Are/were older recordings really so bad?

[rauschwerk]

I'm currently listening to the end of Mahler's 8th acquired this morning as a free mp3 download - Utah Symphony Orchestra conducted by Maurice Abravanel.

Surely you get what you pay (or don't pay) for. Unless a recording has been taken from the master tape, I'd avoid it. The sound on LP will almost certainly have suffered bass cut - very damaging in a work like this! As for YouTube, I would hardly expect the sound to be much good. I bet it is as doctored as hell.

[RichardB]

Of course it isn't surprising that recordings from the late 1950s still hold their own among more recent work - the technology of microphones and loudspeakers has hardly changed since that time in comparison with the developments they underwent in the first half of the last century, tape recorders had reached a high degree of sophistication and stereo recording had been pretty well worked out. The only significant improvement in recording technology since then has been digital recording of course. But when you compare developments over the 60 years before 1960 and over the 60 years since then, it seems that once the technology reached a certain threshold level many people don't find any additional improvements making such a huge difference.

[Dave2002]

....given a system good enough to reveal this. And ears to hear it, of course...

I’ll not go here again, but I’ve had years of ears and systems and there’s not much to choose. I’ll all subjective, depends on so-called ‘listening’ ( which means little).
After a shed load of hi fi I’m now on Edifier Actives which sound no different from ATC / PMC / Castle . I’ve no golden ears, neither has anyone else.

I wondered what the Edifier Actives were - so I looked - https://www.amazon.co.uk/Edifier-R18...079SKK2MN?th=1

While I don't believe in spending shed loads of money on expensive kit, I very much doubt that those speakers are in the same league as some of the "better" ones sold as high quality. Indeed a review of similar models (at the time only available in the US and Australia) suggests that these are speakers perhaps more suited to improving TV quality sound - https://www.techradar.com/uk/reviews/edifier-r1280dbs - so I'm not knocking them as good enough for some purposes - but I doubt whether they'll bring out the best in a really good recording.

However, they might be good enough to distinguish between what I call poor quality recordings and much better ones. The issue I was trying to raise in the original postings was that older recordings are not necessarily "worse" than modern ones - and indeed in some cases on very good equipment might sound as good or better than a modern recording. Also I was concerned that some distributors are selling (or even giving away ...) versions of older recordings which are not representative of the quality which was inherent in the originals. Some cheap collections may offer value as they may allow access to music or performances which would otherwise be hard or very expensive to acquire, but really some are just so dull that it really does raise the question to me of why that is done.

Is it perhaps to show that modern recordings are better - by ensuring that everyone "knows" that older recordings sound significantly worse?

Actually I rather doubt that - many sellers (of almost anything) have next to no interest in the things they sell - that's the nature of commerce.

[Bryn]

Surely you get what you pay (or don't pay) for. Unless a recording has been taken from the master tape, I'd avoid it. The sound on LP will almost certainly have suffered bass cut - very damaging in a work like this! As for YouTube, I would hardly expect the sound to be much good. I bet it is as doctored as hell.

Youtube audio is very heavily compromised by the data rate used. You are very lucky indeed to get anything over 128kbps AAC-LC. Most seem to use 96kbos AAC-LC and some music is crushed into 48kbps, and that's stereo. This, for instance, used 48kbps for its audio component:

[Dave2002]

Youtube audio is very heavily compromised by the data rate used. You are very lucky indeed to get anything over 128kbps AAC-LC. Most seem to use 96kbos AAC-LC and some music is crushed into 48kbps, and that's stereo. This, for instance, used 48kbps for its audio component:

Without knowing the work it's hard to tell how "bad" that is, but it doesn't sound great.

I don't think it automatically follows that low bit rate corresponds to poor quality - but that does depend on what is being recorded. We assume that the transmission/recording system should be capable over a wide range of dynamics and a wide range of frequencies. If those are not present in the original - or produced by the original instrument(s), then quality might indeed be good enough. However much music does have wide dynamics, wide frequency range plus ambience and just about all the subtlety and interest will be lost with low bit rate lossy encoding - whether mp3 or otherwise.

