Steinberg SpectraLayers: Repairing "Congested" Recordings?

[chane]

Some of my favorite stereo recordings from the 60s-mostly recorded in the best Hollywood studios and issued by major labels-sound congested where instruments and vocals sound smeared and confused. This typically happens when the vocals and/or orchestrations get busier in the track. I don’t know if this was due to poor mic placement and/or improper baffling techniques, but it’s hard to imagine engineers like Henry Lewy or Larry Levine making such errors. https://www.discogs.com/artist/259822-Henry-Lewy

https://www.discogs.com/artist/276888-Larry-Levine 1

My Verve edition of the CD album. https://www.discogs.com/master/11184...MTg0ODM3Nzk%3D 1

As for what Spectralayers 10 Pro (version 11 said to be out this year) and my skill level can do for troubled recordings like Look of Love:

1.) If you can hear the congested/smeared sound during the louder, busier passages of Look of Love, might there be any way estimate how many different instruments I’d have to unpack from that mix before I could identify where that disturbance (s) is coming from? Most likely, the less instruments I’d have to (attempt to as a newbie) extract from the mix to find and fix the problem the better chance for best sounding results.

2.) However, what I fear might be the biggest “baked in” problem with Look of Love is that the amount of compression originally applied during mastering to fold the acoustic energy of those instruments into a vintage two track analog recording (for 1/4" analog stereo 4-track 7.5 ips open reel tape and vinyl media) is what may be causing the congestion. And if that’s largely the case then how much-if at all-can SpectraLayers *decompress* two or three stems and/or the entire stereo track? Indeed, IF SpectraLayers can actually do some amount of decompression might it fix much of the audible problem?

[RichardT]

I don’t think Spectralayers is really going to help with this (though I haven’t used it much myself, to be fair). Edit: just checked and the latest version does seem to do stem separation, so I'm wrong!

Recordings generally sound congested when there’s too much going on in each frequency band (too many instruments occupying that band, too much processing). It can also be caused by a lack of detail or instrument separation in the original recordings that hides the transients, so everything sounds like mush.

I don’t know if stem separation will work on these recordings. But it might be worth a shot.

I expect you’re already aware, but there is a copyright in the original recordings that prevents anyone creating derivative works.

[Tomás Mulcahy]

My Verve edition of the CD album. https://www.discogs.com/master/11184...MTg0ODM3Nzk%3D 1

Would be interesting to hear the issue you describe but this link is broken.

[chane]

I'd be happy to upload or email the WAV file; ditto some others.

[chane]

I don’t think Spectralayers is really going to help with this (though I haven’t used it much myself, to be fair). Edit: just checked and the latest version does seem to do stem separation, so I'm wrong!

Recordings generally sound congested when there’s too much going on in each frequency band (too many instruments occupying that band, too much processing). It can also be caused by a lack of detail or instrument separation in the original recordings that hides the transients, so everything sounds like mush.

I don’t know if stem separation will work on these recordings. But it might be worth a shot.

I expect you’re already aware, but there is a copyright in the original recordings that prevents anyone creating derivative works.

Thanks for your prompt reply. Regarding copyright concerns, I’ve always either purchased content (my CD/DVD/BD collections are very large) or borrowed them via my local public library, for which I am taxed to directly support. And pirating content to resell as restored or otherwise enhanced versions isn’t my thing.

From your description of the problem, it seems like this congested sound might have been caused by an insufficient number of individually miked and recorded channels per instrument and/or little or improper baffling to achieve sufficient isolation between those miked instruments. Yes?

If so, might the result be that such “acoustical leakage” between channels among instruments around the studio session floor had caused some degree of intermodulation distortion of instruments bandwidths?

In any case, trying SpectraIayers Pro 10 or 11 will likely be the only way of knowing how well it might be able to fix this stereo track. However, if the congested sound is due to intermodulation distortion than how could merely unmixing the instruments into stem channels eliminate the distortion once the stems are mixed back together?

Otherwise, how else might this stereo track be at least partially fixed, assuming I can get over the app’s learning curve fast enough before the trial expires?

