Here's what SL wrote about it:
"The midrange to tweeter transition region from 700 Hz to 3 kHz for the 1.4 kHz LR4 crossover was now level with the frequency regions below and above for a flat overall response. Extensive listening tests with known program material that I and Don Barringer (on the East Coast) performed with our modified Orions pointed to about the right level setting for the tweeter, but were otherwise not quite on the mark. When analyzing the theoretical shape of the frequency response curve it seemed obvious that the transition from midrange to tweeter was too steep. We therefore left the tweeter at its flat level setting and started to use shelving lowpass filters to shape the response. Immediately we found great improvement . . . "
ORION-3
Is he wrong about what he did ? ? ?
Deward,
No, he's not wrong, but nowhere in that statement does he specify that an electrical -3.2db shelving filter was applied to a "flat" acoustic response to achieve the final result. (You fellas are inferring something that he doesn't actually say.)
The "net" result of the ASP modification is a down shelving action, yes, (yielding a final non-flat acoustic response) but not anywhere near a -3.2db acoustic down-shelving action. It's less than 1db.....relative to the pre-Orion-3 specification/response.
I've posted electrical plots of the ASP before/after response.......I've explained the details of the electrical modification......I've stipulated that SL's description is confusing......etc, etc. John K. appears to be the only one grasping this modification.....and even for him it took awhile. 🙂
Why is this so difficult to grasp?
Cheers,
Dave.
No, he's not wrong, but nowhere in that statement does he specify that an electrical -3.2db shelving filter was applied to a "flat" acoustic response to achieve the final result. (You fellas are inferring something that he doesn't actually say.)
The "net" result of the ASP modification is a down shelving action, yes, (yielding a final non-flat acoustic response) but not anywhere near a -3.2db acoustic down-shelving action. It's less than 1db.....relative to the pre-Orion-3 specification/response.
I've posted electrical plots of the ASP before/after response.......I've explained the details of the electrical modification......I've stipulated that SL's description is confusing......etc, etc. John K. appears to be the only one grasping this modification.....and even for him it took awhile. 🙂
Why is this so difficult to grasp?
Cheers,
Dave.
Perhaps because further down in that link he writes the following:
"Revision 3 of the ORION is now in its 3rd implementation with Version 2. This became a necessary evolution because the frequency response of the loudspeaker can not be flat for stereo reproduction. The response has to roll down towards higher frequencies, but how?"
and presents a graphic showing Ver.0 (the referant in the previous post) at -1.8 dB, and Ver.2 (current) at -3.2 dB for the tweeter rolloff (relative to "Flat Woofer & Tweeter Setting") ? ? ? That graphic is pretty explicit . . . it's not difficult to see why it would be taken literally. You seem to be comparing the present tweeter response curve to some previous (unpublished) response curve . . . SL's own words seem to compare it to "flat" . . .
See my post 471
The 3.2dB are compared to flat and "less than 1dB" is compared to Orion pre rev. 3.
Again, Orion was not flat before because it would have been too bright (and that is kind of normal with such speakers).
The 3.2dB are compared to flat and "less than 1dB" is compared to Orion pre rev. 3.
Again, Orion was not flat before because it would have been too bright (and that is kind of normal with such speakers).
Sure ?
It was the Orion 3.2.1 modifications and SL's analysis of the technical basis of those mods that launched this thread.
As I read (and experienced) the “history” it went as follows:
+ Adding the rear tweeter made Orion+ too bright (if there was no tweeter level change)
+ Lowering the overall tweeter level (to correct the brightness) caused a noticeable midrange dip in on-axis response
+ That was “corrected” (to some extent, in rev3) by (mostly) restoring the overall tweeter level while adding a 1.8dB hf rolloff (among other attempts) with a higher inflection point and some other midrange adjustments
+ That proved not enough, and subsequent revision led to a 3.2dB hf rolloff on-axis (rev3.2) while maintaining the higher inflection point, which is where it stands now.
The “other” obvious solution, keeping it flat on axis and applying the hf rolloff only to the rear tweeter was not done because it would require additional amplifiers and electronics (not available in the asp) or a “passive” shaping circuit on the rear tweeter (fairly easy to do, but wouldn’t that be weird on an otherwise all-active loudspeaker). And there’s no single “right answer”, because so much depends on the behavior of the room . . . particularly the wall behind the speakers.
It all keeps coming back to the question of why we need the roll-off in power response, and whether it is acceptable to get it (in a constant-directivity loudspeaker) by rolling off the on-axis response or better to get it by changing the speaker’s polar response as frequency rises (which is what happens, by the way, in PLUTO).
I thought SL's main arguement was that he started with free field flat (outdoors, on axis) and found that too bright. He realized that a speaker at +- 30 degrees was at an angle where HF hearing is stronger, and so a centrally placed phantom source (equal level both channels) would be too bright. The several dB/decade slope was a correction for speaker position. Flat anechoic response is still his basic assumption of the best starting point.
In this discussion, whenever anyone says "flat" they really should be more precise. Flat on axis anechoic? Flat power? Flat reverberent field in a live room? They all mean different things.
David S.
