Why do vinyl and digital sound different* to each other?

edited May 2013 in Systems
Dave's not allowed to post in this one.
Specs / technical info please.

*Please note choice of comparative.


  • Why not?
  • Prove me wrong Rosam...
  • edited May 2013
    Ok I'll take a stab and kick off.

    - Noise floor. Noise sits at around -60db best case and is often higher. it's usually always audible during quiet periods on a track. This in itself is not an issue but it has to apply some masking effect on low level signals, so is audible.

    - Speed instability. The best TTs are intrinsically several orders of magnitude worse than a CD player. Add to this the effect of record warp and off centre pressings and we can easily get audible pitch 'wandering'. In addition, as the cartridge cantilever bounces the combined mass of the cartridge and tonearm on it's compliant hinge, it scrubs the groove fore & aft, causing high rate flutter to a degree that would make the poorest dac appear perfect.

    - LF intermodulation. Most TT systems output considerable subsonic, which thanks to the audiophile mistrust for filters often reaches the output of the phono stage and this has three effects. Firstly it causes bass cone flap on less than well damped loudspeakers (mostly ported) and his causes audible intermodulation of mid frequencies in particular. Think about it, the radiating surface of your mid range driver (in a 2-way) is dancing around in space. Secondly, the driver coil will be moving in/out of the optimum magnetic flux and this will induce dynamic effects. Thirdly, pushing sub 20Hz signals into a valve power amplifier will use up precious capacity in the output transformer cores, limiting headroom and increasing THD/IMD.

    - Response. Few cartridges have a flat response. Most CD players and dacs do to within 0.25dB.
    Cartridges can be typically +/- 3dB and sometimes wider. This will definitely be audible.
    Add to this the signature of the tonearm and turntable as it reflects, stores and absorbs energy in a non-linear way depending on the construction. The cartridge is bolted to a source of noise  - a time smeared version of the real time signal being traced by the stylus tip and it will read this signal back through the cartridge body, superimposing it onto the main signal. This is all audible.

    - Tracking and tracing distortion. The very best cartridges only mange 2-3% THD/IMD at high frequencies and even then only on the outermost grooves. Most cartridges will be exceeding 5% at HF and many exceed 10% (yes 10%) above 10khz. Suddenly the worst of digital looks stunningly fantastic, no? ;)

    - Inter-channel bleed. The best phono cartridges manage 35dB channel separation. Most sit around 25dB in the mid band which allows audible bleed-through. At HF all cartridges become mono - yes, as the cartridge tip mass resonance is approached your stereo cartridge is in fact mono. Most have no more than 10-15dB separation above 10khz and less at 20khz. This is audible.

    - Wet v dry mix. In professional recording the engineer will balance the wet and dry aspects of the mix. Dry being the direct sound, say the capture from a guitar pickup or keyboard. Wet refers to reverb and non-direct sound, so the sound of a room, or the distant pickup of a vocal. It can also include electronic processing effects such as electronic reverb. With vinyl, the system itself adds 'wet' effects due to the stored mechanical energy present in the various parts of the chain, including the vinyl itself, plus the various other distortions. Relative to the direct pickup from the cantilever in real time these other sounds are out of perfect phase alignment. They audibly change the sound. We generally like wet effects - it's the old singing in the bathroom thing!

    - Mastering. The master for a vinyl record and CD are usually not the same - and nor should they be. This is audible.

    - Lossy conversion processes. No not MP3 :) - but along the same lines. The various electro-mechanical conversion processes throughout the chain, including the whole vinyl production process through to the playback process is inherently lossy and it all adds up. This is audible.

    - RIAA phono stages. None of them are perfectly accurate. Few will decode the RIAA encoding with less than 0.25dB of error. And because the errors tend to produce broad response changes, they are usually audible. One of the primary reasons why phono stages sound different is lack of RIAA conformity. This has little to do with cost.

    - Wear. Every time you play a record you make the resulting sound a little more lossy. High frequencies measurably reduce with repeated play and noise increases.

