Dunno, I'd be surprised if they don't just sum them. The 2.1 active crossovers I've seen sum after the filter.
Apparently some sum before
Last edited:
In a real world with real instruments, overtones and transients enable localization of instruments with deep bass tones. Two weeks ago I was in one of the baroque palaces in Prague and as a part of program there was a concert of a chamber orchestra in its Main Hall. There was one double bass located just at the right corner of the Hall (hall photo attached) and I concentrated especially on the sound of the double bass. It was possible to determine clearly the place from which the double bass sound was coming, it was exactly the right corner, not the middle, not the left, not the uncertain place. I assume it is because of the overtones, higher harmonics and transients associated with the bass melody.
Attachments
.Same applies for live rock/blues/folk mixing.....mic positioning and channel strip panning/sweepable eq (nowadays channel strip delay is available too) are mission critical to getting a live mix that 'sits right'.....at centre line out from the stage (FOH mixing position typically) and importantly (even moreso !) out in every corner of the venue too where 'stereo' information is 'lost/diminished' due to listener positionings/room reverbs/room+ crowd dampings etc.
This is where the value/importance of the sound engineers 'sound check walk' comes in......front of stage/dance floor should sound good and so also should be is every other standing, leaning or sitting spot in the venue (at the bar and behind the bar is uber important too !).
Dan.
Sum to mono centre active sub works well to bottom end augment reasonably full range L&R speakers.....the overtones/delays etc conveying bass positional information is still present and all three speakers work together to provide solid low bottom end fundamentals which don't convey real directional information anyway in the home environment.
The downside is that sub cutoff frequency and level requirement/adjustment can vary from track to track, most certainly from album to album/genre and require tweaking according to the recording mastering and according to 'taste'.
The marketing consensus for LR/Center/Surround/Sub AV systems is that the sub positioning is not important because of supposed lack of LF localisation cues.
In practice with active sub placed between LR FR main speakers, I find that sub LR and especially depth positioning is disarmingly critical despite the typical 'advice'.
For rock/funk/blues/disco etc I find that a well done sub really brings life/fun to the party....typical LR speakers may provide measured flat response but lack in room power response in the lows/bottoms that only center speaker augmentation can provide.
The takeaway is that the 'phantom' lows center does not quite 'cut it' for typical loudspeakers due to various reasons, and correctly done centre channel summed mono loudspeaker reinforcement provides elegant solution despite some opinions already expressed.
Dan.
Last edited:
I agree, that's why I always try to avoid sitting on the far right hand side of the hall, where the double bass tend to overpower everything else.
With due respect, I'd replace your 1st "real" with "classical", and the 2nd with "acoustic". Rock is also real, has been so for way over half a century.
I'll elaborate a little. This is a vinyl rip not using a computer but a 24bit field recorder at 48K sampling. I had been comparing RIAA pre-amps and I chose this LP because the recording is highly regarded and particularly the first track has three distinct sections of increasing dynamic contrast. Even though it is vinyl once recorded it becomes the reference data.
The standard multi-tone is one 65536 time record repeated in a loop and there were questions about the noise floor modulation with real music. Real music fills all FFT bins so added noise is always masked from direct measurement. This is purely a research project to see if anything shows up, the first stage was to see what this extreme comb filter does to the music and see what the floor created by the empty bins looks like. One could conceivably take only the empty bins and an inverse transform and listen to it.
What I posted was the comb filtered file showing the numerical noise of double precision math (~-300dB or so) to show that the noise floor survives back and forth transfer from the time to frequency domain. The second was just (file + 0.0001*file**2) to give ~-100dB seconds to see if a simple non-linear transfer function made a filled in noise floor at an expected level.
Under Windows I used 15 digit FP text files and no music software just to eliminate confounders like SRC, unknown dither, etc. This is an issue with passing files around for folks to play with. This process will not survive format or length conversions well. I could create a data set that would work with a large power of two FFT (like in the LA RIAA article) but I think 1-2M points is the least usable length to capture an appropriate "real" musical event.
Last edited:
Thank you for the link to the article on '101 room acoustics' - most useful, and part of my ongoing studies of audio. The author is American, and is from a big country that has big rooms. I live in a rural part of small country that has small rooms. You PMA, live in a city that has palaces. Let me think out loud here....