[Philbo King]

Some of my favorite stereo recordings from the 60s-mostly recorded in the best Hollywood studios and issued by major labels-sound congested where instruments and vocals sound smeared and confused. This typically happens when the vocals and/or orchestrations get busier in the track. I don’t know if this was due to poor mic placement and/or improper baffling techniques, but it’s hard to imagine engineers like Henry Lewy or Larry Levine making such errors. https://www.discogs.com/artist/259822-Henry-Lewy

https://www.discogs.com/artist/276888-Larry-Levine 1

My Verve edition of the CD album. https://www.discogs.com/master/11184...MTg0ODM3Nzk%3D 1

As for what Spectralayers 10 Pro (version 11 said to be out this year) and my skill level can do for troubled recordings like Look of Love:

1.) If you can hear the congested/smeared sound during the louder, busier passages of Look of Love, might there be any way estimate how many different instruments I’d have to unpack from that mix before I could identify where that disturbance (s) is coming from? Most likely, the less instruments I’d have to (attempt to as a newbie) extract from the mix to find and fix the problem the better chance for best sounding results.

2.) However, what I fear might be the biggest “baked in” problem with Look of Love is that the amount of compression originally applied during mastering to fold the acoustic energy of those instruments into a vintage two track analog recording (for 1/4" analog stereo 4-track 7.5 ips open reel tape and vinyl media) is what may be causing the congestion. And if that’s largely the case then how much-if at all-can SpectraLayers *decompress* two or three stems and/or the entire stereo track? Indeed, IF SpectraLayers can actually do some amount of decompression might it fix much of the audible problem?

Phil Spectors 'Wall Of Sound' records all suffered from this. I suspect this is what happens you have a psychotic producer.

As to fixing it? Like trying to unbake a cake.

[RichardT]

Well, there are lots of possible reasons for congestion! Even more than I listed above. Excessive compression on the master bus could certainly be one cause.

I think stem separation is worth a shot, and then you might be able to clarify the stems with EQ, compression, expansion, etc.

If you post the name of the artist, album and track people here will be able to listen for themselves.

If, as Philbo says, it’s a ‘wall of sound’ production these were made to sound like that. Spector recorded multiple piano players etc so separating stems is going to be hard.

[chane]

If you review the credits for the Look Around album here there's no sign that Phil Spector had any hand in the production, though Larry Levine, who often engineered Phil's tracks, does share credit with Henry Lewy.
https://www.discogs.com/release/1051849 ... ook-Around

FWIW, Lewy would go on to engineer almost all of Joni Mitchell's albums. As for how Levine may have been obliged to mike, mix and record this album any differently than he did the earlier Spector hit tracks I wouldn't know, though i would think that playback of his final takes would have been carefully reviewed by Sergio Mendes, orchestral arranger Dave Grusin and/or A &M label owner Herb Alpert.

In any case, it certainly looks like the main culprit here was the same post-mastering session global dynamic range compression that was all too often applied to master tapes during the vinyl disc cutting phase during that pre-digital era. And as discussed here, it all looks pretty hopeless.

https://forums.stevehoffman.tv/threads/ ... s.1199083/

[RichardT]

Just had a listen to Look of Love from that album (on Tidal) - I don't hear any congestion at all! For the time it's an absolutely lovely recording.

There’s some bass roll-off, but that’s to be expected.

It might be a different master from the CD of course. Do you have access to a lossless streaming service to listen to it there?

[chane]

Just had a listen to Look of Love from that album (on Tidal) - I don't hear any congestion at all! For the time it's an absolutely lovely recording.

There’s some bass roll-off, but that’s to be expected.

It might be a different master from the CD of course. Do you have access to a lossless streaming service to listen to it there?

Thanks so much for the due diligence. But all Sergio Mendes albums throughout the 60s were on A & M records, which, like Verve, were later acquired by Universal. Unless I'm wrong, Verve was Universal's classic jazz label, so at least the digital tape to CD mastering process should be pretty much up there in quality. Tragically, there was likely only one digital copy ever made of each of the Sergio Mendes & Brasil 66 album analog tape masters before this horrible event occurred, however much accidental it truly was.
https://en.wikipedia.org/wiki/2008_Univ ... udios_fire

Therefore, would not only one and the same digital copy be available for any other kind of marketed use?

In any case, besides Tidal, do you subscribe to any other music streaming service which also has access to this track? If yes, please name them and please compare sound quality of Look of Love among them.

[RichardT]

No sorry, I don’t. You could sign up to Spotify’s free tier and listen there - it’s not lossless but it’s pretty good. You would for sure be able to distinguish if the CD was from a different master.

There would have been an analogue master back in 1968, but creating digital masters from that tape may have been done more than once, or a single digital master may have been remastered at some point in its history, after the loss of the tape, to create a new version.

