Have you looked at the AudioControl DM-608, has up to 96khz sample rate.
Yes it designed for car audio but so is the Dayton 408
Yes it designed for car audio but so is the Dayton 408
The unanswered question is whether you modified gain from Fusion amps or not. If not, I can fully understand your remarks. Even for a ~89 dB/W woofer, I found the idle hiss too much...
And my comment was not at all about direct radiators with high power amp vs horns.
Just a note on the Thomann units - I have the cheapest one and it allows 0.02 ms step in the GUI, which corresponds to 48 kHz sampling frequency. The first 3 steps are 0.021, 0.042 and 0.062 ms which corresponds to the theoretical 0.02083 ms rounded up. I tried the GUI for the more expensive ones and it allows 0.01 ms step for the 96 kHz units, therefore offering a delay as a multiple of sample time.
It wasn't hiss at all with the FusionAmps. The gain staging was actually quite easy in the setup--no audible hiss at the listener positions (probably as a consequence of my having had the balance of the signal path already under control when I tested the FAs).
The problem lies elsewhere...and it's a monster of an issue although it wasn't actually identified via measurements. They sounded somewhat similar to the miniDSP Flex using its analog XLR balanced inputs, in fact. They were more like an audio clip of 1960s "distorted psychedelic rock"--coming and going in the transients, that was so popular at that time--basically unlistenable.
Chris
@Cask05 mentioned that "you can't really use with extremely high efficiency loudspeakers (i.e., fully horn loaded, like yours shown above) is the analog inputs--they sound horrible on a Heritage Jubilee. But note that the S/PDIF digital inputs (including TOSLINK optical) that also come with this unit haven't been tested yet. "
I can add to this since I'm using a MiniDSP Flex Eight (via its S/PDIF input) with horn-loaded speakers. I use the MiniDSP Flex Eight with my tri-amped 1986 Klipsch La Scalas and THTLP (horn-loaded) subwoofers.
In the fall, before purchasing the Flex Eight, I contacted MiniDSP tech support asking about the best way to connect my source (a Bluesound Node) to the Flex Eight. They recommended using the S/PDIF input (best choice), followed by TOSLINK (second choice), with analogue being less desirable.
I've been entirely satisfied with the Flex Eight and cannot discern any noise from it. I haven't tried seeing if I can hear distortion with either the analogue or TOSLINK inputs.
@Cask05, unfortunately diyaudio won't let me send you a PM until I have more posts. If you wish to PM me I'd be delighted.
Interesting. Somewhat surprised you had no audible noise with such sensitive speakers. Perhaps because you had the lower power model. Not sure what caused your problems. Perhaps the limiter was activated or the master volume was too low too have proper audio quality yet. You used or tried digital in?
Fedde
A horn-loaded subform would be awesome. we've pretty much turned the K Korner at AK into one, but participation is limited to about 6 refugees from the community.
I'm not sure that you've been paying attention, so I'll try to lay it out more clearly: My system consists of the following:
Denon DP-37F turntable -> Schiit Mani 2
Oppo BDP-103 -> Schiit Bifrost 2
MacBook Pro -> Schiit Bifrost 2
Apple TV -> Schiit Bifrost 2
All sources are fed into a Schiit Freya +, -> DSP Crossover -> Tweeters, x2 Schiit Rekkr 4 W Class A mono blocks,
DSP Crossover -> Mid horns are fed into a modded Glow Audio Amp Two 15 WPC P/P EL84 amp
DSP Crossover -> Bass horns are driven by a pair of Alesis RA-150 150 watt bridged mono amps.
they are then fed into a pair of heavily modded, 105 dB sensitive LaScala clones.
despite my mods, which allow an improved frequency response from 23Hz to 20Khz, the speakers still exhibit many of the same peaks and nulls inherent in the original LaScala design. They sound better than decent, but the frequency curves show that there is considerable room for improvement. This curve in blue, shows the passive crossover, with crossover points at 400 Hz and 6000 Hz, with 12db Linkwitz Riley filters:
note the nulls at 55Hz & 350 Hz and a resonant peak at 175 Hz.
utilizing 24dB Linkwitz Riley filters and parametric eq in the Dayton Audio DSP 408, I'm able to tune the response to +/- 3dB across the frequency spectrum:
the difference is noticeable and a marked improvement over passive. It is not possible, nor is it desirable to go route analog signals to a passive crossover and keep digital signals digital, nor would I want to if I could. The flaws with the DSP-408, notably single-ended inputs and a nominal amount of treble hiss are negligible and something that I can live with, however, for me, the gist of the hobby is building a better mousetrap, gradually improving the system, learning the why's and how's along the way, until a nebulously defined endgame has been reached. To that end, the most obvious point of improvement is in the DSP crossover, followed by the amps driving the tweeters, phono pre and perhaps the amps driving the woofers, in that order, affecting what are likely single digit percentages of improvement.
