Readers vs. Analog vs. Digital

Compiled by Lukas Kendall

Wait, non-tech people, before you tune out, thinking that this is another boring argument of sound reproduction (it is!), I have some information that might be interesting.

First of all, Varese is reissuing 1941 (Williams) in around six weeks. It will not have any extra music, however. I¹m also told the difference between the old 1941 Bay Cities and Alhambra CDs is that the Bay Cities disc had much better sound. The Varese CD will come from the Bay Cities masters.

Last Thursday¹s Variety had a short article on how recent soundtrack albums‹you know, the big pop things‹are selling. Check out these figures of units sold, as of mid-August:

Space Jam‹3.7 million

Men in Black‹1 million

Batman and Robin‹531,000

My Best Friend¹s Wedding‹363,000

Nothing to Lose‹288,000

Hercules‹227,000

Spawn‹183,000

The Lost World‹66,000

Austin Powers‹64,000

Con Air‹29,000

As readers well know by now, MIB, My Best Friend¹s Wedding and Austin Powers have brief score cuts, and Batman and Robin has a cut from Elliot Goldenthal¹s Batman Forever. Hercules has score as well as songs (it¹s a musical, duh), and only Lost World and Con Air are pure-score albums. Lost World, incidentally, is being considered a disappointment for MCA.

Basically, we can deduce the following: song albums, even to dud films, can sell hundreds of thousands of units. Score albums are lucky to crack 50,000 and many won¹t even get near that. So, before you say, how dare some record label put out a song compilation instead of the score that people (i.e. us) ³really² want, consider what you might do if you were a record executive. Let¹s see, sure 250,000 seller one way, sure 25,000 seller the other.

Okay, onto the technical stuff. Two responses came in of some length to the second part of Josh ³Swashbuckler² Gizelt¹s article on analog vs. digital:

***

From Michael Lee, leeji@ecf.toronto.edu:

I just want to point out two technical errors in the article Digital vs. Analog Part Two by Josh "Swashbuckler" Gizelt:

Since the original wave is being cut up, some sounds will "fall through the cracks," so to speak. How much this happens depends on the bit-rate being used.

It also depends on the sampling rate. Currently the CD standard is 44.1KHz. That means a CD recording captures all audio signals up to about 20KHz, close to the Nyquist rate. On the other hand, number of bits (16 in the case of a CD recording) in each sample determines the dynamic range of the recording.

Because the original sound wave is being approximated, the sound has essentially been diced, and when it is put back together, it is stepped. This means that the curve of the wave is no longer smooth, but instead somewhat jagged.

No, the approximated sound wave is not stepped! It is a perfect, smooth sound wave. When recording in digital, the original sound wave is first filtered by a low pass filter with the cut-off at about 20KHz (in order to prevent a phenomenon called aliasing during sampling). Then this new sound wave is sampled and each sample is further encoded to a 16-bit number (this is the digital-to-analogue process). During playback, the 16-bit numbers are decoded and the EXACT sound wave AFTER the filtering mentioned above is reconstructed. So what we hear is the original sound wave without the components at frequencies higher than 20kHz.

From Chris Malone, malone@ist.flinders.edu.au:

I do agree with many things Mr. Gizelt was saying but I'd like to point out a few things:

There Is Rarely "Presence" ...in a digital recording.

I think you'll find there can be. By making a digital recording doesn't mean an engineer will totally change their miking style or mixing intentions. The "presence" that Mr. Gizelt refers to is analog warmth that exists on 'great analog' recordings.

The extent of the problem is lessened (but not eliminated) by increasing the bit rate. The most demonstrative evidence is the metallic sound in the strings in many digital recordings.

Sure but if you heard the recording on ALL sources, analog and digital I think you would find that it would be something other than the digital recording causing the "metallic" sound you mention.

Early digital recordings suffered from "bad" sounds because‹let¹s face it‹it was a new medium and we didn't know how to deal with it properly. As digital recording evolved we are now learning more about how to deal with it. We are still learning about it and if you listened to early CDs they weren't mastered well anyway so you cannot blame a digital recording in that case.

On analog recording:

Generational Loss: When an analog signal is reproduced, each generation it goes through introduces anomalies and information loss, and decreases the signal-to-noise ratio. Anyone who has a copy of a copy of a copy of a tape would know what I'm talking about. By the time you get to the fourth generation, the music is less distinct and the there is tape hiss up the wazoo (of course, none of us have ever copied a commercial tape... RIGHT?).

With each copy of an analog tape (I'm talking open reel at say 7.5 IPS) there is a 3dB increase in noise. When making an analog recording, mixed to analog reels and transfered to an analog master (which can be copied to digital tape for CD) there is a 3dB noise increase each time. Drop-outs and narrowing of stereo image are big problems too. The more you copy a tape, the narrower the stereo image becomes. Also azimuth problems can severely plague a tape copy. Azimuth is the skew of the head in relation to the tape. This can create problems with upper end frequencies and music coming from one channel earlier than from the other. This is why azimuth must be checked and adjusted, if necessary, during mastering. Wow and flutter is also accentuated, but this is where digital is better, negligable wow & flutter figures, just like Compact Disc.

This becomes a serious issue when an analog recording is mixed. Since every time you make a copy of the sound it is further distanced from the original, a recording that requires many mixdowns is going to display serious problems. Recording musicians will recognize this as being the most formidable problem with "bouncing" tracks on their 4-track tape decks.

Tell me about the limitations of 4-track decks!!! ;-) I know exactly what you're saying. Analog recordings have a warmth that usually must be created otherwise in digital recordings. The good thing about analog is that levels can be run 'hotter', into +3dB or even more without any bad tape saturation. If you were to do this with a digital recording it would clip badly and sound God-awful. By recording analog you can make good use of the natural analog compression that occurs when you run levels slightly hotter.

In my view the best compromise is to record analogly, mix to a digital medium and master to digital (CD/whatever). You benefit from the analog properties you mention but you don't have the bad S/N ratio of multiple analog mixdowns and the signal is virtually only 1-generation old.

There really does need to be another letter added to the AAD, ADD, DDD code. It's always XXD because CD is a digital medium but it doesn't take into account analog or digital mastering in the case of those severely limited 1970s master tapes‹Superman: The Movie for instance! ;-) It should be, for Superman AAAD. Analog recording, mixing, mastering and to CD. Star Wars SE would be more like AADD, analog recording, mixing and digital mastering. The stupid re-mixes of the Empire Strikes Back cues would be ADDD, analog recording, digital mixing, mastering, transfer. You get the idea.

In digital reproduction, this is not an issue for the same reason tape hiss and surface noise is not a problem. You are not reading the digital tape itself, only what is encoded on it, so every copy made is identical to the original.

I'd tend to agree, but I think there are differences between two digital copies of a DAT tape for instance. Error checking bits, etc. add small alterations apparently. I've only read of this so I don't know about it first hand.

***

Okay, let¹s take it to rec.music.movies.digital-analog.trivia.technical. Tune in tomorrow for the next exciting episode of "Film Score Daily"!

Lukas@filmscoremonthly.com