I am mentally preparing for my next speaker build - another LeCleachhorn. Only this time with a 180 degree backward curl to the lip of the mouth - maybe (or maybe not!) with an oval shape, as I want to use a single point stereo ceiling speaker as the sound source. My thinking is to do away altogether with the compression chamber, and instead use the horn as a wave guide voiced to focus the soundstage as a true stereophonic bubble of sound within a room.
Not something heavy and massive like what Volvoetre or Romy the Cat or Nelson Pass would make, but something light and portable, and relatively independent of room acoustics.
Other (considered) readings are making me more and more convinced that true monophonic recordings actually contain stereophonic information that otherwise cannot be, or shall we say cannot easily be detected or taken advantage of.
Obviously, I cannot prove this, but I do believe that the cues for stereophonic spatialised information are within the tiniest of sounds masked by the overall signal to noise ratio within monophonic recordings. I think equalising a single point stereo speaker fitted with a double coil mono woofer and twin tweeters inside a horn would make for a very interesting experiment.
Between the world wars of the last century there was a horn speaker built as an international reference for audio sound quality at the Science Museum in London. It was an exponential horn 27ft in length with a 7ft wide mouth, powered by a Western Electric driver. They used it to play live radio transmissions from the BBC Crystal Palace radio orchestra, and have lunchtime concerts in the museum cafeteria. You could even buy tickets. It was very popular, and those who sat and listened would afterwards say that the illusion of the orchestra being in front of them was overwhelmingly convincing - and unforgettable.
My point is, if this what they say they were hearing, where was it what they were hearing? I think it was all inside their heads. That the horn by its sheer size was efficiently capable enough at presenting this low level spatial information, thereby making it real enough so as to become unforgettable. In other words, real memories can only be made from real experiences.
So, it seems it's been established/accepted that NFM is a likely result of non-linear processing & all the devices in our replay chain are non-linear.
The audibility side of NFM needs to be analyzed - the standard answer to this is that masking will render NFM inaudible.
It has been suggested that only by gathering NFM measurements & analyzing them can we begin to correlate what characteristics of NFM might correlate to audibility - NFM spectral makeup, pattern of modulation, amplitude of modulation, etc. which makes for a very complex problem
The aspect that multitone testing using multitone test signals or Scott's great measurement (well done Scott in following this up) don't seem to capture is the modulation nature of NFM. A multitone measurement is a snapshot of multiple IMD products from a fixed test signal. Scott's measurement is a fixed distortion level injected into a modified music signal.
The best collection of multitone measurements I found is from a measurement device & software Reference Audio Analyzer which has a database of m-tone measurement reports for various DACs & other devices (a lot have 10 tone & 50 tone measurements).
This forum page seems to have lost the links back to the m-tone measurements devices that fall into each category that they have identified from this database:
- Uniform horizontal distribution of harmonics,
- Distribution at an angle in the direction of increasing the high-frequency region
- And an intermediate option, where raising at a slight angle.
- Nonstandard Nonstandard
If these links return or someone can decipher the lists of devices whose m-tone measurements that fall int the above categories, it might be interesting to see if people can correlate listening experience with the above categories. For instance, do devices that show IMD m-tone distortions rising towards higher frequencies sound harsh, brittle, sibilant?
(I tired to look at the page source to resolve the device names in the broken links but the page is in Russian so I can't identify the categories)
Anyway, that's one possible area of interest but the m-tone measurements may not correlate to audibility because it doesn't show any modulation aspects to this NF
One characteristic of auditory perception is that our perception is relatively insensitive to fixed noise (below a certain level) - it's easily categorized as a background sound & focus on foreground sounds unaffected. In modulating background noise, however, research shows that it interferes with our perception of foreground sounds - understanding speech in environments with modulating background noise, for instance.
Your comments seem somewhat contradictory, certainly 1930's electronics would have a fairly poor SNR by current standards. I agree though in general these effects are inside the head.
The fidelity of that horn was most likely due to its extreme efficiency. The demands on the driver and the amplifier would have been a very low, it would be interesting to know exactly how low.
Giant Denman Horn goes on display at Science Museum - BBC News
Last edited:
Thank you for pointing out this discrepancy.
Looking back at what I wrote, I realise I was trying to say that the spatial information I was referring to was contained not just in the signal, but also the noise, and the ruthless efficiency of this behemoth of a horn enabled what was a holographic listening experience. Of course, I maybe utterly wrong.
