Jump to content

A K-402-Based Full-Range Multiple-Entry Horn


Chris A

Recommended Posts

9 hours ago, PrestonTom said:

This is new territory for me.

I identified the application and the tool within that application (Audacity and the tool "Clip Fix").  I recommend reading the available material online using Google as your source-finder.  This is not the thread to do that.  I recommend starting a new thread or reusing this one:

 

I should mention the following--if you have clipping in music tracks, they have been damaged.  Unfortunately, that includes about 95% or more of pop and rock genres produced since 1991. 

 

You then have a three-way fork in the road:

  1. keep listening to the clipped tracks and try to ignore the sound that you're getting. 
  2. don't listen to the tracks, and store the discs away to collect dust, never to be listened to again.
  3. try to repair the major damage done during mastering by using something like Clip Fix to reconstruct the main clipped peaks, then run "Normalize" to pull the overall loudness of the track down to some level below that which is above 0 dBFS (I use -0.3 dB for that setting).

The result can be startling in terms of the subjective listening to the tracks.  Clip Fix is a very compute intensive task, taking perhaps a quarter or a half of the real-time music playing duration in order to compute the de-clipped tracks.  I use a macro within Audacity to do this for all tracks on a disc that I've just ripped where I've detected clipped tracks exist:

 

1778959811_ClipFixmacroorchainscreenshot.JPG.6f8e110e46660d597fc863da940f8914.JPG

 

This allows me to start the process and walk away to do other things until the clip fix process is complete. When it is, I have a new sub-directory with fixed tracks below the level of the directory where the original tracks exist--called "cleaned".  I can then listen to those tracks, or just replace the original tracks that were ripped.  The difference in sound of the new tracks vs. the original tracks varies from inaudible to very audible.  In all cases, the new tracks sound the same or much better to my ears, and are less loud when played back by the amount of the maximum amplitude reconstruction that has occurred during clip fix.  Additionally, if you're looking at dynamic range figures for the album and individual tracks (www.dr.loudness-war.info), the buried dynamic range of the tracks can be more accurately assessed after running clip fix.

 

So what does clip fix actually do?  It removes the odd-order harmonics induced in the mastered tracks due to clipping--harmonics that were generally not there during recording of the original music in the venue. Normalize brings the resulting level of the tracks back below 0 dBFS in order to save the tracks without clipping and to reduce their loudness back down to something that is more in line with real music loudness levels.

 

This is all about "loudness envy" and the Loudness War to make tracks louder.  There is no other reason to clip tracks during mastering.  This is the first and most egregious behavior of the record industry that leads me to reject that culture and to distrust all that I receive from it. 

 

Chris

  • Like 2
Link to comment
Share on other sites

By the way, when doing clip fix on multichannel music tracks within Audacity, the command "Normalize" cannot be used--as this only works for stereo tracks.  Instead, the function "Amplify" will return the (negative) level above 0 dBFS that the clip-fixed track has generated, and that value can be used to reduce the overall gain of the track (all 6 channels simultaneously) to below 0 dB to avoid clipping in the saved tracks. 

 

Unfortunately, this requires each track to be declipped one-by-one since the macro function of "Amplify" will not be usable within the macro to find the resulting negative gain level to apply to the track being worked on.  This is a small price to pay for getting de-clipped 5.1 tracks.

 

Chris

  • Like 1
Link to comment
Share on other sites

2 hours ago, Chris A said:

Clip Fix is a very compute intensive task, taking perhaps a quarter or a half of the real-time music playing duration in order to compute the de-clipped tracks.

 

In the early 2000s I designed a motion picture soundtrack extraction system that included real-time clash detection and correction. ("Clash" is where the audio waveforms from two soundtrack channels are modulated so heavily that they overlap -- the effect is identical to what we call clipping.) I am not at liberty to say how I did it, but I can say that the reconstruction is an approximation, and that the approximation sounds considerably better than the clipped waveform. 

 

The problem with audio waveforms clipped for the loudness wars is that "easy-over" compression is used. The effect is that the waveforms are clipped, but the edges surrouding the clips are rounded to reduce their audibility. That also makes it difficult to approximate the unclipped waveforms because the actual onset of clipping is difficult to discern.

  • Thanks 1
Link to comment
Share on other sites

Yes, but not all tracks have those "rounded peaks", and Clip Fix does work well in finding those embedded clipped peaks on most tracks. 

