I've always been a Hi-Fi enthusiast. If I hadn't got into the trade of organ building, there's no doubt that I would have made audio my career in some way. What follows is a comparison, sometimes geeky, exploring the technical strengths and pitfalls of each of the mediums of compact disc and vinyl records.
The vinyl format is fascinating to me, and audibly performs a lot better than the physics would dictate! After 30 odd years of CD audio, the vinyl format refuses to lie down and die. Why is that? I'm even contemplating upgrading my turntable to maximise my enjoyment of the vinyl format. In the following text, I'm going to investigate the vinyl format compared to its CD counterpart. If you don't want to wade through the following technobabble, just go to the bottom of this page to hear some competently tracked vinyl numbers!
There's no arguing that on paper, the CD format for audio recording and reproduction is far superior to vinyl recording. The dynamic range – evident at quiet levels – was a revelation when the CD format came into being in the early 1980's. Let's look a bit further into the way that audio is digitised here.
Shown to the left is my tonearm headshell, proudly holding as its captive, the Denon DL110 which you will be able to hear later.
In line with the normal range of hearing (for a young person anyway), the CD format can respond from 20Hz to 20KHz (20 cycles a second to 20,000 cycles a second) in a pretty linear fashion. Let's look at the current sampling frequency of CD – 44.1KHz. At 20KHz, there are just 2 samples of this frequency taken per cycle – academic maybe, but let's look further down the audio spectrum. At a normal vocal level (around 1KHz), there are just 44 or so samples per cycle – frighteningly few for such an important frequency range. Thats's less than the British mains electricity frequency of 50Hz! During the digitising process, digital noise (called dither) is added to the signal to make the resulting product more linear. This helps to avoid quantizingerrors, which could result in a rather unpleasant sound in the finished product. It also helps to mask the relatively few quantizing points mentioned.
If I'm honest, you'll never hear different frequencies in isolation unless you're using a test record to set up a tonearm/cartridge. The sounds we record and reproduce in music are really intricate wave-forms – you only have to digitise, then stretch this wave-form out to appreciate how complex it really is.
I'm going to take yet another geeky side-step here. When the sampling rate was being established for digital audio, it was understood the sampling rate needed to be at least twice the highest frequency that had to be reproduced – if this were any lower, the system would suffer from a phenomenon known as aliasing. This means that you could capture a frequency, and make it appear a lot less than it is. The easiest way I can think of explaining this, is as follows... Imagine you have a wheel spinning at a constant speed, with a dot painted on its periphery. Now darken the room and put a strobe on it. By varying the speed of the strobe, you can make the wheel appear to spin faster, slower or even backwards!
Before I put digitisation to bed, I'll just say a bit about resolution. In the CD format, there are 16 bits used to encode the amplitude of the signal, This gives 65536 possible levels of dynamic (volume, or depth as it's sometimes referred to). In a truly analogue recording this depth is infinitely variable.
The vinyl format is, in my opinion, a very special case-study into what is possible in physics. The cartridge I currently use (a Denon DL110, high output moving coil variant) has a frequency response which extends to a quoted 45KHz, that's over the sampling frequency for CD's! Indeed, going back a few years, there was a protocol for quadraphonic encoding to vinyl that never took off. This protocol required a 30KHz carrier to be recorded onto the record (to de-multiplex the audio signal and apportion it to the correct channels).
There are swings and roundabouts to the vinyl format however. Because there is such a difference in speed between the outside and the inside of a record, there is a great difference in the potential dynamic range between the two. In days when vinyl was the norm, this resulted in the recording engineers and artists concerned, putting their “biggest numbers” on the first part of each side of an LP! This is not the case for classical music however, and there are compromises to be made to avoid the crescendos (that normally come towards the end of a lot of classical music) being distorted. In this case, the most dynamic passages can come at the worst possible time for the vinyl format. This is especially true in organ music, which has some horrendously impossible-to-follow waveforms, when the instrument is played at full-tilt! Some very judicious cutting levels have to be established for these cases.
In short, the shorter physical wavelength of the groove modulations at the end of each side of a record, combined with any differences of shape between the original lathe cutter and the playing stylus, conspire to make the last part of a record notoriously hard to track, especially when recorded at high levels. In order to minimise these aberrations, every aspect of the replaying system has to be set up just so - and also a stylus without too much wear has to be thrown into the mix. Get all this right though, and the results can be amazing!
CD's, being a digital medium, automatically alter the speed of rotation to keep the data stream coming from them constant. At the beginning, the disk spins a lot faster than at the end (they play from the inside outwards). This means there is no audible difference in performance throughout the whole playback.
Another awkward aspect of vinyl playback, is that the arm on conventional record decks traces an arc, whereas the cutter on the mastering lathe works in a linear fashion. While a great deal of this can be mitigated by having the cartridge offset on the arm, there are portions of the record which are mis-tracked due to this anomaly. Some record deck manufacturers have got over this problem by employing a tangental tracking approach. This ensures the cartridge is always at a tangent to the radius of surface currently being tracked. These decks are quite an engineering feat though, and can be prohibitively expensive.
