Crossover Faults: Harshness & Wondering Image?

[ClassicalMusicGuy]

Hi ASR,

I have been a lurker for quite a while, and this is my first post. I'm an OG audiophile with roots in the 70s and a classical music lover. I have been auditioning new loudspeakers, but there are some audible issues that I repeatedly encounter.

My standard audition music is a classic recording of a violin concerto. I like this particular recording because it uses straightforward miking techniques, has a gorgeous violin sound over most of the instrument's range, and presents a formidable challenge to most speakers in creating a credible orchestral sound. Unlike a typical demo recording, this one only sounds good with really good speakers.

Typical audible problems I have encountered in listening to speakers that get good reviews and are not cheap:

1. Specific violin notes are reproduced harshly. Not a range of notes - just specific ones.
2. The violinist appears to be doing a jig while playing or the stereo image falls apart and smears across the two speakers on certain violin notes.
3. The violinist appears to be doing knee bends because the various notes radiate from different drivers in the loudspeaker cabinet. Upon playing a scale, it is apparent that I am hearing the physical transition between the drivers as the frequency increases. Even if I was sitting many feet farther away, the difference in the timbre of the violin through each driver is apparent.

I reckon that these are symptoms of crossover and dispersion issues among the drivers. I am curious if y'all have had similar experiences and perhaps have been able to experiment and validate the causes.

To a man, every salesman has gotten upset when I point out these deficiencies in their speakers. Nobody has said something like, "You must be a very experienced listener. You need to try our Acme Model 10000 - you will love it." Additionally, I do not recall a reviewer (professional or otherwise) who has commented about similar phenomena in their review.

For me, these flaws make a loudspeaker unsuitable for enjoying the kind of music that I listen to. Is it really too much to expect speakers to convey the performers standing or sitting in a space? Am I the weirdo for hearing these types of defects?

-Brett

 

[dfuller]

Well, it could be any of those things - or it could be the room.

Can you tell us what speakers they are?

 

[LTig]

I also think it could be the room. A very good coax speaker like the Genelec One serious could be a solution to the OPs problem.

 

[wwenze]

Your phenomena is easily explained with room modes.

Just sit at a different location and you will hear different notes being louder and coming from different locations. Front or back, left or right, up or down.

 

[kyuu]

The problem with auditioning speakers in showrooms is that... it's a showroom. They're almost always pretty poor rooms acoustically speaking, and they aren't your room. You also have no control over how the speakers are set up.

Instead of chasing ghosts, your best bet is to find a loudspeaker that has been tested (meaning Spinorama results here or on Erin's Audio Corner) and performs well and which matches your budget, space, and aesthetic requirements, and test them out in your room. Oh, and also get a good room correction system; all domestic rooms have modes that need to be tamed.

 

[sam_adams]

Un

Hi ASR,

Unless you are planning to expand your assertions with more specificity, this may not turn out the way you'd hope it would.

Oh,

 

[Mnyb]

I also think it could be the room. A very good coax speaker like the Genelec One serious could be a solution to the OPs problem.

Or a KEF model coax has precision imaging if the room and recording permits . We most not forget the recording.

Your hifi tries it best to reproduce the recording not the performance or event ( that’s not even a possibility )

 

[Holmz]

Hi ASR,

I have been a lurker for quite a while, and this is my first post. I'm an OG audiophile with roots in the 70s and a classical music lover. I have been auditioning new loudspeakers, but there are some audible issues that I repeatedly encounter.

My standard audition music is a classic recording of a violin concerto. I like this particular recording because it uses straightforward miking techniques, has a gorgeous violin sound over most of the instrument's range, and presents a formidable challenge to most speakers in creating a credible orchestral sound. Unlike a typical demo recording, this one only sounds good with really good speakers.

…

-Brett

It would be good to listen to that track.
Where is it?
Can it be uploaded?

 

[kemmler3D]

Welcome to ASR.

