sibilance on RF-7 (cont'd)

[prego]

I'm openning a new thread to deal specifically with an experiment I did in order to track down sibilance on a dynaco 70 & RF-7 setup.

Background: I've already ruled out source/dac/cables material/wiring(bi or not)/interconnects/amp tubes(not 7199s though)/preamp/pre-tubes/RF feedback/ac-noise/acoustics. It happens on all volume levels, no clipping issue for sure. Crossovers were moded by DeanG and caps & horn were burnt in for at least 20 hrs consecutively by a combination of sine wave sweeping and pink noise in considerably high volume level. The sibilance is still easily noticable on almost all female voice sessions. I've also checked an average solid-state integrated amp (rotel 06) and although sibilance is less exaggerated, it's still there. This issue of RF-7 and sibilance is well known, and I've found some threads mentioning this (for examp: http://forums.klipsch.com/forums/thread/166062.aspx). However, none which I've found had given a path of course on how to resolve this issue thoroughly without dumping the speakers. I will give it a try then.

Now to the experiment I just did. I have a specific segment in a track I got on a cd (of female voice; Katie Meula) which exaggerates sibilance very very noticably, and is taken by some audio critics as a sibiliance test. As an anecdote, Katie was born and raised in the country of Georgia (east europe) and moved into N. Ireland at the age of 8, afterthen relocated to England where she debuted at the age of 14, so this is maybe why her English accent is patologic and great as a sibiliance reference. I've ripped this track using an accurate digital extracton into a 2ch 16bit 44khz normal WAV file. I've used a wave editor (cooledit) and passed the segment through a high pass filter I created, which takes off everything below 2000hz, leaving the rest unaffected (0db) using FFT. I then burnt this filtered segment(*) on a cd in exact digital forma. I took the right speaker's horn out and unwired it from the crossover, and instead I wired it to the right speaker's terminal output of the dynaco directly. The left channel remained unaffected(crossover in action). So I've played the new track and compared the sibilance one against the other (L vs R), and the result is that the channel with the crossover bypassed exhibits no exaggerated sibilance at all (regardless of volume level), while the left channel (crossover in action) sounds as before. Of course that I made sure previously that the same amount of sibilance was coming out from L and R equally. This is the only evidence I got in tracking down sibilance.
Since the crossover is moded (resistance on minus to minus path was reduced from 2ohms into 1.66ohm, so p=i^2*R; see: http://forums.klipsch.com/forums/storage/6/588211/HF.jpg), I initially suspected that. But, given the impressions of the vast majority among forums members who had moded the crossover, this shouldn't be the case.

I'd like to read from people whom once tried an active crossover(instead of the stock), or others that tackled this issue before with this setup, and changed into a decent amp which fixes this issue. My target is to make everyone happy, staying with the RF-7 for smooth 2ch listennig, not necessarily with the amp.

(*) If anyone's interested in the sample then PM me.

BTW: RB-5 exhibits normal sibiliance on the sample.

[leok]

This could be getting into some old territory where several of us continue to disagree. I use Jensen paper-in-oil caps for the direct series elements and Solen "Fast" capacitors for the parallel filter circuits. I Voltage conditioned all caps (subjected them to DC Voltage, both polarities at about half of their Voltage ratings or 150VDC whichever was lowest). I also use power line filtering for all sources and amps.

There is, as a rule, no sibilance distortion/exageration from my RF-7s using any of the three amps listed. I say, as a rule, because there is such distortion occasionally, but not so often that I find it a problem. It's a tough problem but it can be virtually eliminated. The RF-7s have a very clean high end and any distortion there is painfully obvious. I do find, however, that once the system is cleaned up the RF-7s are a delightful speaker.

Leo

[prego]

what about the mod, why increasing current in the horn x-over board was found legitimate, in terms of numbers rather than taste... e.g.: there is the varying impendance of the horn as a function of frequency, and its db response. on the other hand there is current and volts that the amplifier perturbate. essentially the thing that moves the air is power to my understanding, then why the mod. thanks.

[leok]

The circuit involved is a bandpass shunt across the horn to flatten its response at frequencies where it peaks in efficiency. Lowering that resistance was one of the first things I did to tame what I thought was a peak in the horn's output (the resistance reduction I used is a bit less than Dean's). The speaker is so smooth now I should try removing that shunt. On the other hand, the speaker is so smooth now, maybe I should leave it alone.

