same ratio. In this way the CD will be at the same level as other CDs that you were playing so the volume will be the same. Limiting and or compression is applied, so that the CD will have better use of the bits and so will sound better due to greater resolution (EQ might be applied as well).

This compression occurs with almost any CD as the mastering process is considered essential.

The bottom line is that the Redbook format has 96 db but its really not used. In fact the amount of compression used is generally more than you typically see during the LP mastering process. **That* is why CDs generally have less dynamic range. The recording industry makes them that way on purpose. If anyone here is sensing some irony then you have the right idea.

 

[Mike Lavigne]

This is an interesting thread. As far as I know one of the most dynamic recordings ever pressed was the Telarc 1812. It had no where near 120dB of dynamic range more in the 60's and it was very difficult to find an arm and cartridge combo that could make it through without mistracking. Not even going into the demands if put on the rest of your system.

Rob

just get a Herzan....

the reality is is that the main culprit of miss tracking from dynamic music passages is feedback, not the groove topography. with RIAA bass de-emphasis the bass notes are reduced significantly during cutting and then boosted at playback. many early record players were one box set-ups with the speakers part of the box....records had to be cut with caution.

i can tell you that miss tracking the most dynamic passages does not occur with active isolation, even with speakers that are -3db at 7hz and -6db at 3hz 8 feet away.

of course, the top level current state of the art tonearms and cartridges do help too.

these days lots of tt's have excellent passive isolation too.

typically it's the execution of an analog format, not the format itself, which is the limitation. we keep finding more music in those grooves as we approach the ideal of Lp playback.

 

[esldude]

Thanks to Tim for reminding us it is Rudy Van Gelder's birthday. Some quotes from an interview with him in 2011.
http://jazzprofiles.blogspot.com/2011/05/rudy-van-gelder-signature-sound.html


What are your feelings on digital versus analog?

The linear storage of digital information is idealized. It can be perfect. It can never be perfect in analog because you cannot repro*duce the varying voltages through the dif*ferent translations from one medium to an*other. You go from sound to a microphone to a stylus cutting a groove. Then you have to play that back from another stylus wig*gling in a groove, and then translate it back to voltage.

The biggest distorter is the LP it*self. I've made thousands of LP masters. I used to make 17 a day, with two lathes go*ing simultaneously, and I'm glad to see the LP go. As far as I'm concerned, good rid*dance. It was a constant battle to try to make that music sound the way it should. It was never any good. And if people don't like what they hear in digital, they should blame the engineer who did it. Blame the mastering house. Blame the mixing engi*neer. That's why some digital recordings sound terrible, and I'm not denying that they do, but don't blame the medium.

 

[Robh3606]

the reality is is that the main culprit of miss tracking from dynamic music passages is feedback, not the groove topography.

Hello Mike


That might be true but not in this case. The groove topography was extreme. Have you ever seen an original pressing?? I still have mine and it's something to see. The point I was making is there is no way you are going to get 120dB as claimed on an actual vinyl record you can play on your turntable.

Rob

 

[Don Hills]

...
By the time you get to about -45 db, half the bits of the 16 bit word are turned off. The result is that there is less resolution (this is why CDs are not particularly good a low level detail).

This loss of information is quite audible! ...

:facepalm:

 

[Atmasphere]

^^ I recommend you get a digital recorder and play with it.

 

[esldude]

^^ I recommend you get a digital recorder and play with it.

If he does, he would only find that your description still makes no sense. Causing a facepalm all over again for having bothered.


You can read this Vanderkooy and Lipschitz paper for the explanation.

http://drewdaniels.com/dither.pdf

Or this page by Benchmark.

http://benchmarkmedia.com/blogs/new...t-audio-has-more-resolution-than-16-bit-audio

 

[Mike Lavigne]

Hello Mike


That might be true but not in this case. The groove topography was extreme. Have you ever seen an original pressing?? I still have mine and it's something to see. The point I was making is there is no way you are going to get 120dB as claimed on an actual vinyl record you can play on your turntable.

Rob

i have more 1812 pressings than you can shake a stick at....over 10...maybe 15 of them. if i have the Telarc i'll throw it on tonight and report.

i've yet to find 'music' that my cartridge cannot track. some test records can find the resonate frequency of your arm and wreak havoc.....but that's not music.

 

[Orb]

If he does, he would only find that your description still makes no sense. Causing a facepalm all over again for having bothered.