Assuming the room isn't causing trouble, what you're describing can be caused by speakers with peaky frequency response and uneven dispersion. Some very expensive speakers do suffer from these problems, some don't.

As others have mentioned, you can expect some of this kind of variable loudness of pure tones due to modes in most rooms.

If we exclude discussion of the room, a speaker that will probably cure everything you're describing would be the KEF LS60, or maybe some others in their line. Really depends on your budget, but there are solid options (not necessarily KEF) ranging from the mid-3 to mid-5 figure range.

 

[goat76]

@ClassicalMusicGuy

Can you give some examples of recordings of the violin where this problem occurred in the loudspeaker demos you had?

I would like to listen if I can hear the same problem, and maybe we can figure out if the problem already occurs in the recordings. I suspect it’s a big challenge to make a stereo recording using multiple microphones of a violin, as the violinist tend to move the instrument a lot while playing it, making individual tones “jump” between the speakers.

 

[dfuller]

@ClassicalMusicGuy

Can you give some examples of recordings of the violin where this problem occurred in the loudspeaker demos you had?

I would like to listen if I can hear the same problem, and maybe we can figure out if the problem already occurs in the recordings. I suspect it’s a big challenge to make a stereo recording using multiple microphones of a violin, as the violinist tend to move the instrument a lot while playing it, making individual tones “jump” between the speakers.

Indeed, if I'm close miking a violin it's one mic, then others considerably further away to capture ambience (i.e., mostly reflected sounds).

 

[Holmz]

Hi ASR,

I have been a lurker for quite a while, and this is my first post. I'm an OG audiophile with roots in the 70s and a classical music lover. I have been auditioning new loudspeakers, but there are some audible issues that I repeatedly encounter.

My standard audition music is a classic recording of a violin concerto. I like this particular recording because it uses straightforward miking techniques, has a gorgeous violin sound over most of the instrument's range, and presents a formidable challenge to most speakers in creating a credible orchestral sound. Unlike a typical demo recording, this one only sounds good with really good speakers.

Typical audible problems I have encountered in listening to speakers that get good reviews and are not cheap:

1. Specific violin notes are reproduced harshly. Not a range of notes - just specific ones.
…

Which notes?
Are the specific freqs? And what are the crossover frequencies?

…
2. The violinist appears to be doing a jig while playing or the stereo image falls apart and smears across the two speakers on certain violin notes.
…

Is the violinist moving when the recording was made?

…
3. The violinist appears to be doing knee bends because the various notes radiate from different drivers in the loudspeaker cabinet. Upon playing a scale, it is apparent that I am hearing the physical transition between the drivers as the frequency increases. Even if I was sitting many feet farther away, the difference in the timbre of the violin through each driver is apparent.
…

Again the recording and the specific notes would help.
But it could also be the box is being excited.

Single notes could potentially highlight phase problems easier than many instruments and note all being played at once.
I would think that a piano would also sound bad on the bad speakers???

 

[Redacted]

Too many variables at play to assert much....

 

[Keith_W]

Sometimes the fault is in the recording. This fault in particular:

2. The violinist appears to be doing a jig while playing or the stereo image falls apart and smears across the two speakers on certain violin notes.

... is known as "phantom centre image drift". It is caused by phase issues between two microphones or two speakers. It is rare for two identical speakers to have different phase distortion in each unless there was a manufacturing defect, so the fault is usually in the recording. If the musician moves whilst recording music, it alters the phase relationship between the two microphones and can create this effect.

You can find something similar in this recording of Schubert lieder with Ian Bostridge and Julius Drake:

Listen to how Bostridge seems to be in front of you, and when he sings certain notes he suddenly seems to jump to the left or right. Try track 11, "Wilkommen und Abschied".

If you could let me know what recording you are using, I could listen to it on my system and let you know what I think of it. Please specify which track and preferably a time stamp.

And welcome to ASR by the way!

 

[ClassicalMusicGuy]

It would be good to listen to that track.
Where is it?
Can it be uploaded?