The odd thing about some of the damage caused by mp3 is that often it is the bass notes which lack depth, yet they ought to be the ones requiring only a few bits surely. I don't really know why mp3 not only trashes the frequency ranges, but also affects the dynamics.I do wonder if it's the mp3 compression which is causing the problems, or some other - possibly "human" intervention when doing the transfers and conversion to lossy formats.

[Bryn]

https://cs.stanford.edu/people/eroberts/courses/soco/projects/data-compression/lossy/mp3/algorithm.htm

[Dave2002]

https://cs.stanford.edu/people/erobe.../algorithm.htm

Sure - I knew most of that. The real issues for me are how and why mp3 screws up so much.

Some of the changes which mp3 does may be considered relatively benign. For example sometimes it appears to have the effect of smoothing out violin tones - which can be achieved by a filter. Some people may actually like that. I have heard recordings - say from DG - which had a sort of silky tone - not dissimilar to some mp3 compressed audio. However often the artefacts from mp3 are unpleasant - to my ears at any rate. Why it messes with the dynamic range and perhaps how it screws up the bass - I'm really not sure.

Given that the world has largely moved on and most people are not really interested in how and why lossy compressed audio works, and since non lossy formats are now much more prevalent than before, probably most people who want better quality will look for that - in various formats - while others will "put up" with mp3 or other compressed audio - and most of the time may not notice what they're actually missing. However, sometimes things are just so dire that one does wonder what the original might have sounded like.

I'll come back on this when the CD of the Mahler 8 arrives. I'm hoping it will sound orders of magnitude better than the mp3 and to some extent justify my claim that many recordings from the past were in fact pretty decent.

[rauschwerk]

Full digital audio is known to be an inefficient method of encoding music signals. That has been understood for a long time. Hence the evolution of lossless Codecs such as flac. Lossless codecs encode efficiently but crucially they do not discard any audio information to achieve lower bitrates.

As I understand it, mp3 does not deal with transients (plucked instruments, percussion) very well. It compresses file size by eliminating parts of a complex signal which it is believed the ear cannot hear. That is what is meant by 'compression' in this context. It has nothing to do with dynamic range compression. Because of this, mp3 can only really be judged subjectively.

I have encoded a string quartet recording with mp3 at various bit rates, and was surprised at how low a rate I found acceptable. Owners of other ears might well get different results, of course.

[jayne lee wilson]

Full digital audio is known to be an inefficient method of encoding music signals. That has been understood for a long time. Hence the evolution of lossless Codecs such as flac. Lossless codecs encode efficiently but crucially they do not discard any audio information to achieve lower bitrates.

As I understand it, mp3 does not deal with transients (plucked instruments, percussion) very well. It compresses file size by eliminating parts of a complex signal which it is believed the ear cannot hear. That is what is meant by 'compression' in this context. It has nothing to do with dynamic range compression. Because of this, mp3 can only really be judged subjectively.

I have encoded a string quartet recording with mp3 at various bit rates, and was surprised at how low a rate I found acceptable. Owners of other ears might well get different results, of course.

I recall when the lossless FLAC relays ran on Radio 3 from Spring 2017 to the end of the Proms, I made close comparisons on concerts from various familiar venues. Even with the usual 320 kbps AAC (technically and audibly superior to mp3) the leap in quality to lossless (which ran at various bitrates then) was obvious: the tangibility of instruments and ensembles, and the presence of a palpable acoustic they were performing in. This also meant that the climaxes in a Bruckner 4th I heard from the Barbican had a much freer, "soaring" quality, as they had more acoustic space to expand within. How sad they never brought it back...

Call this a subjective response if you want, but it is due to measurable, technical parameters.
(Incidentally - not sure what you mean by "full digital audio" here...."inefficient"? Surely not if its lossless or hi-res...I guess you meant "lossy"..?)...

****
By all means worship at the feet of classic analogue, but bear in mind the early stereo recordings in Boston, Chicago, Detroit and Minneapolis (RCA, Mercury etc) made by such legends as Richard Mohr and Lewis Layton, were using some of the finest equipment available, from the mikes to the open-reel recorders. (Just take a look at the photos of the control rooms). Microphones were placed with obsessive care, as their ambition was to put that orchestra there before you, in your room. They would try anything to achieve that.

