Saturday, 3 June 2017

MUSINGS / MEASUREMENTS: Thoughts on Vinyl LP Fidelity... (And thoughts on Sgt. Pepper's 50th Anniversary Remix.)


Alright... Let's not get worked up about the cartoon :-).

I know, the debate between vinyl and CD (digital) remains one of the greatest "wars" of the audio world; not that it should be a big deal IMO... As usual in most audiophile conflicts, there are many words spilled about the topic, but it is rare to see "data" or actual direct comparisons. For example, look at Wikipedia's entry for "Comparison of analog and digital recording" and we see no actual illustration to demonstrate differences.

Years ago, I wrote a little about this here (and I think it's only fair that I have a digital-preferring Batman cartoon). As you know, over the years there are all kinds of vinyl evangelists going around touting the superiority of LPs over CDs and digital overall (and vice versa although I'm not sure the digital camp is as evangelistic). Here's a nice example of all the awesomeness that is vinyl based on someone's opinion.

As I've said before, I agree that there are some great qualities about having an LP collection. So long as you have the space for the non-biodegradable collection, the beauty of the artwork is wonderful. It's comforting and collectable memorabilia. And likewise the sound can be alluring in the same romantic way. For some, the ritualization of the playback process can bring with it that sense of security and physical engagement as well.

But let's just not beat around the bush, there's no comparison when it comes to fidelity. There's just no excuse for editorials like this one (silly claims of digital "losing some of the very continuousness of presentation" among other false gems). When we're looking at the ability for the system to maintain accuracy / fidelity in terms of what was originally on the source and the system subsequently able to copy, store, transfer, and reproduce the waveform as closely to the original as possible, the difference in capability is truly worlds apart. In this regard, a CD (and digital file, especially hi-res) is obviously able to handle the quality in a way that will not corrupt the playback with noise, distortions, or timing anomalies.

Remember that back in 2014, I took the PlatterSpeed software (Apple and Android) for a "spin" to demonstrate the lack of wow and flutter with digital. In comparison, we see the results using my Technics SL-1200 M3D and my friend's Roksan TMS. To make it easier to compare, here are the results in one large summary image (I used the iPad app)...


Click on the image to have a close up. The obvious take-home message is that when it comes to temporal and by extension frequency stability, LP playback leaves much to be desired even when we're looking at a rather expensive turntable like the Roksan TMS (I) with a set-up cost easily exceeding $5000 on the used market when you add in the cost of that SME309 tonearm and Whest PhonoStage.20 preamp (see the detailed system description here). These days, a used but excellent condition Technics SL1200 M3D isn't cheap either. I'm obviously not comparing the output from cheap Crosley turntables!

As for the LP itself, remember, no piece of vinyl is perfect since it is a physical piece of plastic. Mild decentering of the spindle hole and tiny warps on the surface are par for the course... I'm using Dr. Feikert's Adjust+ 7" Test Record here which as a test disk I'm sure is pressed better than most commercial LPs and looks to be flat as a ruler with no warps. Remember that this test disk is also used for precise adjustments like setting cartridge azimuth.

It has been a long time since I've seen direct objective head-to-head data between LP and digital, so I thought I'd use the test LP to make some FFT overlays to demonstrate the differences. Easily done. Basically I connected the common XLR output from my Emotiva XSP-1 preamp to my Focusrite Forte ADC.

The vinyl playback chain:
Dr. Feikert's Adjust+ LP 3150Hz sine @ 33.3rpm --> Technics SL-1200M3D (stock tonearm with Denon DL-110 cartridge) -->  Emotiva XSP-1 --> Focusrite Forte ADC --> Win 10 laptop
The digital playback chain:
16/44 FLAC of 3150Hz sine (-12dBFS) --> Pi 3 Touch streamer --> TEAC UD-501 DAC --> Emotiva XSP-1 --> Focusrite Forte ADC --> Win 10 laptop
We can then record and analyze the 3150Hz sine wave tone to see just what kind of playback accuracy and quality comes out. Note that since I don't know the absolute loudness of the LP Adjust+ track, I actually "handicapped" the digital resolution a bit by taking out 2 bits worth of dynamic range with the 3150Hz tone at -12dBFS. Here's what the set-up looked like with my Technics SL-1200M3D and recording/measuring gear:


First, using WaveSpectra analyzing the audio FFT in realtime, this is what playback looks like "zoomed" in to the 3150Hz primary tone (frequency scale 1-5kHz):


As a still image, we don't appreciate that in fact because of wow & flutter, as shown by the large 3-way comparison graph, depending on the platter speed at the moment, the actual frequency from the LP playback varies back and forth slightly around 3150Hz. In the image above, I think we can make out the LP playback frequency as slightly higher than 3150Hz (shifted a pixel to the right) whereas the digital tone is precise.

If we zoom out a bit and look at the full 96kHz bandwidth captured with 192kHz digitization:

Notice that even if we ignore the noise level difference, the LP signal isn't clean. We can see the higher order harmonics evident - approximately 2% total harmonic distortion in the signal through the LP --> Denon DL-110 cartridge --> Emotiva phono/preamp. This is orders of magnitude higher harmonic distortion compared to an equivalent digital system (which even in inexpensive DACs these days would be significantly less than 0.05%).

We can zoom in and have a look at the sine waveforms themselves and see the differences between the digital output and analogue LP output (the 2 stereo channels overlaid):

Notice that even though I tried to normalize both waveforms to 100% amplitude, variation in amplitude with the LP playback resulted in a lower amount of gain applied. Notice the DC offset with the right channel in the vinyl playback especially (blue waveform shifted upwards) which adds to a bit of channel volume imbalance with that right channel louder. There's also a relative delay comparing the right with the left channels. The digital playback is clearly more precise in timing, amplitude, and overall morphology like symmetry and precision (smoother sinusoidal waveform).