    - Tip profile issues. You could file this under Tracking and Tracing but it's worth a separate paragraph. In order to accurately capture the highest recorded signals at 15kHz on a vinyl record the minor of the two scanning radii on the diamond stylus needs to be <=3.5um, according to research by the Shure Brothers and others many decades ago. Most elliptical styli come in at 8um, most line contacts at 5um and spherical tips can be anything up to 15um.
    This causes tracing distortion but it also causes the frequency response to change as the stylus moves across the playing surface. A cartridge with typical 8um elliptical stylus will have lost around 3-5dB of output at 20khz by the time it reaches the last track on the LP side due to this 'scanning loss' as it's known. This can be audible.

    - Feedback. should need no explanation but you can file it under 'wet' for effects.

    Well that's all pretty horrible and it's a wonder these blasted vinyl things work at all :)

    That's the bad news.

    The good news is that many of these additives and losses are euphonic - we actually like them.
    They combine to give vinyl it's big, bouncy, colourful sound compared to digital.
    You can like both (I do) but such is the catalogue of differences that the two can never be made to sound the same. What I think we can say is that as/if vinyl replay continues to advance in all areas, then the sonics become closer to good digital. Paradoxically this may not be a good thing for those who love the 'sound of vinyl'. So I guess it could be argued that beyond a certain point, technical progress is leading to subjective deficit.

  • Utterly stunning Rob. Exactly the sort of response that I was after.
    Many many thanks for your time.
  • Can we appoint Rob a 'Chews Agony Uncle'? :D

    Nice post Rob, very interesting read.
  • Haven't got the time to read it now, but that looks like a great post Rob. Thanks
  • I see. What it boils down to is 'digital is right'  O:-)
  • When I have listened to really good vinyl, I have been struck by how similar it is to good digital. Indistinguishable, actually.

    Of course, it could have been the other way around. ;)
  • What factors associated with digital prevebt it from being a perfect benchmark?
  • In my limited opinion the problem is in the recording medium and the data capacity of a CD.

    If a recording has been done digitally the recording is a sample of the sounds (it is not 100%).  If the recording is multitracked each track is a sample and when mixed down to stereo the end result is a sample of a sample.  Dave Grusin who makes wonderful digital recordings tries to record everything as live as possible with few overdubs or mutltitracking to avoid this.  Finally when put on a CD the capacity is less and it is compressed again (we have not got the delights of dvd capacity with longer bit lengths or whatever the jargon is).

    If the recording has been made digitally then the vinyl pressing should sound close if not identical to the CD however, the mechanics of a cartridge and turntable come into play so there may be variations in presentation but the end result is a recording that does not sound as good as an analogue one.

    If the recording is analogue it captures 100% of the sound and a noise floor of the medium (tape hiss).  The more multitracking the more noise floor (although they do use dolby and dbx to boost recording levels and reduce the interference of noise).  When replayed what you hear is exactly what was recorded and when transferred to vinyl it sounds wonderful - natural and open.  If they then decide to put it into a digital format the end result is a sample not 100% the conversion is two way - converted to digital converted back to analogue and the end result is down to the designer and the way the music is presented.

    So the answer is - analogue best replay is vinyl.  Digital and save your money and stay digital.

    To verify this all it would take is for a recording to be made both on a digital tape and old analogue tape at the same time and to compare playback of the two tape machines afterwards.  If this was done first with live (one take recording) and then multitracked and overdubbed recordings a concrete view could be taken of which recording medium sounds best.  

    All I know is to my ears analogue recordings replayed on vinyl sound so much better to me.

  • edited May 2013
    In my limited opinion the problem is in the recording medium and the data capacity of a CD.

    If a recording has been done digitally the recording is a sample of the sounds (it is not 100%). 

    This is a common view but Nyquist demonstrates that it can an be 100%, even with CD.