If the mastering was not done sympathetically (if the mastering engineer wanted to bring things up to some contemporary notion of loudness, for example), this could affect the sound.

[James Perrett]

Having had a listen to The Look Of Love on YouTube I would say that the sound is simply of its time. This is a pop record, not a jazz record, so it would have been intended to be heard on AM transistor radios where there is no real high end so the clarity comes from the high midrange.

It may well have only used 3 tape tracks or 4 at most so plenty of processing and mixing would have happened before the signal hit the tape. It was also possibly only intended to be heard in mono. I've heard a few raw multitracks from this era and often the sound on record is very close to the sound coming from a straight "faders up" mix from the multitrack.

[chane]

Having had a listen to The Look Of Love on YouTube I would say that the sound is simply of its time. This is a pop record, not a jazz record, so it would have been intended to be heard on AM transistor radios where there is no real high end so the clarity comes from the high midrange.

It may well have only used 3 tape tracks or 4 at most so plenty of processing and mixing would have happened before the signal hit the tape. It was also possibly only intended to be heard in mono. I've heard a few raw multitracks from this era and often the sound on record is very close to the sound coming from a straight "faders up" mix from the multitrack.

It's seems quite unlikely that Look of Love and other heavily orchestrated tracks on this album were originally recorded on 3 or 4 track machines. In 1965, the Beach Boys recorded "Pet Sounds" on a Scully eight track tape machine; Motown also began using Ampex or Scully eight tracks that year. https://en.wikipedia.org/wiki/Multitrack_recording And as shown, the entire "Pet Sounds" project was recorded mostly at the same studios as was "Look Around", at least a year later. https://en.wikipedia.org/wiki/Pet_Soundshttps://en.wikipedia.org/wiki/Look_Arou ... des_album) Unfortunately, to ensure trackability on the cheap or even SOTA vinyl players of the times the analog master tapes were obviously hit with high doses of global dynamic compression. The last thing the record companies wanted was to get a lot of returns of 'unplayable' records. So we all paid the price with severely compromised sound in these releases. Not all records had this done, but there were enough of them. Even many classical releases used this heavy handed compression in the name of 'playability', especially at the inner grooves when much classical music is at it's loudest. It was also common practice for the original engineers in classical sessions to 'ride the gain' during loud passages.

These older analog compressors when applied to entire mixes have the signature sound of removing all transients and clarity in the mix - everything starts sounding 'thick' and 'muddy'. This would obviously be most used during louder portions of a song. Because the attack and release times of these compressors was quite long by today's standards, the effect was a total flattening of the sound.

It's all very tragic. If only digital audio had happened a lot sooner.

[RichardT]

I really don’t get what you’re saying here as I see no signs at all of heavy compression (by today’s standards).

There’s definitely some compression but I don’t think it’s global and I’d say it’s quite light.

I’m stepping out of this thread now as I’m finding it quite confusing and frustrating.

[Tomás Mulcahy]

It's seems quite unlikely that Look of Love and other heavily orchestrated tracks on this album were originally recorded on 3 or 4 track machines. In 1965, the Beach Boys recorded "Pet Sounds" on a Scully eight track tape machine

This is not correct. The Wrecking Crew band was recorded on 3 track at Western with Chuck Britz, that tape was then copied to one track of the 8 track and the vocals were overdubbed at Western and CBS Columbia Square.

https://en.wikipedia.org/wiki/Pet_Sound ... ck%20Britz.

I wonder if you have some confirmation bias here, when you overlooked the 3 track recording on Pet Sounds? I would suspect that the number of tracks is a red herring in the alleged problem of congestion. James hit the nail on the head- the intent of the medium was there from the very start of the production process.

Try the Pet Sounds stereo remix, that sounds lush. On the deluxe version you can hear the band recording sessions- the three track was no limitation at all. The players and the room are beautifully captured.

[tea for two]

Weeeell my hearing these days tops out at approx 12.57kHz, so any tunes I do my choice of instruments where I play them is always below.
I just had a listen to Look of Love, their choice of instruments sounds where they played them as well as the Vocalists they kept their top end to the upper mids to my ears : I haven't analysed this in software so I could well be wrong.
So I'd hazard a guess by doing so they didn't require any heavy handed global processing. So Look of Love sounds clear enuff to my ears, not muddy.