To reach that hair-splitting level of improvement, I would like balanced connections, better sample rates and 8-10 bands of parametric eq per channel and above all, no added noise.
I think markw4’s key point is that you’re still going to need an ADC for your analog source(s). There has to be a compromise somewhere. I know that my vinyl setup sounds better direct analog only, but I value DSP EQ of my speakers/room just slightly more than that pure analog experience because I listen to digital sources way more often. So I use a Focusrite ADC for the signal out of my phono pre so I can include it as an option into my Flex Eight. Records still sound very good and I enjoy the experience. This is my best of both worlds.
Given your parameters, that Thomann unit should work very well for you.
Given your parameters, that Thomann unit should work very well for you.
Yes, I tried it after using the XLR analog inputs. No perceptible change in symptoms. As soon as I replaced them with the Xilica XP series crossover and stereo amplifiers (First Watt F3 on HF, Crown D-75A on LF), the problem disappeared.
Chris
gigantic, what is the phase response of your system ? Would you post a screen shot from REW of spl with the phase.
The reason I ask is that phase response is historicly ignored in horn loaded Klipsch speakers due (I believe) to the difficulty of driver time alignment with passive crossovers.
ChrisA (Cask05) brought phase response into focus here https://community.klipsch.com/index...y-effects-of-quasi-linear-phase-loudspeakers/
The reason I ask is that phase response is historicly ignored in horn loaded Klipsch speakers due (I believe) to the difficulty of driver time alignment with passive crossovers.
ChrisA (Cask05) brought phase response into focus here https://community.klipsch.com/index...y-effects-of-quasi-linear-phase-loudspeakers/
ok, I think I get @Markw4 's objection now. @Wirrunna brings up the point that I didn't emphasize enough that the compromise of the 2nd ADA conversion in the DSP crossover is offset not only by EQing the speakers to a flat response, but achieving a quasi-linear phase response and time alignment. People with direct radiating speakers don't really need to take this into consideration, as their drivers are all within 3-6" inches of alignment of each other. In a purely horn-loaded speaker, (a LaScala is horn-loaded from about 110Hz, up and direct radiating below that, but for our purposes here, we'll call it "horn-loaded") there is several feet between the bass, mid and tweeter drivers in a vertical plane. this creates, for lack of better descriptors, a "harshness" that puts many people off of horn speakers. By time and phase aligning the drivers, we get a better sense of space and staging between the speakers, a better sense that the musicians playing are "right there" in the proscenium between and beyond the space, fore and aft, left and right of the speakers themselves. While the sources in and of themselves may sound better without the additional analog to digital conversion in direct radiators, it is well worth the transgression in horn loaded speakers.
Actually, 3-6 inches (7.6-15.2 cm) is a lot of time misalignment, depending on the frequencies/wavelengths involved.
I know Dennis Kleitch (i.e., user "djk" here and on the K-forum, RIP) used to talk about moving a typical tweeter on top of a loudspeaker forward or aft by as little as a half inch (1.3 cm) and the soundstage size and imaging would dramatically change for a listener on-axis with the loudspeaker. That's about 1/4 wavelength of sound at 6.7 kHz--the approximate frequency that tweeters and midranges are crossed in Klipsch Heritage loudspeakers. This is a mental image that's easy for the user to remember.
But if you're talking about midrange--bass alignment at say 400 Hz, you'd have to have more than 8.5 inches (21.5 cm) of physical misalignment of acoustic centers (bass bin--midrange) to hear approximately the same effect.
So direct radiator loudspeakers do suffer from audible time misalignment effects on phase/change of phase (i.e., group delay). But most issues arise in direct radiating loudspeakers not so much because of physical misalignments of the driver housings, but rather to the use of passive "steep slope" crossover filters, which introduce 90 degrees of phase lag--per order of the filters--on the lower frequency drivers.