We used the third movement from the 1967 Deutsche Grammophon recording of Beethoven's Violin Concerto featuring French violinist Christian Ferras with Herbert von Karajan conducting the Berlin Philharmonic. This was recorded by Günter Hermanns at Jesus-Christus-Kirche, which has a 2.5 second echo, direct to four-track using the following assignments:

• Track 1: Left orchestral image (Main left pair + left section spots)
• Track 2: Right orchestral image (Main right pair + right section spots)
• Track 3: Center/Woodwind blend
• Track 4: Christian Ferras's dedicated solo violin spot

No, this is not the ultimate audiophile orchestral recording, but it is a classic performance recorded by a famous recording engineer. I first heard it from an original LP pressing on a carefully setup six-digit price-tag system in San Diego - which sounded amazing and totally sucked me into the music. Since then, we discovered that it was useful for quickly separating the goats from the sheep, the men from the boys, or whatever your favorite metaphor is. Compared to more modern recordings of the same piece, Mr. Ferras maintains a beautiful dark tone throughout - even during the virtuoso "show-off" sections. Modern violinists tend to play "brighter," which is less useful for speaker-testing, since I want to verify that the speaker doesn't get shrill when challenged with complex violin playing. Also, the solo violin is mixed in from a single spot microphone, so there should not be any significant movement of the violin's perceived image location. I have found a couple of spots where I suspect there was a tape splice or damage to the tape such that it loses full contact with the head and the image wonders very briefly, but I am not holding loudspeakers to account for recording defects.

I believe that this digitally-remastered recording is available from all of the major music streaming services, although it can be difficult to find in a search due to the large number of competing versions of this concerto that are available. I recommend the Ferras/von Karajan recording series for those who are fans of this type of music.

 

[Everett T]

Hi ASR,

I have been a lurker for quite a while, and this is my first post. I'm an OG audiophile with roots in the 70s and a classical music lover. I have been auditioning new loudspeakers, but there are some audible issues that I repeatedly encounter.

My standard audition music is a classic recording of a violin concerto. I like this particular recording because it uses straightforward miking techniques, has a gorgeous violin sound over most of the instrument's range, and presents a formidable challenge to most speakers in creating a credible orchestral sound. Unlike a typical demo recording, this one only sounds good with really good speakers.

Typical audible problems I have encountered in listening to speakers that get good reviews and are not cheap:

1. Specific violin notes are reproduced harshly. Not a range of notes - just specific ones.
2. The violinist appears to be doing a jig while playing or the stereo image falls apart and smears across the two speakers on certain violin notes.
3. The violinist appears to be doing knee bends because the various notes radiate from different drivers in the loudspeaker cabinet. Upon playing a scale, it is apparent that I am hearing the physical transition between the drivers as the frequency increases. Even if I was sitting many feet farther away, the difference in the timbre of the violin through each driver is apparent.

I reckon that these are symptoms of crossover and dispersion issues among the drivers. I am curious if y'all have had similar experiences and perhaps have been able to experiment and validate the causes.

To a man, every salesman has gotten upset when I point out these deficiencies in their speakers. Nobody has said something like, "You must be a very experienced listener. You need to try our Acme Model 10000 - you will love it." Additionally, I do not recall a reviewer (professional or otherwise) who has commented about similar phenomena in their review.

For me, these flaws make a loudspeaker unsuitable for enjoying the kind of music that I listen to. Is it really too much to expect speakers to convey the performers standing or sitting in a space? Am I the weirdo for hearing these types of defects?

-Brett

What speakers where you previously using where this issue was absent from?

 

[ClassicalMusicGuy]

Welcome to ASR.

Assuming the room isn't causing trouble, what you're describing can be caused by speakers with peaky frequency response and uneven dispersion. Some very expensive speakers do suffer from these problems, some don't.

As others have mentioned, you can expect some of this kind of variable loudness of pure tones due to modes in most rooms.