For me this is always the first step to getting a clear mix for my stuff : choice of instruments sounds, which octaves they are played at : so that they complement have clarity from the outset : don't have to do jiggery pokery later on or at least minimal.

[Hugh Robjohns]

Some of my favorite stereo recordings from the 60s-mostly recorded in the best Hollywood studios and issued by major labels-sound congested where instruments and vocals sound smeared and confused....it’s hard to imagine engineers like Henry Lewy or Larry Levine making such errors.

I wouldn't categorise this characteristic as an error. I would suggest it is an artefact of the technologies and the sonic fashions of that specific era.

Those engineers and their peers elsewhere all worked hard to achieve the best sound they could given the facilities available to them, and the expectations of their record label clients and record-buying public. Many of them did a phenomenal job, with releases which set musical quality benchmarks that still hold up well today. They have a common sound character which defines the sound of the 60s... just as the music from other eras have distinct sound characters.

Modern equipment and release formats allow far greater fidelity and dynamic range than was available on the 1960s, but it would be very unfair to judge those vintage releases on the basis of current capabilities.

[chane]

I really don’t get what you’re saying here as I see no signs at all of heavy compression (by today’s standards).

There’s definitely some compression but I don’t think it’s global and I’d say it’s quite light.

I’m stepping out of this thread now as I’m finding it quite confusing and frustrating.

I'm sorry about this and please know I am in no way second guessing your evaluation of this recording. In fact, I'm heartened by your positive comments about the sound quality of Look of Love. No, I don't have Tidal; at present I only have VLC player, an older Benchmark DAC and an old pair of Sennheiser 650 headphones. Happily, my total hardware budget is large so no issues finishing the speakers which Pierre and I have designed for Troy Crowe to build and for biamping the system later on.

But once those speakers are here I'll be done with headphones. And if you've ever heard any sizable number of 1960s pop or even many classical and soundtrack recordings over horn type speakers like Pierre's and those I'm building you would immediately understand how the same recordings can sound so different compared to conventional speakers. There is a very wide difference in the sound quality of many recordings which are often highly accentuated by the resolving ability of horns-even more so than by electrostatic speakers, as horn speakers are far more efficient and can easily produce wider dynamic range and yet with very low distortion. These differences tend to be smoothed over and homogenized by your typical direct radiator speaker system which makes most recordings sound at least 'nice'.

It was for the cheaper variety of those cone type speakers and mono systems-plus the further limitations of AM radio-that recording engineers were obliged to mix and master for. Sadly, that's still probably the case to a large degree, what with the ongoing Loudness wars which producers and/or artists continue to insist on waging.

But I certainly will listen to Look of Love on Tidal or via another high res service, and more than once throughout the coming months, as big and nice changes happen with my hardware.

[chane]

It's seems quite unlikely that Look of Love and other heavily orchestrated tracks on this album were originally recorded on 3 or 4 track machines. In 1965, the Beach Boys recorded "Pet Sounds" on a Scully eight track tape machine

This is not correct. The Wrecking Crew band was recorded on 3 track at Western with Chuck Britz, that tape was then copied to one track of the 8 track and the vocals were overdubbed at Western and CBS Columbia Square.

https://en.wikipedia.org/wiki/Pet_Sound ... ck%20Britz.

I wonder if you have some confirmation bias here, when you overlooked the 3 track recording on Pet Sounds? I would suspect that the number of tracks is a red herring in the alleged problem of congestion. James hit the nail on the head- the intent of the medium was there from the very start of the production process.

Try the Pet Sounds stereo remix, that sounds lush. On the deluxe version you can hear the band recording sessions- the three track was no limitation at all. The players and the room are beautifully captured.

I'll concede if accurately corrected, though according to the Wiki Brian Wilson and/or Chuck Britz did make use of "eight" track machines during those sessions. https://en.wikipedia.org/wiki/Multitrac ... ng#Process

[sonics]

This topic triggered my attention. I'm always interested in hearing examples of older engineering done right, and when I read the names Henry Lewy and Larry Levine I was intrigued.

This music was new to me, in the sense that I'd never listened to it as the sole focus of my attention before. I missed out on those LA dinner parties in the 1960s where it might have been played!

I listened to a few streams and identified three different versions, and a couple of level-adjusted variations. I can't be sure if that was playback or mastering.
The oldest release sounded like a genuine original transfer from tape. Another has been processed for noise removal, and a third had been remastered badly and has clear phase issues.