So for fourth order filters, you get a full wavelength of delay of the lower frequency drivers relative to the higher frequency drivers. (This is why Al K's steep slope crossover filters introduce problems in Klipsch Heritage loudspeakers: they screw up the phase response/time alignment of the loudspeaker even more--for the dubious advantage of reducing the width of the crossover interference bands in the loudspeakers. TANSTAAFL.)
It's just that horn loudspeakers can have much larger time misalignments due to both the physical offsets of the acoustic centers and steeper slope passive crossover filters--that makes time alignment in fully horn loaded loudspeakers a real issue. But once you acoustically measure and correct for acoustic time misalignments by using DSP via higher frequency driver delays (and also avoiding higher order crossover filters--IIR filters, that is), suddenly there is a "magic" that appears in the soundstage image that's difficult to describe.
Chris
I use a Najda DSP X/O pre to do what you describe work by 5 way system.
4 front loaded horns and high efficiency ribbon tweeters.
As standard it is single ended.
Only 8 analogue outputs, but there was an expansion board offered for a short while to make that 10, plus an extra analogue input too.
Analogue volume control, nice DACs implementation, ability to change OPamps easily - I use Burson.
I have a 2nd one that is fully balanced.
Unfortunately Najda dev is no more since 2018😢
So just in case, I too look for something that has 10 outputs, I use my SP-10 rarely these days, so could even live without analogue in.
I time / phase align all drivers Inc the 2.4m long tapped horn subs - so face similar to you.
Also apply PEQ here and there to take room modes.
If my Najdas died I be in fix!
I've never been fully convinced by miniDSP although never owned them.
It's comments by those who have compared them to Najda and other solutions that have put me off - perhaps silly but there we are!
CamillaDSP or possibly Octavia? But seem to be lots of work, when you read the threads.
My experience using Rpi and Volumio for streaming solution was largely positive.
The search goes on.
4 front loaded horns and high efficiency ribbon tweeters.
As standard it is single ended.
Only 8 analogue outputs, but there was an expansion board offered for a short while to make that 10, plus an extra analogue input too.
Analogue volume control, nice DACs implementation, ability to change OPamps easily - I use Burson.
I have a 2nd one that is fully balanced.
Unfortunately Najda dev is no more since 2018😢
So just in case, I too look for something that has 10 outputs, I use my SP-10 rarely these days, so could even live without analogue in.
I time / phase align all drivers Inc the 2.4m long tapped horn subs - so face similar to you.
Also apply PEQ here and there to take room modes.
If my Najdas died I be in fix!
I've never been fully convinced by miniDSP although never owned them.
It's comments by those who have compared them to Najda and other solutions that have put me off - perhaps silly but there we are!
CamillaDSP or possibly Octavia? But seem to be lots of work, when you read the threads.
My experience using Rpi and Volumio for streaming solution was largely positive.
The search goes on.
Im about four pages into your thread about it on the community and it has me rethinking my approach. I set the slopes at 24 db and addressed the eq and it was dramatically better from the ALK universal passive crossover. I then set the time alignment and it too was a noted improvement. Now, after I finish your thread, I suppose I should look at the natural curves of my drivers and set about doing further measurements (much to my s/o‘s chagrin) and adjustments to see where that gets me.
One question that seems relevant, even though we’re not discussing it directly: I set 24db Linkwitz-Riley filters to reduce, if not eliminate driver overlap and without time & phase alignment, it appeared to make the sound more coherent. In taking your approach with first order, unnamed filters, am I to throw concerns about driver overlap out the window? I’m using a JBL D2225H woofer and PRV D2200PH mid driver- the JBL extends to 2kHz before rolling off, while the natural high pass for the PRV D2200PH and WG4550 is around 400Hz. Should I not worry and focus on phase?
No, You still have other choices to make. One is to get the driver/horns closer together to get the center-to-center distance between tweeter, midrange, and bass bins as close as possible (especially true for Jubilees with separate K-402s and KPT-KHJ-LF bass bins), another is to lower the crossover points to a frequency that minimizes polar lobing, but not so low that it causes poor sound quality from the compression driver. Another is to get proper time alignment/phase response through the crossover regions, and then to EQ to attenuate the so-called stop bands of the drivers (instead of using higher order crossover filters). I use something I call "zeroth order" crossover filters in everything but the center MEH: I don't use the canned crossover filters that come with the DSP crossover at all--just PEQs and shelving filters.