Because we know the turntable system is significantly less accurate in the time domain, we can average out the spectrum over a few seconds to appreciate the "temporal smear". Here is the spectrum averaged over 5 seconds (using Adobe Audition 3 FFT for analysis):


Note the logarithmic frequency scale. What we see is that clearly the LP playback in green is very much noisier than the 16/44 digital playback. At around 1kHz, the difference in noise floor is around 20dB worse than digital. The noise floor difference gets worse with lower frequencies using my setup (a nice warmish sounding surface noise plus whatever amount of rumble). Again we see the ~2% total harmonic distortion in the playback from the LP.

Because we're seeing the spectral average over 5 seconds, we also see that the 3150Hz primary signal is taller with the digital playback. Due to the LP playback inaccuracy, there's "spectral spreading" resulting in that "skirting" at the base of the primary signal which is absent with digital playback. Over time, because of imperfections deviating from exactly 3150Hz, some of the frequency "energy" is spread around. According to PlatterSpeed, my Technics SL-1200M3D deviates -9.5/+8.8Hz which when low-passed still stays around -0.8/+0.5Hz (-0.02/+0.02%) over a minute; this is certainly quite good as turntables go but obviously much "looser" compared to digital playback stability from the TEAC DAC!

Conclusions:

I hope this post has provided an interesting perspective into the differences in objective performance of an analogue turntable / LP system as compared to a 16/44 digital playback using the same pre-amp output.

From a price perspective, digital is "cheap"! My 2013 TEAC UD-501 can be had anywhere from US$400-600 these days (I could have just as well used the <US$90 SMSL iDEA from last week to show a similar output fidelity). In comparison, consider how much it would take to buy a minty quality used Technics SL-1200 turntable these days (of course you could consider the inexpensive SL-1200-like options like the Pioneer PLX-1000 for ~$700), then add another US$200 or so for the Denon DL-110 HOMC cartridge I'm using here, and then there's a phono pre-amp... As we have seen, in the last year, Technics has been cashing in on their "reborn" SL-1200G series ranging from US$4000 to an "affordable" $2000 with the recent SL-1200GR. Does it really make that much sonic difference compared to a vintage model in the context of the limitations in a turntable system?

From a performance perspective, sure, I could spend much more for a better cartridge, we could spend >$100,000 for a turntable, countless dollars on isolation platforms, expensive pre-amps, etc... But let's face it, as much as you clean your LP, there will be surface noise worse than the silence of digital zeroes. The signal to noise resolution from an LP cannot beat 16-bit digital much less get anywhere close to 24-bits dynamic range. And since no plastic / PVC / nonbiodegradable vinyl LP is ever perfect (take note environmentalists), good luck finding a disk with a perfectly centered spindle hole to minimize the temporal distortion!

Remember too that a direct drive turntable like the Technics already will perform better than the majority of belt-drives in terms of speed stability, so the results I'm showing here are better than many other 'tables out there.

As for digital playback and time domain, remember how over the last few years, I've been saying that jitter really doesn't matter? Consider this 24-bit J-Test result when I use a Squeezebox Touch to send the signal to my TEAC UD-501 using the thinnest piece of plastic TosLink cable I own:

It's an example of the imperfection of S/PDIF and the jittery embedded clock. Looks nasty, right? Remember, in a 24-bit J-Test we should not see sidebands or a jitter modulation pattern at all. All those peaks around the 12kHz primary signal are reflective of data jitter.

But look what happens when I record that J-Test signal (yellow) at the same amplitude as the other signals from the Emotiva XSP-1 preamp, and overlay it on the 5-second log graph:

Notice that with the volume attenuation (a rather normal -19dB on my preamp) to match the amplitude of the other signals, the noise level of the system actually drowns out the jitter anomalies around the 12kHz primary tone. When I do my DAC measurements, although I can show the jitter anomalies to quantify differences between devices, it doesn't mean that jitter will be audible whatsoever during actual playback once it goes through your signal chain! And notice that in a real system, this signal still has to be sent to an amplifier and your speakers. This is why I've said all along that if anyone gets worked up about digital timing jitter, he/she should absolutely stay away from vinyl!

Timing anomalies, resolution limitations (in real life, analogue is far from "infinite resolution" - insignificant rationalizations by Steve Guttenberg here), and harmonic distortions are clearly technically much worse with LP playback. Another way to say this is that digital is more "transparent", way less "coloration" is added to the sound. Let's not forget that I have not even mentioned vinyl imperfections like ticks and pops from dirt and groove anomalies, "mono bass" for technical reasons, or inevitable reduction of linear resolution in the inner grooves ("inner groove distortion").

I trust that reasonable vinyl lovers can appreciate the fact that a digital system does provide significantly higher fidelity. I hope the images above put into context the magnitude of differences regardless of subjective preferences. Given the countless contributors to sonic quality with vinyl playback (wide range of speed stability with turntables, all kinds of cartridge characteristics, electrical matching of low voltage signals, potential need for step-up transformers, preamp RIAA compensation accuracy...), in comparison, digital can achieve high fidelity with much less fuss.

Oh yeah, before I forget, raise your hands also if you enjoy taking out your micro screwdrivers, maybe magnifying glass, tonearm tracking weight scale, and protractor for cartridge alignment/set-up. I don't think I see many hands up.

However, just because technically it is superior does not mean that the encoded music takes advantage of this fact about digital. I readily admit that I have many albums especially from the 1980's where the LP version sounds much more "full bodied" rather than the "thin" sounding CD, likely due to limitations of the early digital mastering techniques. Likewise, over the years there are many LPs with audibly more dynamic range than their digital counterparts - ironically perhaps because of the inherent limitations of vinyl and the need to avoid excessive signal amplitude. Remember that it's a bit of "give and take" because even with a dynamic master, limitations in temporal stability, surface noise, material imperfections can still distract from the enjoyment of an LP. Of course, I do not discount the idea that some folks actually like the "euphonic" distortions (which for me is the mark of the "euphonophile" rather than the audiophile :-).

Having discussed and shown the graphs and charts, let me also say that personally, I like vinyl. It's actually rather impressive and a testament to the precision and painstaking evolution of turntable technology that we can even achieve the quality of LP playback we have. I still purchase mostly used titles and in fact earlier this week I went into the bargain bins to grab some almost mint U2, Tom Cochrane & Red Rider, Warren Zevon, and Sinéad O'Connor. Good to "recycle" plastic. It's fantastic that I can pick up some early pressings of childhood and adolescent favourites for literally pennies on the dollar (luckily I have some very good used album stores close to home)! I'm very comfortable with saying that I like collecting LPs for the memories, nostalgia, and album art; the sound quality for the most part isn't the primary draw.

