I'm the first person to admit that digital still isn't capable of 100% perfect reproduction... but it's 99.8%* closer than anything else we've ever had available, compared to the 90%* on a really good day of analogue tape, or 70%* of vinyl.
If the material calls for a warm and fuzzy analogue feeling, record it that way... the digital format can easily capture and reproduce it!
*Made up statistics... but they feel about right! ;)
Hugh's guesstimate of 90% for analogue tape on a good day relates of course to the losses at each
tape generation. In practical analogue production it was usually worse. Just mixing down a multitrack meant a rerecording - one more generation, so we're already down to 81% - assuming an ideal transfer.
The mighty Nagra portable recorder was mentioned. It was primarily designed for location film sound track location acquisition. But a downside not always considered was that prior to the Nagra and similar 1/4" machines, the location dialogue etc was recorded to sprocketed magnetic tape which could then be placed directly on the Steenbeck editing table for sound mixing - no generation losses. So the new Nagra system, great as it otherwise was, required that extra generation of analogue tape losses.
Then as mentioned, that mixing stage itself required a rerecording - yet another generation. In practice, by the time a feature film came to our local cinema, we were hearing maybe fifth generation analogue tape sound, which could sound pretty rough.
When listening to digital versions of old film sound tracks we are often hearing the sound far better than we ever did in our cinemas, at least if production staff are skilled and had access to the earliest generation well preserved analogue recordings. There was a point to the little AAD, ADD, DDD symbols on early CD's...
It's easy to forget that DAW's now allow us close to lossless tracking, then lossless mixing and user copies. Even a good mp3 can be very close to the data uncompressed mix it was made from, if prepared with skill.
For fun, try doing what I did for a living in the 80's: audiobook production for the print handicapped - all analogue tape. For economy, user copies were cassettes running at half normal speed. Then to protect the valuable original master (first generation) from wear and accidental damage, we normally duplicated from a half speed cassette sub master (second generation). So the clients usually received half speed, third generation cassette copies - high speed duplicated at each stage. No third head "confidence monitoring" of duplication quality either.
Because of the very slow tape speeds and narrow tracks, losses that might be insignificant on pro gear at 15ips became critical at less than 1 ips: a dirty or worn high speed duplicator head, a slight azimuth misalignment, a slight over or under bias - or a combination of those. The highest frequency possible on the user copies was 5kHz. (Dolby HX Pro would have been nice). Actually often quite listenable on speech although some voices survived better than others.
The importance of getting it right at source? Critical. Band limiting, compression, de essing, big pop filters on mics for the worst voice over artist with the worst mic technique. Yet today we have the luxury of choosing which top shelf mic to use on our voice. With this audio book production, post production wasn't normally practical. You had
to get it "right" at source, for it was all downhill from then on.
Ah the romance of analogue tape production! And analogue tape was a great advance over the disc cutting that preceded it.