Latterly, labels such as BIS and its sister label CPO have tried to be equally uncompromising with digital gear, in dynamic range, acoustic presence and so on. The recent release of the live Schubert Symphonies on CPO (Michi Gaigg/Orfeo) has its thrills & spills musically , but the sound has a spaciousness and tangibility Mohr and Layton would surely have admired.

[Bryn]

I recall when the lossless FLAC relays ran on Radio 3 from Spring 2017 to the end of the Proms, I made close comparisons on concerts from various familiar venues. Even with the usual 320 kbps AAC (technically and audibly superior to mp3) the leap in quality to lossless (which ran at various bitrates then) was obvious: the tangibility of instruments and ensembles, and the presence of a palpable acoustic they were performing in. This also meant that the climaxes in a Bruckner 4th I heard from the Barbican had a much freer, "soaring" quality, as they had more acoustic space to expand within. How sad they never brought it back...

Call this a subjective response if you want, but it is due to measurable, technical parameters.
(Incidentally - not sure what you mean by "full digital audio" here...."inefficient"? Surely not if its lossless or hi-res...I guess you meant "lossy"..?)...

I'm not sure what you intended by this. FLAC always results in varying data rate. It depends on the audio data it has to work on. Quiet ambient requires very little data, for instance.

[jayne lee wilson]

I'm not sure what you intended by this. FLAC always results in varying data rate. It depends on the audio data it has to work on. Quiet ambient requires very little data, for instance.

I've measured more consistently higher rates off of downloads (when I used to buy them) than than the 400-500 kbps that those FLAC webcasts often ran on, but I could still easily hear the upgraded SQ on R3 (even in quieter ambient moments...). Thanks for your correction, I should have been clearer, sorry.

[Bryn]

I've measured more consistently higher rates off of downloads (when I used to buy them) than than the 400-500 kbps that those FLAC webcasts often ran on, but I could still easily hear the upgraded SQ on R3 (even in quieter ambient moments...). Thanks for your correction, I should have been clearer, sorry.

Somewhat related, the 320kpbs ascribed to the BBC Sounds data rate for its AAC-LC stream is the average data rate, not a fixed one. It appears to vary between circa 200 and 400kbps (using Tools/Codec Information/Statistics in VLC Player).

[cloughie]

Of course it isn't surprising that recordings from the late 1950s still hold their own among more recent work - the technology of microphones and loudspeakers has hardly changed since that time in comparison with the developments they underwent in the first half of the last century, tape recorders had reached a high degree of sophistication and stereo recording had been pretty well worked out. The only significant improvement in recording technology since then has been digital recording of course. But when you compare developments over the 60 years before 1960 and over the 60 years since then, it seems that once the technology reached a certain threshold level many people don't find any additional improvements making such a huge difference.

RB - I fully agree with all that!

[Dave2002]

I have encoded a string quartet recording with mp3 at various bit rates, and was surprised at how low a rate I found acceptable. Owners of other ears might well get different results, of course.

I can identify with that.
Can't remember all the details but I have had perfectly acceptable (to my ears) mp3s for some instrumental combinations at low bit rate - but anything complex needed at least 225 or thereabouts.

[rauschwerk]

(Incidentally - not sure what you mean by "full digital audio" here...."inefficient"? Surely not if its lossless or hi-res...I guess you meant "lossy"..?)...

What I meant was that an more efficient coder results in fewer bits for a given audio signal (don't know if that's the generally used terminology). By 'full digital audio' I meant PCM with 44.1kHz sampling rate and 16 bits/sample. This was used for CDs because PCM was a well established technology and hardware for the coding and decoding existed in the 1970s. I define flac as more 'efficient' than unadorned PCM because it uses a lower bitrate.

I think that older recordings were simply less consistent in quality than modern ones. Last night I listened to Scriabin's Poem of Ecstasy (Houston SO/Stokowski) recorded on 35mm film by Everest in the late 50s and transferred to CD under the supervision of Seymour Solomon. It was listenable, but nobody could possibly have mistaken it for a modern recording.