    You will get the introduction of other artefacts such as noise and distortion but the fundamental audio signal is recreated perfectly intact. Where digital systems fall down is at very low signal levels which these days is below -90dB and in terms of noise mostly occurring at double the sampling frequency, so >40khz.
    Both can be made trivially small in magnitude by careful design. Of course it gets better if higher bit and sampling rates are used.

    On Dave's point about Digital being better, I would agree where transparency is the sole criteria, or technical performance if you like. However I would say that from my experience in listening to many hi-fi systems, the notion of transparency isn't high up the priority list and people tend to build their systems to sound pleasing rather than measured against any idea of perfection. Against that backdrop I'd hesitate to call one medium better than the other.

    I'm trying to avoid the X is better than Y type discussion and focussing on identifying the differences in line with what I think are Doc's request in his OP.

  • On Dave's point about Digital being better, I would agree where transparency is the sole criteria, or technical performance if you like. However I would say that from my experience in listening to many hi-fi systems, the notion of transparency isn't high up the priority list and people tend to build their systems to sound pleasing rather than measured against any idea of perfection. Against that backdrop I'd hesitate to call one medium better than the other.

    I'm trying to avoid the X is better than Y type discussion and focussing on identifying the differences in line with what I think are Doc's request in his OP.
    My point about digital being better was in response to Ben's completely unjust swipe at me in the OP ;-)

    I think that the point about how we build hi-fi systems is pretty much spot-on. I don't really claim digital is better than analogue, but people who build systems that I don't like the sound of, tend to go the analogue route. 

    Back in the day when LP12s ruled the roost, I really struggled with analogue sounds. But I know I could happily live with a well-sorted 401, Voyd or Rock. One day I may get a turntable again, but only to access music that's only available on vinyl.

    I hate messing about with records and their related paraphernalia.
  • I'm trying to avoid the X is better than Y type discussion and focussing on identifying the differences in line with what I think are Doc's request in his OP.
    Yes Rob. Yes. Thank you.
  • Dave. Yes. analogue faffing is a major turn off. Definitely an unsightly feature on the vinyl landscape.
    edited June 2013
    Many of the points Rob mentions are valid, but OFTEN all but inaudible with a well engineered TT replay system and a decent recording so it is rather misleading to conclude that vinyl must always be inferior. That simply isn't the case.

    Red Book was derived precisely because there was no standard initially for digital mastering and even with Red Book, many recordings are so far below the ideal standard that they don't deserve the tag "hifi".

    Whereas Vinyl replay has many variables that can induce distortion, digital has many variables in the recording and mastering process that can and do utterly destroy any semblance of dynamics and tonal accuracy.  There are good recordings and in these days of hi-rez downloads, there'smore was to benefit, but it all depends upon how the original was mastered.

    This CD V's digital nonsense will rage on for decades I suspect but it really bothers me that few people see the wood for the trees, ie its the recording and mastering, irrespective of the final replay system, which determines the potential for a great performance.

    Vinyl may lose up to 5KHz after very few plays of a new LP but lets face it, how many of us have perfect hearing above 15KHz and could tell the difference?

    The medium to my mind is less important than the mastering.

    Having said that, I always revert back to a well recorded LP.  It just sounds better, whatever the reasons or drawbacks.
  • I agree with Paul (except for the very last part - I have no TT at home).

    Such differences in mixing and mastering easily show through on a smart phone, playing an MP3 internet stream through earbuds.

    With a decent mix to start with, even though its been compressed for transmission/streaming, the results are entirely enjoyable, while other recordings are clearly inferior to begin with.

    I'm not suggesting for a moment that this is 'good enough', but it adds weight to the suggestion that the recording itself is an order of magnitude above the replay medium.

    I do believe that digital replay is generally more accurate, also, given the anecdotal evidence, it seems that vinyl is seen as capable if superior results with sufficient equipment, and careful setup.

    I find the idea that analogue generally has well sorted mastering but has room for tweaking the replay system; and digital may suffer some poor mastering but the replay system is generally robust and accurate, quite compelling.