Finally, I think vinyl playback in the age of ubiquitous digital adds to one's "character". And some adherents/evangelists are certainly "characters" indeed :-). To have a nice turntable, decent collection of LPs, and adhering to the obsessive ritual of good vinyl "hygiene" (I still use the 3M suction cup washing technique for the dirtiest albums) inevitably inducts one into the anachronistic order of audiophilia... Totally cool if this is something you want to get into.

As usual, I encourage you to do your own "real world" tests. I'd certainly love to see objective comparisons using other equipment!

Enjoy.

----------------------------------

New this week is the arrival of the remastered Sgt. Pepper's Lonely Hearts Club Band 50th Anniversary Remix. Yeah, it sounds pretty good. When digitally remixing the work, there's the opportunity to keep resolution higher with less/no generational loss (a nice "unveiling" of instruments and voices), lower noise, achieve better soundstaging with much improved studio technology, more precise centering of the lead vocals, etc... These are all notable - have a listen to the remixed "Within You Without You" as a nice example of the improved resolution. I agree with some other listeners that the bass "weight" has increased substantially as well. So far, pretty good overall.

But... (Sadly, there is a but here...)


Above is the 16/44 CD "Lucy In the Sky...", clearly dynamically compressed with a DR8 using foobar dynamic range meter plugin. Let's be honest, with 50 year old source material, this is not true high resolution - impossible. Imagine, in a perfect world with the Super Deluxe Edition Blu-Ray, we should have a 24/96 hi-res version that would squeeze out all the dynamics there is in those original tapes, right? Sadly, this is no perfect world and the opportunity to hear the album aimed not at car and mobile listeners has been squandered. The "high resolution" version on the Blu-Ray is just as compressed as above with all those "thou shalt not go above this point" peak limitations.

For those with a surround setup, make sure to spend time with that multichannel 5.1 mix (I agree with Mark Waldrep - it's great)! Even though dynamic compression has also been used, you get more breathing room with more channels, the surround effect is impressive, very enjoyable, and obviously painstakingly re-created. Kudos!

Looking at both the 24/96 stereo and 5.1 24/96 audio, objectively it's actually more like 16/48 resolution. Again, not surprising given the age of the source material... IMO feel free to downsample/dither for your music server - you're not going to hear a difference.

Oooo.... What have we got here?


Why that's of course the Audioquest Dragonfly Black... Now with MQA "rendering" ability through a firmware upgrade (not "decoding"). I'll put this device through its paces in the days ahead and maybe have a further look / listen / examine what MQA "rendering" means... Stay tuned.

Have a great week ahead everyone. Hope you're all enjoying the music!

ADDENDUM (June 3, 2017):
1987 first CD release of Sgt. Pepper's, waveform for "Lucy In The Sky With Diamonds". A time before "Loudness Wars" when digital music still had good "headroom".


50 comments:

  1. Nice article! But the Lucy in the Sky thing confuses me. Which version is what? Where did you get the uncompressed version shown in the upper pic?

    ReplyDelete
    Replies
    1. Hi Techland.

      The "LSD" waveform display is the new 2017 50th Anniversary Remix that came out last week. That's the CD rip. The Blu-Ray 24/96 stereo looks almost exactly like that (peak limited at -0.1dB or so).

      You have to go back to the 1987 CD release to see a time when there was not peak limiting applied. Not saying that the 1987 version sounds as clear or "unveiled as the new remix of course. (ADDENDUM above.)

      Delete
    2. I think you don't understand my confusion because you know all the details ;-)...but they are missing here, IMHO. The first wave from display would show left and right channel. As there is heavy peak limiting ONLY on the right channel it must mean that the 'left' channel (upper part of the pic) shows a different version of LSD than the lower version. And you are indeed talking about different versions without clearly saying what is what. If this is the same song and indeed shows left and right channel, then this quite unusual waveform seems to need a note saying that.

      Delete
    3. I see what you mean Techland.

      Each of the "LSD" waveform displays is the whole song left and right channels (top, bottom).

      I selected this track because it's so easy to see the peak limiting in the right channel much more than the left. And of course we don't see the same in the 1987 release where all peaks are allowed to extend to their natural termination.

      Delete
    4. Wow! That's too bad. I was about to rush out to buy the 2017 remix.

      The 2009 edition, while recorded a lot louder than the 1987 one, still doesn't clip like that (I'd post the wave form, if I could). You would think that, given the lengths they went to, to produce this remaster, they would have been more circumspect about the 'loudness wars'.

      Sigh. Maybe the 2027 release ...

      Delete
    5. It's just weird that the remix's right channel 'looks' so much worse than the left. Is that down to the original mix and remix putting certain instruments mostly in one channel or the other?

      Delete
    6. Those very loud segments belong to the chorus. Looks like they really saturated the amplitude in the right channel more than the left. Oh well.

      Delete
  2. Great Blog

    Hi Archimago. I really appreciate reading your blogs. The Vinyl / PCM comparison is great.

    Juergen

    ReplyDelete
    Replies
    1. Thanks Juergen.

      Appreciate your input over the years!

      Delete
  3. There is no question that digital playback is far cheaper and much closer to technical sonic perfection than the vinyl playback medium.
    However IMHO technical perfection does not always sound better and there are certain attributes with certain less technically perfect reproduction systems that yield an overall more pleasing sound- despite their imperfections.
    For example, tube amplification. Without question tube amplification is technically inferior to solid state amplification- S/N, frequency response linearity, THD, etc. But many (including myself) find more pleasing reproduction using tube based amplifiers because (for example) they offer a much more expansive, engaging soundstage.
    Playing vinyl on many occasions can sound better than its digital counterpart despite its obvious limitations.
    "Ritual" associations, mastering and recording quality aside, I have found repeatable sonic tendencies when using very high quality digital and vinyl playback sources that often tend to favor the vinyl medium (despite the fact that I WANT to love the digital).
    Vinyl often sounds smoother, more natural, less edgy, more organic and overall more pleasing.
    Digital at its best sounds amazing, but often seems "hyper real", analytical, edgy, hard sounding and forces a focus on the sound rather than the music.
    These are tendencies I have found with each medium and a very generous sample size.
    I often prefer the sound of vinyl despite its imperfections and often despise the sound of digital despite its technical perfection.