If we exclude discussion of the room, a speaker that will probably cure everything you're describing would be the KEF LS60, or maybe some others in their line. Really depends on your budget, but there are solid options (not necessarily KEF) ranging from the mid-3 to mid-5 figure range.

The worst "offenders" in violin listening have been:

• Harbeth 30.2: Smeared stereo image of the violin on specific notes. During the crescendos, the image collapsed and the overall tonality was harsh
• Sonus Faber (various mid-range models at multiple dealers): Overall bright tonality with wondering imaging of the violin. With a three-way floor stander model, it was possible to hear the violin sound transition between the tweeter and midrange - thus changing its tonality. Unbearably harsh at certain high-frequency notes.
• Falcon MD10 & MD20: Harshness of certain violin notes was surprising as well as the unstable stereo image
• Revival Atalante 3: Overall bright tonality with harsh high-frequency notes plus unstable violin image
• Wharfedale Super Linton: Could easily hear the violin part transition between the tweeter and midrange. Unstable violin image.
• Dynascan: Various mid-line bookshelf and floor standers. Unstable, unconvincing stereo imaging.

Speakers with a stable image on violin solos:

• Estelon: Unsure of which model, but a very expensive tall three way design.
• KEF Reference 1: Convincing orchestral sound and stable imaging
• Dali Rubicon 2 & 8: Stable image and clean high frequency characteristics, but unrealistic "bright" tonality
• Falcon LS3/5a Gold Badge: Stable image and excellent spacious transparency, but with a colored tonality and not enough bass to reproduce tympani's and cellos

Inconclusive from my listening:

• Fyne (high-end models): Very different results depending on particular audition situation. Imaging is present but it is different than most other speakers; I don't know to describe it. Although I wouldn't say that there was a wondering image illusion, there were times when the violin would seem to be located exactly at the throat of one of the loudspeaker's coaxial drivers rather than floating as a phantom image as it typically would.
• KEF R3 Meta: Overall harshness and splattering of image across the soundstage on some high notes

I was especially surprised with the KEF R3, and so we switched back and forth several times between a Reference 1 pair sitting next to them. On any classical music containing significant high-frequency violin energy, the the R3s were harsh and lost clarity. On pop/rock tracks, we were unable to identify a similar phenomena, although the imaging on vocals was much better with the Reference 1.

My hypothesis is that the deficiencies are caused primarily by driver resonances and mid/tweeter dispersion issues over the crossover region, typically 2.5 -3.25 kHz. Violins and operatic sopranos are perhaps the most common emitters of fundamental frequencies in this region - the use of the highest octaves of a piano/keyboard are relatively rare and harmonics from non-operatic singers have a much lower relative amplitude than the fundamentals from an expertly played violin or a diva.

In other words, using violin music (or sopranos) for testing loudspeakers is probably a special case that is particularly challenging and relatively uncommon by most customers. Therefore, my derivative hypothesis is that most loudspeaker manufacturers do not optimize their design for the upper ranges of violins but rather focus on typical male & female vocals, the guitar frequency spectrum, bass impact, etc. I suppose that I can't blame them for making a design compromise that impacts a minority of likely customers.

Thoughts?

-Brett

 

[Redacted]

So all those speakers were tried as A-B tests in the same room? You must own a hifi shop!

 

[bmc0]

typically 2.5 -3.25 kHz. Violins and operatic sopranos are perhaps the most common emitters of fundamental frequencies in this region

The very top of the soprano range is about an octave below that. Squillo (singer's formant; prominent in both male and female operatic voices) is in that range, though.

 

[Everett T]

To the OP, have you tried any large format two ways like the JBL 4367 where the crossover is below 1000hz? I personally don't think there is an issue with the crossover region in general but the audition room for some of the speakers you tested. I'm still curious what speakers you're coming from that played you test track correctly?

Dennis Murphy plays in the DC Orchestra and his voicing and choice of drivers would work for you I'm guessing. His company is Philharmonic Audio