You can read this Vanderkooy and Lipschitz paper for the explanation.

http://drewdaniels.com/dither.pdf

Or this page by Benchmark.

http://benchmarkmedia.com/blogs/new...t-audio-has-more-resolution-than-16-bit-audio

The benchmarks is not a good example IMO (waaay too generic-simplistic on dither,16-bit, and also DSD), agree with your context and point though.
I guess Benchmark never heard of PDM MEMS microphones that connect direct to DSD associated signal processors within AD/DA hardware solution.
Anyway raised this aspect of DSD before in another thread ages ago; DSD may be thought of as 1-bit PCM but technically it is not (PDM MEMS microphones simplest example why not and closer to DSD/1-bit system than PCM).

Cheers
Orb

 

[Don Hills]

The benchmarks is not a good example IMO (waaay too generic-simplistic on dither,16-bit, and also DSD), agree with your context and point though. ...

John sums it up well in the conclusion, though.

... When dither is properly applied, there is no advantage to long word lengths other than improving the noise performance. Quantization errors in a properly-dithered digital system produce nothing other than random noise. Properly-dithered digital systems have infinite amplitude resolution.

Long word lengths do not improve the amplitude "resolution" of digital systems, they only improve the noise performance. But, noise can mask low-level musical details, so please do not underestimate the importance of a low-noise audio system.

There's a set of samples illustrating the point at the bottom of this article:
http://www.tnt-audio.com/sorgenti/dither_e.html
They're all 4-bit 44.1 KHz sample rate.

Remember that for a realistic indication of how audible the noise in the samples is, you need to turn down your volume to 72 dB below your normal listening level.

 

[esldude]

John sums it up well in the conclusion, though.


There's a set of samples illustrating the point at the bottom of this article:
http://www.tnt-audio.com/sorgenti/dither_e.html
They're all 4-bit 44.1 KHz sample rate.

Remember that for a realistic indication of how audible the noise in the samples is, you need to turn down your volume to 72 dB below your normal listening level.

Yes, that is an excellent example of the value of dither, especially shaped dither. I wanted to include the link in my post, but couldn't quite remember where I had heard that demo. So good that you knew of it and posted.

 

[Orb]

Yes, that is an excellent example of the value of dither, especially shaped dither. I wanted to include the link in my post, but couldn't quite remember where I had heard that demo. So good that you knew of it and posted.

Although one should use flat TPDF in this context (music and PCM), and there is enough anecdotal evidence to suggest (appreciate this cannot be taken as fact until further studies) we are sensitive to dither solutions-how process is applied and potentially reiterated; not just the work done by Keith Howard and a few others taking his approach but also many respected audio engineers who would not necessarily understand the maths-mechanisms involved with the various dithers.

Don,
again they are being oversimplistic and using a specific context IMO that does not reflect real world; please show me how a properly dithered 16-bit system has infinite amplitude resolution or how real world 16-bit and 24-bit properly dithered test tones for a digital system have exact same performance/trait (they do not btw).

But yeah I agree dither IS critical for audio in many situations-scenarios just that it is made out to be the holy grail of perfect digital generally and it is not (much more complex).
Sorry for digressing, I will leave it at that as we have plenty of existing threads already that go into detail and with many papers linked in them by many of us.

Just to emphasise we are not disagreeing, my objection is more to do with the generalisation/context in parts for one of the articles linked earlier.
Orb

 

[Don Hills]

... we are sensitive to dither solutions-how process is applied and potentially reiterated; ... many respected audio engineers who would not necessarily understand the maths-mechanisms involved with the various dithers.

Indeed. I am aware of the debates amongst audio engineers over when to dither and the effect on the audio of the various forms. There's perhaps a little snake oil around the various trademarked flavours of shaped dithers, but there's general agreement that they are audibly different.

... please show me how a properly dithered 16-bit system has infinite amplitude resolution or how real world 16-bit and 24-bit properly dithered test tones for a digital system have exact same performance/trait (they do not btw).

They do perform the same when you get down to the same ENOB, but I see your point. As for "infinite amplitude resolution", it works the same for digital as for analogue - it's a balance between amplitude resolution and temporal resolution. In the presence of noise, infinite resolution requires infinite time.

... Sorry for digressing ...

It's my fault as well. It's a distraction from the original point - that in a properly implemented digital system, reducing the number of bits doesn't reduce the resolution. It just raises the noise floor. It's a non-intuitive point that can be hard to get across. Doubters could perhaps ask themselves that if reducing the bit depth reduces the resolution, how does DSD work?