 

The goal, of course, is to dramatically reduce the harmonics generation due to the clipping--which is mostly bass and mid-bass clipping.  Rounded peaks also reduce the generation of those peaks, but then again, someone would have to explain to the consumer why the clipped peaks are there to begin with.  Even the MP3 players now have a default "on" replay-gain-like gain control for each track to reduce the loudness to proper levels.  Unfortunately, they didn't implement real-time clip fixing at the same time--so almost all consumers are stuck with what they get from the record companies (RIAA).  Generally, the MPAA member companies (i.e., the movie soundtrack industry) historically haven't gone along with clipping for loudness-war loudness of tracks.  Unfortunately, I've found that's apparently changing.

 

Unfortunately, too that the higher frequency bands are also swept away by clipping the tracks, so in most cases, 30%-50% of the waveform in most pop/rock tracks produced today are gone forever (i.e., not just the peaks, but all the music in the tracks since heavy mixing and mastering compression results in almost all the peaks being clipped nowadays).  Does this affect the sound of the remaining track?  Yes.  However, there is nothing that I know to do about it, short of writing a very nice letter to the record company, asking them if you can instead download the mixdown tracks instead.  (Good luck with that, BTW...)

 

Chris

  • Like 2
Link to comment
Share on other sites

In the early days of CD mastering, if a submitted track had even one clipped peak, it would be rejected by the mastering house. Nowadays clipped peaks are standard operating procedure.

 

Take a look at "Hell Freezes Over" sometime.

  • Like 1
  • Thanks 1
Link to comment
Share on other sites

Greg, actually the DTS multichannel version of Hell Freezes Over (by the Eagles) looks good with no clipping.  I haven't demastered it by looking at the modification date of the files, so I'm pretty sure they're good. 

 

Chris

 

 

Link to comment
Share on other sites

3 minutes ago, Chris A said:

Greg, actually the DTS multichannel version of Hell Freezes Over (by the Eagles) looks good with no clipping. 

 

I only have the CD version, and it's full of clips.

Link to comment
Share on other sites

From the DTS 5.1 version, track 11 (Take it Easy):

 

1165702944_TakeitEasyAudacity.thumb.JPG.9a60403032f12ae726be90b9a9d95d6a.JPG

 

The track assignments are 1) left channel, 2) right channel, 3) center, 4) subwoofer (LFE), 5) left surround, and 6) right surround.

 

Chris

  • Like 1
Link to comment
Share on other sites

1 minute ago, Chris A said:

From the DTS 5.1 version, the same track (Take it Easy):

 

On the CD, Track 1 is Get Over It. No matter; the whole CD is similarly clipped.

Link to comment
Share on other sites

23 hours ago, Chris A said:

The best jazz multichannel albums that I've got (and among the best recordings that I own) are the Yellowjackets Time Squared, Altered State, and Lifecycle multichannel SACDs, along with Xiomara's La Voz multichannel SACD.

Thanks Chris, I just ordered Altered State. It should be here in short order.

Link to comment
Share on other sites

1 hour ago, rickmcinnis said:

Chris A would you give me an idea of where you placed your ports in relation to the mouth.

The ports are located no more than 5" from the throat entrance plane (I measured to the 2" compression driver throat screen) along the central axis of the horn.  If you measure along the sidewall, this is 6.5" inches from the throat entrance plane to the beginning radius of the woofer port (off axis port).  My ports were cut with a 2.25" diameter hole saw, and were lengthened along the side wall-->-top/bottom wall crease until the ending port radius reaches 9.5 inches, in two places for each woofer (two ports per woofer).  The total active woofer area to port area should be about a 10:1 ratio, i.e., "compression ratio" which is a rule of thumb. 

 

You can use a 2" diameter hole saw and simply lengthen the off-axis ports until the 10:1 compression area ratio is reached.  Place the ports along each crease of the horn wall as is shown in the pictures I've provided--NOT ACROSS THE CENTER MID-WALL OF THE HORN. 

 

Achieving the 10:1 compression ratio in a rule of thumb, but it strongly affects the low frequency efficiency/and subsequent cutoff frequency of the woofers.  I apply EQ to my K-402-MEH to extend its SPL response to very low frequencies as shown in the first post of this thread:

 

K-402-MEHwithDanley-StyleCrossoverFilters.jpg

 

You can read the -3 dB cutoff frequency for yourself...along with the controlled phase growth down to that frequency and below.  In this configuration, the K-402-MEH functions as its own subwoofer, relying on the combined woofer area of its two 15" woofers to achieve usable output.  The K-402-MEH picks up greater than 12 dB of deep bass extension when placed in full room corners (eighth space) relative to the mid-wall (half space) location shown above. 