    ReplyDelete
    Replies
    1. Hi Anthony.

      Yeah, I have no problem with subjective preferences. My dad is very much into his tube amps, vintage speakers, although these days he doesn't play much vinyl any more since I set up a Squeezebox Touch in his system and he can stream off my digital library across the city.

      I agree with the sense of "smoothness" with LP playback. Whereas I enjoy a high-res digital classical symphony as sounding "real", I don't hear that "smoothness" as sounding real to be me though. I hear it as "euphonic coloration".

      If I use a digital photography analogy, I liken vinyl to the dreamy smooth tone of say a portrait after going through a digital plug-in that smooths out facial pores, wrinkles and other imperfections. Something like Imagenomic Portraiture... Yes, it can make portraits look wonderful, fantastic for wedding photography when you want to portray the idealism of the day! There's a time for this and a time for high-fidelity - warts and all.

      On a related note, some of the best vinyl I've heard are vinyl rips to digital :-). Add a small amount of noise reduction, some click and pop removal... Smoothness with the benefits of ease of manipulation in the digital domain!

      Delete
    2. Archimago,

      "Smoothness" is probably not the best description of the attribute I hear listening to vinyl that makes me prefer it so often because it implies a rounding off of things, treble etc. that gets rid of the warts.
      What I am hearing in more detail is a more life like naturalness in tone and timbre that only comes with vinyl and not other analog sources (for some weird reason). I listen to wind or percussion instruments for example and they sound closer to real life to me. Their sound has a "ring" to it that clicks the brainwaves into identifying it as close to the original. I hear a trumpet playing a sustained note when listening to vinyl and it's a "that's it" moment. It sounds right, correct, authentic.
      I hear this more often with vinyl. I do not hear this as often even with analog reel tapes which are making a push at audio shows. Digital playback and analog tape sound similar IMHO.
      Unfortunately to achieve this with vinyl (for me) required double the investment vs. digital for the vinyl system including turntable, cartridge and phono preamp, careful setup and adjustment etc.
      Is vinyl perfect? Heck no, it has all kinds of sonic issues and inconveniences and I can completely understand why people do not bother. Yet somehow the sonic virtues manage to break through those barriers to myself and others who hear the same way.

      Delete
    3. Thanks for the note Anthony...

      I'm wondering do you have a LP recommendation for wind instruments that's a favourite of yours? I'd certainly love to keep a lookout for it and experience how it sounds.

      Delete
  4. You can make digital sound vinyl-like or analog-like anytime with some DSP if that's the end goal. So much for purity of sound I guess.
    In the end whatever brings you happiness is good. I am just tired of hearing about perfection when the imperfect is the goal.

    ReplyDelete
    Replies
    1. But not quite. The appeal of the sound of vinyl lies in part in its ability to reproduce natural, pleasing timbres- the quality and character of the tone and pitch and not just a rolled off frequency response- which is the quantity of the tone. Altering digital reproduction with DSP affects the quantity of the frequency output but it retains the same quality. I have heard DACs that try and mimic the "sound" of vinyl and it does not sound good. It takes away digital's forte which is exquisite neutrality and detail.

      Delete
  5. One thing that vinylphiles don't seem to 'get': you can digitally record the output of an phono preamp (as Archimago has done here), and play that back, and it will sound *exactly* like playing the vinyl record itself (unless there's something badly wrong with your record/playback chain). That itself demonstrates what's going on re: LP vs CD -- it's not that digital adds or subtracts from the sound, it's that 1) the mastering is different and 2) vinylphiles (and tubephiles) are euphonophiles. A digital recording of the vinyl output captures both the mastering and the 'euphonic' distortion. Faithfully. The problem is not digital.

    ReplyDelete
  6. I believe we do "get" that. It just isn't relevant about which format sounds better more often.

    ReplyDelete
    Replies
    1. Unfortunately, many vinylphiles don't acknowledge that you can make an identical digital copy of a vinyl disc. They really do think that as soon as you digitize something you change the sound completely. They claim the technology is broken. None of them have actually tried it out and done a blind test – they just regurgitate what they read in the audiophile press: "Digital can never sound as good as analogue". None of them have looked into how the technology actually works – they just regurgitate what they read in the audiophile press: "Digital is a bunch of square waves", etc.
      On the website Computer Audiophile there was recently a big debate, because someone had written a guest editorial where they also quoted Stereophile writer Art Dudley saying that vinyl was superior because digital technology wasn't capable of this and that, which was something that was quickly disproved in the comment section. In the end, the editor of the website posted a retraction of the article (although he didn't write the article).
      But Anthony, I'm glad to see that you acknowledge that you simply like the sound of the vinyl material and your playback chain rather than claim superior fidelity. That's what certain audiophiles do, and it's impossible to argue with them.
      I have compared around 800 albums on vinyl and CD, and some albums I would never, ever buy on CD. It's especially CDs from the 80s. They sound cold, thin, shrill and clinical. Some also have a very unpleasant timbre, we agree about that.
      But it must be down to either poor mastering or (less likely) poor converters in the 80s, or in the case of reissues, master tapes in poor condition, if you can take an analogue master tape or a vinyl record and make an identical digital copy that nobody can tell apart in a blind test. And many vinylphiles really do think they could easily tell them apart ("It's so goddamn obvious!"), and as far as I know, the ones who have tried have all failed (and I know of at least 5 tests ranging from 1980 to just a few years ago, the latter being done on $200,000 speakers and a $130,000 amplifier and a $150,000 turntable).
      I do acknowledge that certain analogue or digital masters might have unpleasant timbres, but when you transfer it to vinyl it will have a nicer timbre, because the vinyl material as well as the playback equipment changes the sound. Certain cartridges also have a very different timbre to other cartridges. I've even tried to add record hiss to a digital file and found that I somehow liked it better, as there seemed to be more "atmosphere". But again, it has nothing to do with transparency and fidelity, but if it sounds good, it sounds good :-). So, I'm less strict than many objectivists, but I don't buy into the opinion that some people hold that "almost any vinyl record in the entire world will sound better than the CD version" – far from. But there should be more focus on what sounds good, because some records do really sound great while the CD counterpart sounds not so great, often because the album was poorly produced.