One final point on the vinyl format. When a recording is mastered on the cutting lathe, the treble level is boosted and the bass level is cut... in pretty massive amounts. The reason for this is two-fold:-
1 – To keep the groove spiral as tight as possible, the bass level is cut. This is to restrict what would be enormous excursions on the physical wave-form, else the recording time would be really short (especially at 45 RPM). The bass is boosted on the replaying equipment.
2 – To keep the surface noise down, the treble is boosted. On the replaying equipment, the treble is cut along with the noise. In boosting the treble, significant energy levels are recorded onto the master (and all the pressings thereafter).
One negative effect of this high frequency boost, is that when played on some equipment, sibilant sounds – especially the esses and tees used during speech and vocals, can sound distorted and blurred. This is more often noticeable towards the end of each side of an LP.
But isn't vinyl currently mastered from digital sources anyway?
The answer here is almost certainly yes, but using professional equipment, with at least 48KHz sampling rate (sometimes a great deal higher), and at 24 bit resolution. There is no way that this can be produced in the current CD format. This 24 bit depth is 256 times more precise than the normal 16 bit CD depth.
So which is better, vinyl or CD?
Mmmm... Well... There is a great difference of opinion here, as this is a very subjective thing. You have to go a long way to better a clean vinyl pressing – I have to say that I'm just about in the vinyl camp right now. For technical specifications, CD is the best on paper. For musicality however, the vinyl route still seems to be the best. Audiophiles the world over seem to hold with this opinion too!
The truth is though, that vinyl has to have everything in its favour. There is nothing worse than a bad pressing, with pops and surface noise that interfere with the content. Another annoying thing is that record manufacturers don't always get the centre hole in the middle. That can lead to a phenomenon called wow. This means the pitch goes up and down with each revolution.
There is a format which should have both the standard CD and vinyl formats quaking in their boots – the DVD Audio format (DVD-A). This can offer sampling rates of 192KHz (for stereo), at 24 bit depth... serious stuff, but this format isn't commonly available for solely audio reproduction.
Below are some audio samples from vinyl. These even stand the scrutiny of being listened to with headphones! ...
These samples are in MP3 format. They are 320K samples, which makes them the highest quality currently available for this format. It may be an idea to save these samples, then listen to them on some decent gear (or use headphones on your phone/tablet)!
Some needle drops of a few vinyl recordings. All from 33 RPM albums...
If you waded through the technobabble above, you would have seen that vinyl can be somewhat brittle towards the end of a side. Here, the last bars of Beethoven's Ninth Symphony is very capably tracked by the DL110. The second side of the record (which this is from) is 45 minutes long, achieved by a very tight spiral in the quieter passages!
*Chicago Symphony Orchestra and Chorus, Conductor Sir Georg Solti. Soloists... Jessye Norman, Soprano; Reinhild Runkel, Contralto; Robert Schunk, Tenor; Hans Sotin, Bass.
Some naked samples from my needle drops...
These samples are snippets of the original raw transcriptions, directly off the turntable. There is no compression at all. These are at 24 bit depth. The percussive sounds, especially the ride and crash cymbals in the Joan Armatrading sample, sound really crisp.
...I'm still not convinced yet Matt - what's your candid opinion?
Okay, I can't be more frank than this... When Compact Disc was finalised as a viable medium for audio reproduction, the minimum possible technical criteria were sought. There were manufacturers that would even have accepted 14 bit encoding at the time (only a quarter the resolution of the present 16 bit protocol)! If the bar was set just a bit higher when digital audio made its way into our living rooms, vinyl would now be totally dead in the water. Put slightly differently (at the risk of employing tautology), if there had been 24 bit encoding with at least 96K sampling rate, this CD vs vinyl debate simply wouldn't exist.
There are CD players which claim to be 20 bit systems. These players don't read the discs as a 20 bit stream (as 20 bit CD's don't exist), but interpolate their 16 bit data in a very clever way, so the content would appear to have 20 bit resolution. This is much the same way that Blu-ray players can upscale a DVD to appear visually better. We all know however, that a genuine Blu-ray disc is much better to watch, with so much more visible detail - Quod Erat Demonstrandum!
In all the above, I'm not saying that compact discs are a bad thing. It's great to be able to listen to music without the distraction of the extraneous noises which come with a bad vinyl pressing - CD's are more consistent in that respect. As said earlier, this was a comparison, pointing out the strengths and weaknesses of the vinyl and CD formats. Many people find today's very compressed formats more than acceptable (MP3 files, and the vast content that can be downloaded from iTunes). Indeed, MP3 samples sound really good, considering a great deal of the (unimportant) detail is stripped from their content. That's another story that can wait for another day though!