I can hear the use of compression in the recording process, and probably some in the mastering stage, but I don't think the best of those recordings has been ruined. I don't think that those versions would gain anything by radical remastering. I'd do it slightly differently, certainly.

Do you really think that the engineering you've heard (for vinyl release, obviously) is that bad?
Have you compared other versions to make sure you're listening to your preferred release?

(Side note: I was surprised to hear the doubled word-endings on the vocal performances. Messy!)

Weeeell my hearing these days tops out at approx 12.57kHz,

Oh, TfT, you do make me larff...

(My italics.)

[Tim Gillett]

These days we can take it for granted that we can make an essentially identical (digital) copy of a recording. Not so pre digital. "Generational loss" was a given. The mixdown copy of the session tape was by definition poorer quality, and any copy of that copy worse still, and so on. So just the act of post session mixing degraded the sound.

"Duplication" of recordings and maintaining sound quality in duplicates used to be a far bigger deal than it is these days. To preserve the session tape for posterity, that session tape was used as little as possible, meaning most of the wear and tear occurred intentionally on the copies.

The holy grail for remastering engineers has always been finding intact original session discs or tapes to work from, or the next generation mixdown tape.

The "congestion" you heard on certain reissues of version may be the increased distortion, baked into the best copy they had access to. The Pet Sounds CD reissue from Mark Linnet is a good example because we can hear the distortion in the later copy compared to the much cleaner sound from the original from which the analog copy was made.

[chane]

This topic triggered my attention. I'm always interested in hearing examples of older engineering done right, and when I read the names Henry Lewy and Larry Levine I was intrigued.

This music was new to me, in the sense that I'd never listened to it as the sole focus of my attention before. I missed out on those LA dinner parties in the 1960s where it might have been played!

I listened to a few streams and identified three different versions, and a couple of level-adjusted variations. I can't be sure if that was playback or mastering.

The oldest release sounded like a genuine original transfer from tape. Another has been processed for noise removal, and a third had been remastered badly and has clear phase issues.

I can hear the use of compression in the recording process, and probably some in the mastering stage, but I don't think the best of those recordings has been ruined. I don't think that those versions would gain anything by radical remastering. I'd do it slightly differently, certainly.

Do you really think that the engineering you've heard (for vinyl release, obviously) is that bad? Have you compared other versions to make sure you're listening to your preferred release?

(Side note: I was surprised to hear the doubled word-endings on the vocal performances. Messy!)

I’ve been a huge fan of this very particular flavor of bossa nova/jazz music since my teens. It happened here on Long Island at 5 Marsak Lane. I was babysitting for this couple and they had a Fisher stereo with one of the better tube receiver models and floor standing speakers. While browsing their albums, I found these two. https://en.wikipedia.org/wiki/Herb_Alpe ... asil_%2766
https://en.wikipedia.org/wiki/Equinox_( ... des_album)

Luckily, the two young kids remained asleep upstairs and I had time to play at least half the tracks on one or both albums. I was absolutely captivated, and not only by the musical style-and especially the tight and delightful arrangements-but also by the apparent sound quality. I certainly had no gear sounding that good back then and the mix and/or apparently low (clipping) distortion and apparent sound stage (and air?) of several, if not most of these tracks were far better than any of the pop and rock music I’d heard, though that would be hardly surprising.

But the unmistakable and crucial difference between those albums and those to come, beginning with https://en.wikipedia.org/wiki/Look_Arou ... des_album) , was that merely by partnering with arranger Dave Grusin (who at the time did TV themes like this https://www.youtube.com/watch?v=XlyU1h0QL8U ) , the sound scape size of Brasill 66 tracks had to jump from a small jazz/vocal combo to a full orchestra, and the consequent need to mash down those far great sound pressure levels to make them “vinyl” friendly on most turntables and AM radio.

No, as I mentioned to RichardT, I haven’t compared different releases. AFAIK, there are no currently pressed vinyl versions of any of these albums, and I completely switched from vinyl to CD long ago. I do plan to check out “Look of Love” on Tidal, but it’s pointless to do so until I finish upgrading my hardware within the coming months. Subwoofers and amplification are ready; need to choose a DAC from my short list. What remains is choosing among three compression drivers, and how Troy and Pierre choose to filter that driver’s HF response and apply some degree of voicing to the passive crossover.