In this discussion, whenever anyone says "flat" they really should be more precise. Flat on axis anechoic? Flat power? Flat reverberent field in a live room? They all mean different things.
David S.
I think you have to take into account the source material, too. As a simple example, suppose you have a jazz trio that was recorded close-miked. There is little ambient room information on the recording, so I want my system anechoic flat to reproduce the sound as if the trio is in my room.
If, on the other hand, the trio was distance-miked and there's lots of room ambiance on the recording, I want my system flat at my listening position. I want to minimize my room's affect so I can hear the room the trio was recorded in.
In the first case, I'm moving the trio into my room. In the second case, I'm moving myself to club where the trio is performing.
Not much is distance-miked any more except orchestras and some audiophile live recordings. Studio recordings are generally close-miked and then sweetened with reverb to make it sound like there is ambiance.
Personally, I start with flat at my preferred listening position and then eq to what sounds subjectively good to me. Is it accurate? Who cares, I'm trying to make my ears happy, not my RTA.
-dr_vega
I've also been saying this for years, and graphically show the impact earlier in the thread.
That stereo has this error is beyond argument. It just is. How any specific recording compensates, if it does, is unknown. If it doesn't, the playback will show the error. There's no way without DSP to compensate on stereo playback except maybe setting a frequency response curve somewhere in the playback chain that is some compromise between 30 deg and front center.
This issue largely goes away with a real center channel.
Dave
Recordings are still mixed and mastered. Any arguments there? Until they are not, they are compensated for. Maybe if more audiophiles would do some recording, they would understand the issues with mics, patterns, and distance/room interactions or lack of. In the face of overwhelming evidence it's hard for me to imagine why it's so hard for intelligent people to understand. Belief has a powerful grip. Engineers/producers are listening to these recordings at least dozens of times before they put them for sale. Sometimes it is hard to tell. I'll give you that. Not all of these people are deaf idiots--though some of them are. Stupidity is more dangerous than evil right? And there are idiots in every field. If their speakers had this HRTF mod, I'd agree. Never the less, speakers need to be fairly uniform in response(this has been shown several times as we all know but some forget) and none of my speakers sound as if they need this adjustment and they are all pretty much Flat. Making a better room is the icing on the cake.
The Orion's compensation is to fix the Orion's problem without tweaking the room. It is not a universal target for all speakers. The Orion is a narrow bass/midbas pattern with a wide MR and lower treble pattern dipole speaker. Radically different than most studio monitors. No surprise that its FR/PR might need some tweaking and bravo to SL for figuring it out and having the gut to openly admit it. Improving it now even further it appears.
Dan
The Orion's compensation is to fix the Orion's problem without tweaking the room. It is not a universal target for all speakers. The Orion is a narrow bass/midbas pattern with a wide MR and lower treble pattern dipole speaker. Radically different than most studio monitors. No surprise that its FR/PR might need some tweaking and bravo to SL for figuring it out and having the gut to openly admit it. Improving it now even further it appears.
Dan
An Interesting discussion because I have seen all this myself. With a very wide HF directivity an listening axis that is flat is very bright. I design for a listening axis fall-off with frequency of about 3 dB from say 3 kHz to 15 kHz. Nothing extreme, but definately necessary. I also find that in a small room a flat LF response will sound lacking and a + 3 dB per decade rise below about 200 Hz is about right. This of course is for a steady state condition, while the HF above is based on a semi-anechoic condition. In a steady state measurement in a real room the fall-off tends to be greater because of the greater HF absorption in most rooms.
Wow, that sounds like a lot. But in a small room measured with an omni mic, it probably isn't.
I think the little bit of low bass boost that most prefer has to do with the relatively high level at which recordings are often mixed and mastered. Doing this myself, I can understand why it's done this way--you can hear more minute details. However, I don't see why they can be mixed/mastered that way, then turned down to a more normal listening level and rebalanced. Anyway, it's easy to twist knobs either way or get a decent receiver.
Dan
Dan
yes 😛
No... but they generally have very dead studio compared to the average homeowners listening environment.
Not all.... but certainly many. 😀
It obviously varies greatly.... there are many great recordings but not so much consistency.
Agreed
SL would probably do what every other recording engineer does.... he would record live sound using the most neutral mics possible or in the case of studio recording he would likely besing relatively small studio monitors in a fairly dead mastering room of the recording studio.
I'm guessing but it is likely that at the end of the day or the next day everyone would go home and listen again;
- SL on his Orions in his lively living room (with DSS filter)
- Bono (U2) would listen on his iPod 🙄
- and the Metallica would listen in their vehicle on the freeway. 😱
Later everyone would probably argue about what should be changed to ensure it still sounded good in a great variety of situations/environments.
Not at all... I think it's great that you participate in a lively debate such as this. Someday maybe we can argue (in another thread of course) about how best to compensate for the acoustic impulse delay of a 10" woofer vs 6.5" midrange. 😉
Really?The phantom center usually sounds duller to me, not brighter.
http://www.diyaudio.com/forums/multi-way/172806-flat-not-correct-stereo-system-19.html#post2363393
- Status
- Not open for further replies.