 

Note that I don't need 120+ dBSPL output at 20 Hz from the center only loudspeaker (I've also got TH subwoofers in the corners with -3 dBSPL response down to 14.5 Hz), so this configuration is very handy to have to avoid having to place another subwoofer in the middle of my L, C, R frontal array of loudspeakers in order to fill up mid-wall infrasonic room modes.  If you have three K-402-MEHs across the front of a room, with the outside loudspeakers in the room corners, the need for separate subwoofers is sharply reduced.

 

Chris

  • Like 1
Link to comment
Share on other sites

1 hour ago, rickmcinnis said:

From what I read thickness of the ports has great impact on the upper frequency range of the woofer section so i wonder if using 3/8 inch LEXAN would be any better than sticking with 3/4 inch of a wood product?

Bricks would work better if you could figure out how to attach them to the horn walls. Stiffness is the name of the game.  Lexan (polycarbonate) isn't very stiff in my experience--tenacious yes (total bending energy to achieve buckling)--but the bending modulus isn't that high.  You're looking for high bending modulus.

 

The way to compensate for the thickness of the material used in the walls between the front of the horn and the woofers is to chamfer (countersink) the edges of the ports into something more like a knife edge on the woofer (reverse) side of the horn.

 

 

  • Like 1
Link to comment
Share on other sites

1 hour ago, rickmcinnis said:

These have a longer snout that the driver Chris A used which makes me wonder if this could have a big impact on the placement of the ports?

No.  Zip, no effect.  You need to read some horn theory/physics to understand what is occurring, and why this is true (Huygens principle).  The only time depth of compression driver is important is when setting up channel delays, i.e., the acoustic center will be further back for the deeper throated compression driver, but this doesn't affect the placement of ports in the horn.

 

Wavelength=slitwidth.gif

 

Chris

  • Like 2
Link to comment
Share on other sites

2 hours ago, rickmcinnis said:

1.Are you doing anything unusual to seal the mouth of the horn to the box

Sticky foam-backed tape, just like Klipsch uses on the K-402 to seal it against  the KPT-305 box.

 

2 hours ago, rickmcinnis said:

2. I assume you are not hearing any noises from the box.  One wonders if the two woofers tend to eliminate some box problems - kind of like djk's PPSL bass cabinets?

     a.  Did you add any bracing to the box?  Are you hearing any need to worry about adding bracing?

Again, what happens behind the horn ports generally stays there (unlike direct radiating woofer boxes).  You can put absorption in the back volume to control the biggest resonances, but in general, the concern should be about the stiffness of the horn itself, not really the box.

 

2 hours ago, rickmcinnis said:

 b.  Did you add any damping to the walls of the box like bituminous felt pads or something like that?

I added the equivalent on about a half of an Auralex Sonofiber panel (lint-based).

 

2 hours ago, rickmcinnis said:

c.  Are you doing anything to connect the walls of the horn to the box?  Do you see any advantage or disadvantage to attempting this?  Would there be any advantage to damping the walls of the horn since the ones I am getting is the lightweight version?

If you clamp the K-402 to the box at each front mounting hole that's already there, you'll have clamped the horn quite well (just like the KPT-305 mid-bass module uses). 

 

3 hours ago, rickmcinnis said:

I will not be asking them to go below 100 hertz.,  I am hoping this will minimize many potential box problems by  reducing the pressure on the walls of the box. 

This is not a good reason to cross over almost 3 octaves higher than the K-402-MEHs can easily handle.

 

3 hours ago, rickmcinnis said:

When I get ready to cut the ports I am going to send you a picture of where I plan on placing them and asking for your approval.

You don't need approval from me, but I'll certainly provide an opinion if you ask for one.

 

4 hours ago, rickmcinnis said:

THANKS, again, Chris A.

You're welcome.

 

Chris

  • Like 1
Link to comment
Share on other sites

48 minutes ago, Chris A said:

but in general, the concern should be about the stiffness of the horn itself, not really the box.

 

The thought just occurred to me about perhaps reinforcing a horn with fiberglass and resin. As long as you are not trying to sell them so you don't have to worry about keeping them stock is that of possible benefit?

  • Like 2
Link to comment
Share on other sites

Guest
This topic is now closed to further replies.
×
×
  • Create New...