Optimistically, what I learned here today from Ampexed is that biamping these speakers later and adding an adjustable three band passive filter will offer additional voicing may likely help compensate when playing over compressed recordings. Though I’d also want to hear Pierre and Troy’s thoughts on this.
https://forums.stevehoffman.tv/threads/ ... 083/page-2

[chane]

These days we can take it for granted that we can make an essentially identical (digital) copy of a recording. Not so pre digital. "Generational loss" was a given. The mixdown copy of the session tape was by definition poorer quality, and any copy of that copy worse still, and so on. So just the act of post session mixing degraded the sound.

"Duplication" of recordings and maintaining sound quality in duplicates used to be a far bigger deal than it is these days. To preserve the session tape for posterity, that session tape was used as little as possible, meaning most of the wear and tear occurred intentionally on the copies.

The holy grail for remastering engineers has always been finding intact original session discs or tapes to work from, or the next generation mixdown tape.

The "congestion" you heard on certain reissues of version may be the increased distortion, baked into the best copy they had access to. The Pet Sounds CD reissue from Mark Linnet is a good example because we can hear the distortion in the later copy compared to the much cleaner sound from the original from which the analog copy was made.

When I built my HTPC two years ago, I went with a Rocket Lake Xeon processor and compatible motherboard which supports ECC memory, which does help to reduce the chances of bit loss when saving, copying and downloading files. The NAS that I hope to finally get around to building late this year will use an error detecting/repairing/preventing BT or Z file system to manage the multiple storage drives. Bernie Grundman was right; digital recordings can and do degrade over time via bit loss, though probably less often and more slowly than analog recordings.

You make a very good point where the audible "congestion" heard from the CD (or earlier vinyl) album of "Look Around" may be the added distortion from tape machine and/or active mixing board preamps when the master tape was copied to a "working" or "safety" master used to during vinyl and pre-recorded tape production. Furthermore, and any such distortion would be even more evident (and actually additive) with the "Look of Love" track than some of the other less elaborate album tracks because the Grusin orchestral arrangements almost certainly required more tape and mike preamp channels.

When I do finish most of my hardware upgrades I wonder how this will sound compared to my 1990 mono and 2001 stereo CD editions.
https://www.hdtracks.com/#/album/5df142 ... c09bc163fd

[tea for two]

I’ve been a huge fan of this very particular flavor of bossa nova/jazz music since my teens. https://en.wikipedia.org/wiki/Herb_Alpe ... asil_%2766
https://en.wikipedia.org/wiki/Equinox_( ... des_album)

I was absolutely captivated, and not only by the musical style-and especially the tight and delightful arrangements-but also by the apparent sound quality.

But the unmistakable and crucial difference between those albums and those to come...was that merely by partnering with arranger Dave Grusin...the sound scape size of Brasill 66 tracks had to jump from a small jazz/vocal combo to a full orchestra, and the consequent need to mash down those far great sound pressure levels to make them “vinyl” friendly on most turntables and AM radio.

I adore this famouse Brasil 66 tune just as is : Vocals Piano UprightBass Drums Percussion.
https://m.youtube.com/watch?v=BrZBiqK0p ... JhZGlsIDY2

How to mess up a brilliant tune composed for other instruments : orchestrate it.
I can't tell you the amount of thymes I have thawd orchestra messed this up, heavy handed, unnecessary.
There are untold songs tunes for which using an orchestra can show poor judgement
poor taste : maketh not a modern composer composotion sophisticated.

For clarity or lack of in orchestral recordings pressed onto vinyl here's some Mercury Living Presence 1949-1960s recordings of Classical tunes :
which highlights clarity varies.
https://m.youtube.com/playlist?list=PL2 ... zxS_nG8sEn

[Hugh Robjohns]

...horn speakers are far more efficient and can easily produce wider dynamic range and yet with very low distortion.

Hmmm. I'll go with 'far more efficient' ... but that's hardly relevant in a domestic hi-fi setting. In comparison to a good modern monitor speaker the 'wider dynamic range' claim just isn't true, and compression drivers feeding horns are notorious for high levels of harmonic and intermodulation distortion — although the best modern designs can be quite good these days. I did some work with Meyer a while back developing a bespoke compression driver for acstudio monitor which delivered an amazing performance, but it was both physically large and very expensive.

To be fair, I have heard some good-sounding horn-based hi-fi speakers and studio monitors, but I've heard many more poor ones...

Hi-fi fashions come and go, and horn-based monitors are at least consistent with 1960s productions and low-powered valve amps.