Leo

[prego]

thanks leok. can you please provide some theory reinforcement. I'm curious in gerenal about experimenting with the LRC circuit over the board. What do you mean by what I thought was a peak in the horn's output , were measurments taken to show that the mod flatten the response and not just the that peak and/or that it generated peaks elsewhere.

[Audio Flynn]

It is hard to commment on your issue when you do not have a equipment profile filled out. LEOK has a plenty of experience with RF-7, tube amps and networks.

Stream of concousness:

-consider power filtering

-How extensively has your ST 70 been modified?

[Zen Traveler]

I'm openning a new thread to deal specifically with an experiment I did in order to track down sibilance on a dynaco 70 & RF-7 setup.

.. The sibilance is still easily noticable on almost all female voice sessions. I've also checked an average solid-state integrated amp (rotel 06) and although sibilance is less exaggerated, it's still there.

...Now to the experiment I just did. I have a specific segment in a track I got on a cd (of female voice; Katie Meula) which exaggerates sibilance very very noticably, and is taken by some audio critics as a sibiliance test.

..... My target is to make everyone happy, staying with the RF-7 for smooth 2ch listennig, not necessarily with the amp.

I am mostly a multichannel connoisseur and love the way female voices sound on my system. That being said, I played my Katie Melua CD, "Piece by Piece" and although it sounds nice, I understand why it could be a sibilance test. I played it in 2 channel on RF-7's, RB-75's and KLF 20's, and wonder how it would sound on non Klipsch speakers...Isn't the phenomena on that disk in the recording and not the fault of the playback?

Insofar as not liking the way female voices sound on the RF-7--I have heard this comment before and as much as it baffles me, I have to think that maybe RF-7's aren't for everyone. [^o)]

[prego]

I'm not sure it's a matter of taste. They simply underperform at several elementary tests, such as reproducing accurate voice, IMO.

[Zen Traveler]

... That being said, I played my Katie Melua CD, "Piece by Piece" and although it sounds nice, I understand why it could be a sibilance test. I played it in 2 channel on RF-7's, RB-75's and KLF 20's, and wonder how it would sound on non Klipsch speakers...Isn't the phenomena on that disk in the recording and not the fault of the playback?

[^o)]

[prego]

no no and no

the phenomena is of the rf-7 peaking in that range.

second, I was refering to your statement of whether rf-7 is for one and not the other

[leok]

I am certain the problem you are having with sibilance is not the "peak" I refered to, and at this point I am curious as to whether or not I would hear a peak if I removed the resistor in my RF-7s. I suspect I was addressing a distortion problem by messing with the frequency response. Since then I have made several much more important changes in reducing distortion. So, over the next few days I will remove the resistor and report back.

Leo

[Guest srobak]

Too much "S" in your Klipsch?

De-essers For The Win

http://www.guitarcenter.com/BSS-Audio-DPR-402-Dual-Compressor-De-esser-Limiter-186305-i1125372.gc

[prego]

Problem solved.

Just to make sure and prepare evidence, with giving respect to people, I will report soon.

As a teaser, I can assure that it will certainly amaze many people that hold the RF-7.

[TheEvan]

My breath is appropriately bated (NOT baited).

[jacksonbart]

he is the master

[Deang]

Edit: Post removed.

[Deang]

They are from Israel, so since English isn't their first language, we need to cut them some slack -- but yeah, pretty funny -- the Hans & Franz thing actually ran through my mind too!

[Trey Cannon]

They are from Israel, so since English isn't their first language, we need to cut them some slack -- but yeah, pretty funny -- the Hans & Franz thing actually ran through my mind too!

I deleted that post Dean. I know he was trying to make light of the thread, but I thought it would just make things worse.

I did find it funny as well, but I am not sure others would.

Thank you for understanding.

[Deang]

Not a problem at all, thanks Trey.

[DrWho]

If the sibilance problem with the RF-7 isn't minimum phase, then no

amount of EQ in the crossover can fully remedy the problem... [^o)]

It'd be nice to see some measurements of what's going on. I've tried

some myself, but have not been able to find any kind of peaks.

Resonances wouldn't be accurately visible when the measurement is taken

in a room either.

Does anyone have any impedance response measurements of the RF-7? The

differences between output impedances on amps can certainly be the

difference when peaks are concerned. We've been doing a lot of

measurements in the lab lately and output impedance changes that result

in as little a 0.25dB changes in the frequency response are actually

audible. In fact, they're enough to make the difference between ear-bleeding and smooth... [] I think it has more to do with the damping of cone

movement than the actual frequency response change though...