      Delete
  7. I have created a google doc of the 2009 24 bit file, at:

    https://docs.google.com/document/d/1yJIwUjm3jieSLsqj46TF-F7mU3Hvv3cHUlTnMyep-SU/edit?usp=sharing

    You can see that it is better than the new file, but has more clipping than the CD version.

    Ian

    ReplyDelete
    Replies
    1. Thanks Ian.

      Yeah, different mix but certainly they didn't push it to the same amount. It's interesting that the *left* channel in the 2009 and also 1987 that I showed above is louder during the chorus rather than the right channel in the 2017 50th anniversary.

      Delete
  8. I agree that the stereo mixes are a little bit too loud, i've noticed on songs like Lovely Rita & Good morning there's an amount of distorsion on the top frequencies when you're listening at low level. but on the bluray, the sound is fine and there is no distorsion at all. i hate the loudness war on old recordings.

    ReplyDelete
  9. CD contains format specific euphoric distortion too. Please check distortion vs amplitude in AD - CD - DA process. You will see that distortion distribution is very different from what was tipical in analog recording/mixing/publishing era. Both routes have advantages and disadvantages you can't compare them looking only at the half of the thing like here.

    ReplyDelete
    Replies
    1. Hey Nobody,
      Do you have links to this information? Would love to specifically explore the topic and see if the distortion in the AD/CD/DA changes you're speaking of are in the ballpark of vinyl limitations.

      Of course, like any technology, I trust the audio engineers are aware of the limitations of their digital audio recording and production system such that steps are taken to mitigate problems like clipping... And these days, 24-bits would be ubiquitous and will make a huge impact on studio productions in limiting distortion.

      Delete
  10. The clipping in the remixed "LSD" actually makes me angry. I have recently listened to "remastered" versions of the Beatles albums, too, and I concluded sadly that it was making them sound too 'dry' and edgy.

    It looks as though it could just be that they have been screwing up the sound - I really trusted them to not do this! If so, can we trust them to have made good, clean, non-clipping digital archive copies of the original multitrack tapes? I assumed they knew what they were doing but I'm not so sure now!

    Time to dig out my old Beatles CDs, I think.

    ReplyDelete
    Replies
    1. Some further info. A kind person on my blog has commented with a link to a video about how they created the remixes. The clipping was deliberate policy for the digital version, but they left the vinyl version unclipped i.e. they have got it completely the wrong way round in terms of the capabilities of the two mediums.

      Delete
    2. It should also be pointed out that clipping (even "soft" clipping) is not just a change of dynamic range, it is also the generation of gritty distortion.

      Delete
    3. >The clipping was deliberate policy for the digital version, but they left the vinyl version unclipped

      Of course the clipping was deliberate policy.

      The question is why the $%#@ would they adopt that policy.

      It seems like a bizarre audiophile parody of reality: the masses want DR-compressed CDs. Audiophiles will pay a premium to get less-compressed vinyl.

      So, despite the inherent limitations of vinyl, the vinyl masters end up being of higher quality. That's just insane....

      http://dr.loudness-war.info/album/list/1?artist=Beatles tells a pretty lamentable story.

      Delete
    4. Thanks for the DR Database link blog.

      Yes, it is sad isn't it?

      I can comprehend that the record companies may want to release compressed masters for the mobile headphone crowd. But what I find abhorrent and unforgivable is to use the same mastering with hi-res downloads and in the case of Sgt. Pepper's the Blu-Ray 24/96. Pawning these off as somehow significantly superior; and the ridiculous audiophile press going along with it as if they hear *that* much difference while literally wasting the consumer's money and storage space.

      Even if the compressed 24/96 version might reduce some harshness compared to the 16/44 (I'm not convinced this makes much difference), that's really not what audiophiles should be getting if paying premium prices over a simple CD release!

      Delete
    5. I think it is important to emphasise that the DR database and your DR tool don't really work and mainly mislead. Your reliance on them is disturbing. I expect it of most audio internet fans but not of you! ;)

      Some discussion of this issue: http://www.stereo.net.au/forums/topic/54861-dynamic-range-scores-a-cautionary-note/ Read the opening post, but most of the ensuing discussion is argumentative nonsense.

      And in particular I commend you to read this, from within the above thread: http://www.stereo.net.au/forums/topic/54861-dynamic-range-scores-a-cautionary-note/?do=findComment&comment=1071439 (I quote you!)
      http://www.stereo.net.au/forums/topic/54861-dynamic-range-scores-a-cautionary-note/?do=findComment&comment=966872


      Also, regarding the 'obviously clipped' sections, the above links discuss the dangers of looking at whole-song waveforms on a scrunched timeline. Try expanding the timeline until you can see individual samples and tell us if they still look clipped, and for how long (a few samples would be completely undetectable).

      cheers

      Delete
    6. Hello tnargs,

      I use the DR as an easily accessible way to get an idea of the loudness; that's all. I don't think there's an over-reliance and I am aware of the ability to mislead. Sure, a system like R128 could be better but I certainly don't see a concern with anyone asking that their "hi-res" music have a DR value >10! Especially for a "classic" rock album created in 1967 with the sonic signature of the times.

      The recent Beatles is excessively loud with no variation between the CD and 24/96 release; and that's my main issue with it.

      If you do a search, I don't think I accused it of clipping at all which is of course a technical term and you're right, I don't see it when zoomed into the waveform.