 

[FrantzM]

<snip>in a properly implemented digital system, reducing the number of bits doesn't reduce the resolution. It just raises the noise floor. It's a non-intuitive point that can be hard to get across. Doubters could perhaps ask themselves that if reducing the bit depth reduces the resolution, how does DSD work?

Thanks Don.. This has too often been repeated in this thread and others.

This said I am still waiting for the 120 dB Dynamic range LP ... My calculations may be off, the collective will correct my assumptions...
Let's take the case of an MM magnet with say 3 mV in a 47 KOhms load. Let us suppose that it is capable of going as low as 0.1 mV... as its lowest output .. 60 dB of dynamic range means that the cartridge must be capable of outputtin 1000 times the 0.1 mv which would be 100 mV which is already very high... To reproduce 120 dB, the cartridge should be able to produce 1,000,000 times 0.1 mV ... about 100,000 mV or 100 Volts ... Do you know any cartridge capable of such? I don't know how low a MM can go we can assume .01 mVolt it still would mean a cartridge capable of 10 volts output for ... 120 dB dynamic range...

 

[Orb]

Thanks Don.. This has too often been repeated in this thread and others.

This said I am still waiting for the 120 dB Dynamic range LP ... My calculations may be off, the collective will correct my assumptions...
Let's take the case of an MM magnet with say 3 mV in a 47 KOhms load. Let us suppose that it is capable of going as low as 0.1 mV... as its lowest output .. 60 dB of dynamic range means that the cartridge must be capable of outputtin 1000 times the 0.1 mv which would be 100 mV which is already very high... To reproduce 120 dB, the cartridge should be able to produce 1,000,000 times 0.1 mV ... about 100,000 mV or 100 Volts ... Do you know any cartridge capable of such? I don't know how low a MM can go we can assume .01 mVolt it still would mean a cartridge capable of 10 volts output for ... 120 dB dynamic range...

Frantz,
you can only in theory get 120db+ dynamic range with 16-bit PCM IF going with noise shaping+dither, in reality the ideal and what is recommended/used with music is flat TPDF.
Cheers
Orb

 

[Atmasphere]

... When dither is properly applied, there is no advantage to long word lengths other than improving the noise performance. Quantization errors in a properly-dithered digital system produce nothing other than random noise. Properly-dithered digital systems have infinite amplitude resolution.

<sigh>*There's* something they don't teach in school... </sigh> I'm sorry but that statement is pretty funny.

We can pop a hole in that one pretty quick. Let's limit the word length to 4 bits (since there is "no advantage to long word lengths", we will prove or disprove that by using a short word length). Please show how a 4 bit word can describe infinite values. (Hint: 16 values are possible with 4 bits. With 16 bit you get 65536 values.)

Thanks Don.. This has too often been repeated in this thread and others.

This said I am still waiting for the 120 dB Dynamic range LP ... My calculations may be off, the collective will correct my assumptions...
Let's take the case of an MM magnet with say 3 mV in a 47 KOhms load. Let us suppose that it is capable of going as low as 0.1 mV... as its lowest output .. 60 dB of dynamic range means that the cartridge must be capable of outputtin 1000 times the 0.1 mv which would be 100 mV which is already very high... To reproduce 120 dB, the cartridge should be able to produce 1,000,000 times 0.1 mV ... about 100,000 mV or 100 Volts ... Do you know any cartridge capable of such? I don't know how low a MM can go we can assume .01 mVolt it still would mean a cartridge capable of 10 volts output for ... 120 dB dynamic range...

The 0.1mv is by no means the lower limit of the cartridge! In addition, a typical MM is usually rated at 5mV FWIW. It can make quite a bit more than that, as the measurement is usually taken at a certain frequency, and due to the slope of the curve, an octave can affect the output by 6db. The output level is given for reference purposes only and does not describe the upper limit of the cartridge or the playback system. For a MM phono circuit, its good practice to have your overload limit at around 150mV if that gives you any idea. So the range of a MM is considerably wider than described above.

The reason you don't see 120 db on LPs has more to do with the playback than the record side. The trick is to keep the stylus in the groove. Too much amplitude and it mistracks or can jump out. This was a problem on the original RCA 'Fountains of Rome' conducted by Fritz Reiner, the organ pedal tones at the end of the side caused cartridges of the day to jump out of the groove.

Again, the main reason you see more dynamic range on LPs is because more compression is used in digital media. This is a record industry thing and its not going to change with any speed- not at least until cars are quiet enough inside that the compression isn't useful. No-one expects LPs to be played in cars...