Also, the British at this time were outstandingly good with electronics, which greatly benefited the American post war economy.
As you know George, I have the same odd phono stage as you (at least for MM) which happily adjusts when you series connect the cartridge so the output is the same, so no worries there. People who have not shed the burden of inductance may indeed have the problems you list. 🙂
TOS, your listening situation is unique, but not that far away from many amateurs and professionals here. A single horn most probably will never give you a wide frequency response. It appears almost impossible at the low end, due to size constraints, and to high frequencies due to driver constraints. You can have a great 'midrange' and that is what you have. The problem was well known in the 1930's when cinema horns were designed, and to get wider bandwidth, multiple horns were put together (like they are today) each specializing in a specific frequency range. This is obvious by the historical work of many famous loudspeaker manufacturers, especially from the past examples over the decades.
Now the problem is that multiple horns tend to damage the sound about as much as they actually improve it. I know this for sure, because I lived with K-horns, first mono, then stereo for about 15years. I also designed full range horn loudspeakers with John Meyer (now of Meyersound) for a couple of years, where we had to interface 3 horns, hopefully seamlessly, back in 1974-1975. We certainly improved on the K-horns, but we still had far to go, but not for not trying to throw money, time, and engineering skill at the problem. It is just a series of almost impossible to manage problems by the horn approach, (or any other).
I gave up on my K-horns about 35 years ago, mostly because I found highest quality direct radiators (or electrostatics) to give a more accurate rendition of the human voice, and I still use direct radiators today. However, direct radiators have the problem of FM (Doppler)distortion that is always there, and it does seem to add a veil (or cloud) in the very low background. Only electrostatic speakers match horns in removing this, but then they have their own problems, as horns can sound effortless, and electrostatics can sound 'constricted'.
Stereo just aggravates the problems with horns, because they usually will not stereo image well due to path length differences. So giving your listening situation, your degree of investment, and what kind of music you prefer, you must chose the optimum speaker type.
Now the problem is that multiple horns tend to damage the sound about as much as they actually improve it. I know this for sure, because I lived with K-horns, first mono, then stereo for about 15years. I also designed full range horn loudspeakers with John Meyer (now of Meyersound) for a couple of years, where we had to interface 3 horns, hopefully seamlessly, back in 1974-1975. We certainly improved on the K-horns, but we still had far to go, but not for not trying to throw money, time, and engineering skill at the problem. It is just a series of almost impossible to manage problems by the horn approach, (or any other).
I gave up on my K-horns about 35 years ago, mostly because I found highest quality direct radiators (or electrostatics) to give a more accurate rendition of the human voice, and I still use direct radiators today. However, direct radiators have the problem of FM (Doppler)distortion that is always there, and it does seem to add a veil (or cloud) in the very low background. Only electrostatic speakers match horns in removing this, but then they have their own problems, as horns can sound effortless, and electrostatics can sound 'constricted'.
Stereo just aggravates the problems with horns, because they usually will not stereo image well due to path length differences. So giving your listening situation, your degree of investment, and what kind of music you prefer, you must chose the optimum speaker type.
This stuff is fun to talk about in any case. The only technical measurements I could find say 32Hz - 6KHz with a very small "sweet" spot and the compression driver was electrodynamic and made by Western Electric. I can see how the effect of standing in front of this is part of the experience.
It should not be too hard to replicate something like this, you make the small end out of metal work and the big end could be cast into your foundation like Dick Burwen did. You could bury the small end in your yard with a small access box for the business end. I suspect much of the effect is the classic exponential horn "sound" in spades.
Pretty sure Keith Johnson et al don't actually give a sh8t about mono compatibility. For sure anything for the masses needs to worry about how it sounds on a sonus, but I've never seen any mention of that in interviews with the people who produce acoustic music for the consumption of the high end crowd.
I'll test my Telarc 1812 CD in mono 😀
There was that "Royal" something or other a few years ago? seems to be flaky new age something now,
gizmodo.com/5025867/horn-subwoofer-takes-up-crazy-mans-entire-basement
gizmodo.com/5025867/horn-subwoofer-takes-up-crazy-mans-entire-basement
Last edited:
- Status
- Not open for further replies.