      Subjectively when listening, some parts like that chorus in "LSD" was too "in my face" and I thought could have been toned down, allowed to be a bit nuanced in the vocals and sounded less harsh. For example Ringo's 3 hits on the bass / floor tom just before the chorus vocals sounded on the one hand much deeper and dramatic in the new mix, but I'm hearing some fuzzy distortion to it compared to the 1987 release (which is much leaner but sounded more realistic). I can't help but wonder if toning it down a few dB's and allowing the peaks to extend might have reduced the fuzz yet sounded just as deep and dynamic.

      Ultimately, for folks with hi-fi systems with plenty of power to spare on the power amps with speakers that are not afraid of the volume knob up higher, wouldn't it be nice to feed the system music with higher DR's? Especially for "high-res" Blu-Ray 24/96 versions!?

      I really enjoyed a few years ago when Paul McCartney released albums in high-res as the "Unlimited" versions as recognition that with high-res, maybe we can give the more discerning folks something different, if not technically better. That's what I would have loved to see with this.

      Delete
    7. BTW, this is also the point of my suggestion years ago that there be "Standard Resolution" CD and "Advanced Resolution" Hi-Res masters:
      http://archimago.blogspot.com/2015/12/musings-wisdom-of-simplicity-re-hi-res.html

      Heck, even if one knows that the "Advanced Resolution" doesn't need 24-bits and 96+kHz sampling rate, at least knowing that an audible effort was made not to push volumes up for headphone-in-crowd listening, could make the extra expense at least be worthwhile.

      Delete
  11. Thank you very much for this brilliant blog, which I only found out today. I bet most audiophiles are ready to burn you at the stake, haha! Keep up the good work and greetings from Lisbon, Portugal. Pedro.

    ReplyDelete
    Replies
    1. Hey there Pedro from Portugal.

      Considering it's the 500 year anniversary of Martin Luther and the Protestant Reformation (among other historical instances of stake burning), I hope the audiophile world doesn't need to engage in such barbarism. :-)

      I'm sure certain individuals, websites, manufacturers and publications do not like what I have to say. I believe they have to wake up and realize that perhaps a good number of audiophiles are actually science based and objective individuals when it comes to knowing or wanting to know about what they're interested in purchasing!

      Hmmm, I wonder if polls might be useful to get a sense of how audiophile winds are blowing...

      Delete
    2. Well perhaps I should've written "a lot of audiophiles" instead of most. :)

      Delete
    3. :-)

      Either way, I do hope that in time audiophiles choose knowledge and understanding rather than emotionally-based opinions and faith-based beliefs... We all have a role to play with transforming the hobby I think.

      Delete
  12. As usual - I'm late to the comments party! I am a hybrid between music-phile and audiophile. Basically it's the music that matters to me, but I DO need a certain level of audio quality to really enjoy it. So I have a decent system - a mix of tube and SS that I have put together with much experimentation and A/B testing over the years. I have had (and listened to) an enormous quantity of lp's, having been a large dealer/exporter to Asia. So I can say, with confidence, that sonically, the vast majority of lp's Suck (with a capital S)! Out of probably 50,000 lp's that I have chosen to give a listen (and over 100,000+++ that I skipped because I know better), I have kept about 2,000 lp's so far. So maybe 1% meets my audiophile threshold for repeated listening. But IMO it's not necessarily the fault of the medium as much as the care/skill of the people in the production chain. Of course the vinyl has physical limitations and inherent distortion that digital doesn't - but it can be darn great enough with the right mastering/cutting/vinyl etc. I have some lp's - especially test pressings, 12" 45 rpm's, and lacquers - that are AMAZING sonically. Not perfectly accurate from an "absolute sound" criteria, which is impossible with vinyl - but just plain amazing to listen to. That's what I really want. Life is too short to worry about whether it's accurate - I prefer AWESOME!
    I mainly play and collect records... but not because I have an illusion that they are better than digital. Rather because they're fun and the ones I keep are "good enough". It's fun to hunt for them. It's fun to hear music you never knew existed. The covers are fun. And it's fun to get an occasional $100 or $200 gem for $1. And too, many great ones are not available in digital or only in badly done, nasty-sounding digital.
    That being said - IMO, I can take any lp, record it at 5.6 DSD with software-based RIAA correction, load into Izotope RX, then Adobe Audition, export to 24/96 - and produce a digital copy that is far better listening than the "original", by removing vinyl noise, tweaking eq, etc. Then digital playback, at high volume (which I tend to favor :) ) doesn't have acoustic feedback issues like turntables do. Again, I love playing vinyl, but I find it hilarious when reviewers worry about digital jitter while their tonearm is vibrating like the handle of a belt-sander!
    Now, reel to reel tapes are another story... :) they are imperfect in different ways than vinyl (as is the playback equipment), but a higher percentage are good to begin with and their imperfections are quite repairable. I have been digitally "remastering" my tape collection and the results are fantastic.
    btw - I maybe disagree with Archimago re: Arthur Salvatore... I find the opinions on his LP list to be extremely accurate - and I have listened to most of them. Far better than the TAS gibberish.

    ReplyDelete
    Replies
    1. Thanks for sharing Chip. Wow. That's a lot of hours spent on LP playback! Interesting estimate as well of 1% meeting your expectations of sound quality.

      No debate, clearly the engineering skills and production quality makes a huge difference. And that kind of experience and ability would be doubly important for vinyl pressing!

      Re: Arthur Salvatore... I certainly don't disagree with everything he says (just like I don't typically disagree with *everything* any other audiophile writer says either). I can't fault his LP selection since much of it is subjective, but also parts of his "Audio Philosophy" certainly makes sense. Some things are certainly questionable such as his argument about digital having "very high sound-floor" which seems to be fundamental to his view of the digital/analogue divide. Way too simplistic (and IMO wrong) by not taking into account the big ticket item you pointed to - the quality of the mastering whether it be vinyl or digital.