    However, contrary to logic and received 'wisdom', I believe also that digital replay systems are 'tweakabe', and as such are capable of equalling analogue given decent recordings. There seems to be more to this game than 1st and 0s.
  • I wanted to avoid the entirely valid "which is better debate" here. I was interested in why the respective businesses of retrieving information from vinyl and digital sources appear to yield different results. Why don't LPs and CDs and SSD all sound the same?
    edited June 2013
    It's a good question Ben and there's a fair few reasons!

    Firstly, if theoretically the best LP pressing, mastered perfectly following a good recording is considered, at best you are looking at a dynamic range of say 60dB.  For CD/SACD, theoretically, they should be capable of 90dB (but rarely are).  

    Whilst a mediocre LP might achieve 35 to 45dB, a mediocre CD OFTEN achieves perhaps 20 to 30dB, so the medium is under exploited for various reasons (laziness of mastering engineers, loudness wars etc etc) but that aside, it will sound different (louder and more edgy in many cases) than the average LP.

    One reason that it's hard to produce an LP with a dynamic range as low as CD is that the stylus would jump out of the grooves if over-compressed (ie tracking would be awful) so the effective limit for a reasonably "tight" LP might be say 30dB which is actually better in practice than a mediocre CD/SACD.

    LP replay relies on cartridges to trace the grooves and convert that potential energy into electrical energy and the performance of a cartridge tends to vary between carts/brands more than the perceived differences between many DACs these days.  Its also subject to how well set up the cartridge is (faffing about as you'd say!).

    A well set up quality low output MC cart playing  a top notch LP on a top notch replay system should be capable of far better results than the average CD recording albeit with caveats.

    They will certainly sound different because if different carts/tonearms/phonostages sound different from each other (and the differences can be startling) then they will certainly sound different to digital sources!  Not necessarily better, just different.

    A well designed DAC on the end of a stable transport playing a first class CD recording should in theory be better...more dynamic range, better pitch (hence harmonic) accuracy and lower noise floor (inky blacknesses!).

    The reality though, irrespective of what measurebaters might say is that there also exists differences in DACs. Even identically specced DACs often sound different to me because not everything is explained by the handful of specs alone.  Some are clearly better than others, some produce a warmer sound (ok...more distortion), some produce better separation; some seem to be more detailed or have more solid bass etc etc. There's reasons connected with the circuit other than the DAC end of things for this and often it;s the gain stage on the end of the DAC (which can introduce noise or not reproduce the frequency range as perfectly as perhaps they should in terms of flat performance...ie they can be slightly rolled off at the extremes perhaps?).

    There's more scope for vinyl sources to differ from each other though and digital does not require careful set up!

    I take great care with vinyl set up and use carts which I know (from listening and specs) to be as neutral as possible, but there's still major differences between my decks.  Pitch stability, tracking accuracy, flux damping, surface noise transfer, azimouth accuracy, bias accuracy etc etc etc...it's not hard to see that because so many variables are involved, vinyl will always sound different from vinyl, let alone digital!

    The real beauty though is that there's far greater scope to tune the sound to your liking, accuracy to one side, so that you can really engage emotionally with the recordings.  It's a learning curve though and digital is way more convenient.

    I listened to a £10 DAC no larger than a box of matches yesterday and was blown away with just how good it was. Thats the biggest difference for me...whilst differences no doubt exist between mediums, digital has the advantage that it's a more affordable "in" for those starting out or on a budget.

  • Thanks Paul.
    The idea of variables in the vinyl setup causing variance *between* vinyl sources resonates with me. Certainty my experience is that playing about / changing / upgrading vinyl front ends has been the most gratifying for me over the years as there have been such diffrences / improvements to be found. There aren't many things that I've changed in my vinyl front end that have made only a little difference.
    Paul, in your sagely opinion, is there an "optimum" vinyl setup (theoretical or otherwise) based on the variables you list, or is it all about trade offs and comprimises?
    I only ask these questions out of interest. I go through periods of vinyl listening, and am always pleasantly surprised each time I rediscover the old LPs. But also I am struck that the vinyl "sound" in my system is different in my system to the "sound" of my digital front end. I accept that my TT setup is almost certainly not perfect (neither is my digital setup I'm sure).
    edited June 2013
    What you say is probably something a lot of vinyl die-hards can relate to Ben.

