I'd have to say, sreten's idea is a compelling one. Tweeters do seem to come in hard when crossed higher to a larger woofer.
sreten said:The "dip" is a compensation for the sudden increase in off axis
response when the tweeter comes in above the usually large
midrange units in the BBC's case, nearly all are 8" or larger.
Its not needed and doesn't work well with small midranges.
sreten.
With my own speakers, I heard a large, sudden level increase (subjectively) at the xo point and slightly above with a combination of a 6.5" mid and an 1" dome when did a stepped sine wave test. Moreover the on-axis tweeter level (gated) and at the listening spot tweeter level was even a few dB down compared to the mid. Even, the tweeter had lower harmonic distortion and better CSD response than the mid driver. The xo point was relatively low at 2kHz where the mid is yet not beaming seriously. Now I not hearing that subjective level mismatch with a waveguided tweeter. The room's RT60 is below 0.3 s and fairly even across the spectrum, if anyone suspects this.
Reducing the tweeter level around the xo point and where their dispersion is very wide (compared to the mid driver), maybe it might help somewhat, but it's only a cosmetic solution I think.
Last edited:
Maybe it is some sort of a compromise that does both at the same time to some degree.
Regards
Charles
Edit: YSDR was quicker.
Last edited:
I went through a phase of crossing low and shallow. Both based on hearing a gap. Clearly it can be an issue.
What you say next is interesting, I can relate. There are a couple of secondary sources that make guessing about this uncertain.
YSDR and phase_accurate, as soundbloke said earlier:
"Essentially our hearing is less sensitive to laterally originating sound sources in the frequency band in question. When such energy is encoded in a stereo recording and then replayed from (predominantly) the front, the reproduction will appear overly-bright. Hence a dip compensates for the apparent excess energy in this band (to a degree dependent on all the factors I have described previously)."
"Essentially our hearing is less sensitive to laterally originating sound sources in the frequency band in question. When such energy is encoded in a stereo recording and then replayed from (predominantly) the front, the reproduction will appear overly-bright. Hence a dip compensates for the apparent excess energy in this band (to a degree dependent on all the factors I have described previously)."
But then, why not compensate that brightness at the mixing/mastering stage? Why we need to compensate with the loudspeakers? Isn't mixing/mastering done with speakers?
If I understand your sayings correctly, of course.
Last edited:
I was playing around with my DSP crossover last night and I always came back to the flattest frequency response as being the best sounding. I tried a lot of things like tilted response and also an upper midrange dip that give some nice warm sound. Especially on recordings that sound thin in the bass range. But if judged over many different types of recordings a linear frequency response (+- 1 dB between 100 Hz and 10 kHz) sounds the most accurate to me on average (purely subjective of course). Especially with modern bass-heavy tracks the tilted response is quickliy getting annoying (although I like lots of bass). With the BBC dip I do not only have the feeling that the phantom images are furher away but also that their lateral positions change.
But I must also say that my speakers have less disperion in the HF range than those with dome tweeters. For those a tilted response might indeed have advantages. It is intended to build an analog crossover as soon as everything is sorted out by using the DSP one. And that analog active crossover may include some Control for tilt and a tuneable BBC dip.
Regards
Charles
But I must also say that my speakers have less disperion in the HF range than those with dome tweeters. For those a tilted response might indeed have advantages. It is intended to build an analog crossover as soon as everything is sorted out by using the DSP one. And that analog active crossover may include some Control for tilt and a tuneable BBC dip.
Regards
Charles
Some compensations can be done at the mixing and some can not, listening to music throu a stereo speaker setup is more complicated then it might seem to be. We might do many corrections to our loudspeakers without really knowing why we do them, we found them to sound good but do not understand why, so we apply resons or explanations that we only think is correct or accurate
I have been through this many times, and cannot follow the logic.
From YSDR;
"But then, why not compensate that brightness at the mixing/mastering stage? Why we need to compensate with the loudspeakers? Isn't mixing/mastering done with speakers?
If I understand your sayings correctly, of course."
I agree, the situation at the mixing mastering stage is almost identical to that in the listening room, and so the principle, if right, should apply in both situations, and any 'correction' surely would need to be done in the former process to sound right to the mixers and masterers.
From clef;
"Some compensations can be done at the mixing and some can not, listening to music throu a stereo speaker setup is more complicated then it might seem to be. We might do many corrections to our loudspeakers without really knowing why we do them, we found them to sound good but do not understand why, so we apply resons or explanations that we only think is correct or accurate"
This is effectively using a subjective judgement, and as we all know subjective hearing is fickle and often inaccurate.
As previously stated, in the original BBC designs there was no attempt to design the responses to be anything other than flat, they became so when Rogers manufactured them, for seemingly unaccountable reasons.
Gundry's son has stated that his Father would not do that.
Last edited:
@ phase_accurate
This correlates with my findings as well.
With controlled directivity, flat is the best to me too.
With the uncontrolled directivity pattern of a dome tweeter, the room/environment is critical.
In my situation, the normal dome version sounded too bright, even if I dialed back the tweeter a few dB.
This correlates with my findings as well.
With controlled directivity, flat is the best to me too.
With the uncontrolled directivity pattern of a dome tweeter, the room/environment is critical.
In my situation, the normal dome version sounded too bright, even if I dialed back the tweeter a few dB.
The logic is that the source material remains preserved. Compensating for head shadowing by any means is not energy preserving.
A simple dip will never be accurate, subjectively or objectively: It is a compromise that depends on the spatial arrangement of the sources being recorded and on the recording acoustic.
For the reasons stated by others previously in this thread, there is a certain amount of serendipity in targetting a dip - and quite possibly its origin was one of chance as per many far more important discoveries.
It is also worth noting that the practise was not used throughout the BBC, where, for instance, the equally high-fidelity minded Transcription Unit did not use the same loudspeakers and did not exploit any such dip.
I feel that there is almost no engagement between our beliefs.
If I am sure of my position in a debate, I support it with a clear coherent argument, showing the steps and deductions in that argument, each a logical follow-on from the last, and forming a coherent set of steps leading to a conclusion.
I am also aware of how sophism can result in failure in the structure of argument.
"It is also worth noting that the practise was not used throughout the BBC, where, for instance, the equally high-fidelity minded Transcription Unit did not use the same loudspeakers and did not exploit any such dip."
My guess here would be LS 5/5s, at least that was what was used there when I visited.
Are you by any chance either now, or have been, a BBC research chap?
Ethen Winer, whom I respect seems to think that any dip was a deliberate implementation, whereas much information seems to indicate that no dip was intended, and that the original design was for a flat response as previously stated, but deviated from by Rogers. This view is also shared by Alan Shaw.
If I am sure of my position in a debate, I support it with a clear coherent argument, showing the steps and deductions in that argument, each a logical follow-on from the last, and forming a coherent set of steps leading to a conclusion.
I am also aware of how sophism can result in failure in the structure of argument.
"It is also worth noting that the practise was not used throughout the BBC, where, for instance, the equally high-fidelity minded Transcription Unit did not use the same loudspeakers and did not exploit any such dip."
My guess here would be LS 5/5s, at least that was what was used there when I visited.
Are you by any chance either now, or have been, a BBC research chap?
Ethen Winer, whom I respect seems to think that any dip was a deliberate implementation, whereas much information seems to indicate that no dip was intended, and that the original design was for a flat response as previously stated, but deviated from by Rogers. This view is also shared by Alan Shaw.
Last edited:
The Transcription Unit used soffit-mounted ATC SCM200s. I am not sure the unit still exists, however.
No.
I would put my money on a "retrospective invention" and that serendipity.
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.