      Delete
  13. As an additional follow-up from a vinyl proponent I just received a new remaster LP by Jazz piano player Oscar Peterson (The Oscar Peterson Trio, Live in Cologne 1963). It included a free MP3 download.
    After listening to the LP a number of times I downloaded the MP3 files and uploaded them to my Sony HAP-Z1ES file player.
    To compare the sound of the LP vs. the MP3 is valid because the mastering is identical. Even playing the MP3s through an excellent sounding upscaling digital file player, the sound quality of the vinyl is absolutely superior and more listenable than the MP3 files. The sound of the LP is more open and transparent and has a more extended treble. Basically much more presence in the midrange and treble that makes for a superior, more life like clarity and sound (admittedly my vinyl system is more expensive).
    Many may have already realized this with respect to MP3s and will cite their inferiority vs. FLAC and DSD files. However, for those that did not....
    Is the digital community prepared to say that MP3 files are junk?

    ReplyDelete
    Replies
    1. As your mp3 download is most likely not a recording of the vinyl record, a better way of concluding whether mp3 is junk or not is to take the CD and rip it in mp3 format yourself and then do a blind test between the CD and the mp3. Did you try this? I've done it with Foobar's ABX plugin (free player), and I couldn't pass a blind test between 320 kbps mp3 and wav. Most other people have failed as well. But I easily passed between 128 kbps and 320 kbps. But I'm not an mp3 peddler per se – I prefer CDs (the actual physical disc) and don't see any point in "settling" for less if you actually have a CD player.

      Although I have no idea what equipment you use, then I do know from experience that a lot of analogue equipment, cartridges especially, is specifically made to NOT be transparent. Certain Grado cartridges have a dip in the treble because certain people like that. Then there will be less "presence" with such a cartridge and people conclude that vinyl sounds "warm" (meaning less treble). Many other cartridges, especially very expensive ones, have a lift in the treble to give more "presence". It's basically equalization. So if I took a CD and ran it through an equalizer I would get the same effect, and I have tried it.

      These last couple of days I have actually run recordings of my own Rega Exact through the equalizer in Wavelab and made audibly practically identical copies of a Shelter 501 cartridge, a Zyx R100 Fuji cartridge, a Soundsmith Aida cartridge, and lastly a Lyra Skala. Actually, for the Lyra one I didn't even have to use EQ, but there was a slight difference – the Lyra one had a slight treble dip compared to the Rega Exact, which already has a treble dip (it's not flat). All this was done with just one change in EQ (either a spike at 5 kHz for the Shelter, a spike at 3.5 kHz for the Soundsmith, or a spike at 6.6 kHz for the Zyx). I was actually a bit shocked to especially find that the Lyra and Exact were so similar, although the Lyra costs 10 times as much!

      So, my point is simply that it's likely that your cartridge has a treble lift, as this is very, very common, especially among MC cartridges. The only way to estimate transparency would be to compare the master tape to the CD/mp3 version and to the vinyl version, played with transparent equipment.
      Unfortunately, often the frequency response measurements that are included with certain cartridges are misleading as well. They often show a completely ruler flat line, but when owners of these carts make measurements, the cartridges are not flat. It's not just down to averaging and smoothing used on the graphs – it's much more than that. So, I trust a jaggedy, uneven line from an amateur more than a ruler flat line by a manufacturer.
      But I don't judge you for preferring a treble lift – many people do, myself included in certain cases (my stereo system could also be called slightly bright).
      It is also possible, maybe even likely, that the cutting engineer used EQ when cutting this particular vinyl disc. Yes, some discs are cut flat, and while I don't dare estimate how big a percentage of discs is cut flat and how big a percentage is cut with the use of EQ, then I do know that it's very common to use EQ.

      As for Oscar Peterson, I have his album "Canadiana Suite" on the original UK 1965 vinyl edition, the CD edition from the late 80s and then a Japanese vinyl edition from 1976, and the Japanese vinyl edition is the best one – really a beautiful sounding album. But they probably did *something* when they cut it. And that's okay, 'cause they made it sound better :-). But we just have to remember that they still did something. Alternatively, they did *something* when cutting the disc in 1965 or when mastering the CD in 1989.

      I also posted a comment to one of your comments above, and as I also said there, I'm glad to finally encounter a vinylphile who's actually reasonable, sensible, honest and calm about his preference :-).

      Delete
    2. Thanks Anthony and Anders,
      Great discussions guys! Appreciate the thought put into the comments. Clearly many of you guys have had years of experience determining preferences for yourselves.

      There are just so many variables in play when it comes to preferences, especially when we're talking about vinyl playback with all its combinations and permutations of hardware over the decades. It's one thing to claim personal preferences and another for some writers to declare silly idealism about vinyl playback...

      Like I said, I have no problem with believing that some LPs sound better than their CD/MP3/hi-res counterpart. Mastering versions, EQ, production chain differences can account for these differences which we as end consumers would never know about.

      Anders - would love to see your work if you have a blog or place you're posting these tests with different cartridges!

      BTW: For those who haven't seen the article, here's an example of the kind of extremes some LP people go to; the controversial (to say the least!) "Hot Stamper" from Better Records:
      https://www.wired.com/2015/03/hot-stampers/

      The idea of very different sounding pressings is a good reminder that vinyl is a highly imperfect medium. That's certainly not good in my point of view. From the article, if indeed "The goal is flat frequency response, getting as close as possible to the sound on the original master tape," then we really should be just aiming for the best digital transfer shouldn't we? Especially these days when the majority of LPs come from a digital master...

      Delete
    3. I don't have a blog with my findings, but I can send the sound files to you by Wetransfer or something if you like - maybe you could even turn it into a post on this site one day :-).
      I haven't done any measurements or more objective testing, so my 800 comparisons between vinyl and CD are purely subjective, but I have kept a log of it with comments (although often in Danish, as I'm from Denmark). As for the frequency response on cartridges, then I've looked up many measurements, but they are in general difficult to find, unless you're looking for a common cartridge.

      If you see this, Archimago, and while we're on the topic, are you able to re-upload the soundfiles from the cartridge blind test you did a while back? You did it for me once before, but I didn't see it in time, and then the files were removed due to inactivity :-/.

      I saw the wired article about hot stampers a while ago, and I'm not too sure I give much creedence to that guy. I'm wondering how much of it is really just in his mind. In any case, hundreds of dollars for a Fleetwood Mac album that is $2 everywhere else is just ridiculous.