    Getting the best out of a vinyl front end is not (sadly) as romantic as the medium itself!  It's actually a systematic and logical process.

    The main thing, and one never to be forgotten is that whatever you listen on, correct set-up isn't just about getting the most from the sound, it's about preserving the condition of your valuable music collection.  Badly set up cartridges destroy LPs.  It's also about taking care of the LPs by cleaning them thoroughly every now and then (for me, about every 4 or 5 years).

    The most obvious start point is budget. let's assume that you already have a TT.  Of equal importance are the arm you use (good low resonant and stiff design with excellent bearing surfaces and fine tolerances) and the cartridge itself.  To get the very best out of an LP, there's no real argument other than using a low output moving coil cartridge (I've written a brief blog on the subject HERE

    However, not everyone has that sort of money and these days, there are some very good MM cartridges for reasonable sums of money.  You need to ensure that you do your sums regrading cartridge/arm compliance.  You want the resonant frequency of the arm/cart combo greater than 9Hz and less than 13Hz ideally although occasionally matches just outside of this range will work very well, so it's a guide.  The sylus profile is important in several ways:  

    firstly it determines (partly) how well the cartridge will track;  secondly it has a bearing on surface noise and lastly, it has a bearing on detail retrieval.

    Following cart selection, set up has to be much fussier than most people appreciate.  Most people these days use either Stevenson or Baerwald null-point protractors as for many 9 inch arms they allow the best compromise between average distortion and lowest distortion across the LP. You have to experiment with this.  Whatever one you choose to use or that is recommended for your arm, its vital to get set-up spot on and for this a test LP is very useful (something like the HFN Test Record which also includes a useful protractor).  Deviations of half a mm on alignment will result in fairly significant increases in distortion, ditto bias which isn't set properly.  Azimuth must be spot-on.  Once these things are set, vertical tracking force should initially be set to the heavier end of the recommended scale for your cart, especially if it's a new cart.  You're less likely to cause LP wear through increased stylus tip resonance at the heavier end plus trackability improves.  the height isn't as critical as people make out since LPs are cut with rake angles that vary slightly.  A few degrees difference in rake angle translates to quite a large difference in either tail up or tail down for the tonearm, but keeping the arm level is good enough.  

    Some people claim that everything snaps into focus with a few mm change in arm height.  That's utter nonsense and suggests something is badly set up elsewhere!  Even if it were the case, it will change between LPs and LP thicknesses!

    You can see just how many variables are involved before considering the TT itself, the power supplies, and the phonostage and cables.  For moving magnet cartridges it is essential that only low capacitance cables using generous screening are used.  Total circuit capacitance between cart and phonostage output should be no greater than 250pF ideally otherwise HF will be rolled off and the sound will be dull.  Given that most phonostages use 150pF loading, that leaves 100pF max for the leads, so you should be aiming for a lead offering circa 65pF/m ideally. Its not such an issue with MC carts for other more nerdy reasons!

    TT should be very well isolated from floor and air born vibration, pitch stability should be good (test using a slow piano piece...you'll soon pick up any Wow/flutter).  Well engineered decks don't have to cost the earth.  
    Ensure RCA sockets are cleaned every now and then too on TT/phonostage/preamp sections (it makes a difference).

    Follow this basic guide and you might be quite surprised just what improvements can be had from your existing set up.  I can provide a recipe for an LP cleaning solution if anyone would like it.
  • This is rapidly becoming my favourite 'chews thread eva. Thanks Paul, your info tastes almost as good as your mini muffins.
Sign In or Register to comment.