      Delete
  14. "However, just because technically it is superior does not mean that the encoded music takes advantage of this fact about digital. I readily admit that I have many albums especially from the 1980's where the LP version sounds much more "full bodied" rather than the "thin" sounding CD, likely due to limitations of the early digital mastering techniques. Likewise, over the years there are many LPs with audibly more dynamic range than their digital counterparts"

    BANG! that's all here, no need to comment :) If you love music you will always look for better sounding support/media, and if you love music released before digital era you definitely need vinyl and the closest release to original tapes...that's all, there's nothing magic about vinyl, it's just that they are the closest thing to master tapes (now lost or degraded). But I think that, in 1000 different ways, computer/digital mixing and mastering in some ways make innatural sounding, you need just a simple click on a mouse to destroy a music recording...maybe in vinyl years there much less chance of making mistakes and a lot of skilled and well-paid people working in the industry.
    Music is not just computer measurements and technical illustrations, fortunately ...

    ReplyDelete
    Replies
    1. Yes, some old albums would be closer to the source on vinyl because the master tapes are in poor condition now. Other albums are still closer to the source on CD, because the master tapes are in good condition.
      Although mastering engineer Steve Hoffman is not my favourite source, he nevertheless said that the problem mainly arose when they switched from natural lubricant to artificial lubricant in the 70s and that caused tapes to deteriorate quickly. So older tapes still sounded fine, while tapes from the 70s onwards sounded not so good. I even read about a Dire Straits recording that started to deteriorate during the recording process, but I don't know if that's really true or not.
      However, I know that Moody Blues' "Days of future passed" from 1967 (before artificial lubricant) was already in so poor condition in the early 70s that it had to be remixed. The original LP sounds dark, murky and old, while the CD version and the SACD (from the new mix) sounds much, much nicer. But again, it's a new mix, so...
      I have heard many, many albums from the 60s and 70s that honestly sound anywhere from poor to terrible on the original vinyl edition and then sounds great on CD.

      To make any general statements we would really need to hear the master tapes, in pristine condition, but we can say for sure that vinyl is not closer to the source in any way, although certain people like to think (and scream) that it is. But vinyl might sound "better" if the master didn't sound good to begin with. Unfortunately, many albums are not well-produced. For other albums, some strange mastering choices have simply been made. Hoffman also remastered several Nat King Cole albums, and "The Nat King Cole Story" on Analogue Productions is a good example of how the song "Smile" sounds great on that album, and less than stellar on the album "The Very Best of Nat King Cole" on Capitol from 2006, but that 2006 version still sounds better than the original LP (I've had both the original US LP and the UK 10"). According to Hoffman, the engineers who remastered Nat King Cole's work in the 90s and onwards tried to get rid of that "old" sound.
      So, sometimes some not so great people are put to work on destroying the music. But it's purely that, and it has nothing to do with the use of computers and digital, because you can take any source and copy it to digital and you wouldn't be able to tell them apart. Several blind tests have shown that. You might FEEL that it's "fake" to have a digital file rather than a physical tape, but if you can't tell them apart, it goes to show the transparancy of digital, whereas tests have shown that most people can hear the difference between an analogue master tape and an analogue tape copy, as well as a vinyl cut done from the original tape. Neither analogue tape or vinyl are transparent or "high fidelity". But it can sound great under the right circumstances.

      Delete
    2. By the way: I've heard loads of CD reissues that have just been dubbed from the vinyl record - either because the tapes were lost, or because the reissue company didn't care, or didn't really ask for permission.
      I even have heard a couple of CDs that were probably made from casette tapes! Luckily, some of the most expensive records I ever bought ($200 and up) were done well on CD and sounded better than the record.

      Delete
    3. I didn't think there was any question that vinyl is noisier than digital. Nice to see it visually, but you can easily hear the difference.

      You can spend just as much on digital equipment as analog. I can find a cheap turntable and I can find an expensive DAC.

      What should be analyzed if possible is why an analog sourced vinyl record has certain sound characteristics while a digitally sourced recording has certain sound characteristics. There are differences in how these two technologies sound to the human ear. Can those differences be quantified and measured? One is not better than the other from a listening perspective, they're just different.

      Delete
    4. Hi John,
      While you could be right about analogue vs. digital sourcing of the music, I'm not sure how one could test this without many other variables in the mix.

      These days, almost all new recordings are digital of origin. And as for old material, do we actually have any albums which are the exact same mastering where:
      Sample A: direct analogue master tape --> LP
      Sample B: analogue master tape --> digital intermediate --> LP

      Short of having one's own cutting lathe, I'm not sure how the test can be done. In any event, the transfer of that electrical signal to a physically imperfect medium will being anomalies to the original signal. It's ultimately all about whether one likes those distortions or not I think!

      Delete
    5. There are some examples of new analog sourced recordings. The most recent Beatles mono albums are analog, all of the Music Matters Blue Note remasters are analog, much of Analog Productions albums are analog. However, most remasters are cut from a digital source. It's unfortunate because there is a clear difference in the sound. The Berliner Philharmonic recently recorded some Brahms performances direct to vinyl but that was a rare occurrence.

      What is it that makes analog sound full or fat and digital sound thin? If there is there a lot more of the performance in the digital recording shouldn't it sound fuller? Is it the noise we hear in the analog source that makes it sound full or smooth?

      Delete
    6. John, sadly, your comments are full of statements that are highly presumptuous and completely contrary to fact. And when the factual errors are removed, the remaining statements are trivially explained with psychoacoustics. In other words, bias. Your ideas that digital has a characteristic sound don't stack up. You are attributing psychological phenomena to sound waves.

      Delete
    7. This comment has been removed by the author.

      Delete
    8. @tnargs why such attempts with words like "sadly, presumptuous, contrary to fact, factual errors, trivially explained" to demean? Now I'd like you to back up those statements with your facts.

      If you think digitally sourced or digitally played back music sounds no different than analog I'll leave you to your opinion. I'm not looking to go down a path of trying to convince someone there's a difference when they can't hear it for themselves. And if you think I'm biased for thinking that, then I'll contently live in my bias with the rest of the world.

      Delete