By doing these things (especially using PEQs to accomplish crossover duties instead of higher order crossover filters), the net result is much better than just using higher order crossover filters, with its attendant penalty in phase growth. I achieve zero phase growth across the crossover points without using FIR filters (i.e., much more compute intensive digital filters that usually require more expensive DSP crossovers or PCs running in the signal path). I can show you the steps taken and the resulting performance, here: https://community.klipsch.com/index...speakers/page/3/&tab=comments#comment-2379562
The beauty of DSP crossovers is that you can experiment to your heart's content and spend nothing but time. Once you're done, you can recall the last settings that you liked the most and then leave it alone--until you get another brainstorm to try out or learn about something else that you hadn't considered before.
I know that most here have somewhat short attention spans for personal stories, but I think that one might help:
In December 2007, bought a pair of "Professional" (cinema) two-way Jubilees from Klipsch. These didn't have DSP, and they even didn't have DSP settings published openly. I didn't have experience using REW and calibrated microphone then. Roy had another owner send me settings for another DSP amplifier (XTI-1000). Noisy and chattering highs above 12 kHz abounded. But I was "in business".
About 18 months later Roy accessed me to a small, private area of the K-forum ("Just Jubes") that allowed me access to information about DSP crossovers, settings that Roy had initially developed for the 1st-Gen Jubs in the chamber, and other info on alternative HF drivers, etc. I bought a pair of relatively inexpensive higher power DSP amplifiers (Crown XTi-1000s) which turned out to be terrible for the task--but which got me sound from my large investment in loudspeakers (~$7K, new from Hope). I later replaced the XTi-1000s with an EV Dx38 (bought used) and used Crown D-75As (rack-mount control room amplifiers--which turned out to be keepers). The learning curve at this time was steep, but satisfying. I wished that I already knew this stuff, however, so I could apply to the problems that I was hearing.
It took me a year to acquire the TAD drivers (I had a well-paying full-time job then to afford them.) I fiddled with loudspeaker room placement and rudimentary equalizers upstream. Life was good, but I still felt like I wasn't really getting even 50% of what I could out of the loudspeakers.
Fast forward to 2011--when I took the time to fiddle with REW, a small "mixer", and a measurement microphone: an ECM8000, which required phantom power and microphone calibration, which I didn't have at the time. It was a disaster--since I was using the sound card in the laptop in the measurement loop. Nothing worked right. I dropped the effort and just used Roy's settings. By then, I had added DIY Cloned SPUD tapped horn subwoofers behind the Jubs (fc 14.5 Hz). Integration of the subs with the Jubs was by ear and a guess. I acquired a First Watt F3 for the TADs, then I punted--I didn't have the time or energy, holding down a full-time job and an adjunct teaching position (developing all my own curricula from scratch because there were no books on the subject). Time was at a premium.
Fast forward again to 2014: I was "retired" (i.e., looking for the next thing to do but not having to work for a living). I started in earnest working with REW daily and developing a feel and a familiarity for it-experimenting to find the limits of what it and the Dx38 could do. I started to develop my own settings for the Dx38 crossover--and things were a LOT better. I was learning now at about triple or quadruple the pace as before. Things were better, but not really good (sound quality-wise).
I started demastering my music using Audacity in 2015, to fix the sound quality of really dear recordings--that sounded really bad. Whole cloth stuff, again. I met with success and failure--in largely equal doses. (I still demaster all my CDs that come in--as required.) I experimented with MEHs, and developed the K-402-MEH, which exceeded my wildest expectations (and still does, even today). Then I got a Xilica XP crossover and a USB microphone (self calibrated) about two years later. Now sound reproduction started to take on a life of its own. Learning gobs weekly, sharing what I could online (some people on forums don't like to see others learning and advancing in their hi-fi experiences, I think).
And then April of 2019 came around: I discovered the effects of phase flattening (the link I post just above). I was floored. Not only did I have to completely redo all of DSP settings for the entire setup, but I also had to redo all of my demastered music files on external hard drive (about 20,000 tracks) because the phase-flattening discovery caused all of my music to sound like it had 3-6 dB more bass than it did before flattening the phase response.
_________________________________________________________________
Why the long-winded story? I listen to music now, and try to help others get their systems dialed in. Especially without them having to spend thousands of bucks (or tens of thousands, as the case may be now) to achieve relative success--their 99% solutions. So it took me only 15 years to figure out how to do it, and perhaps another few thousand bucks of gear (the software was shareware/freeware at no required cost). That's why I share what I do. I don't want others to have to learn it the hard way--like I did. Fifteen years of effort, starting and stopping, is too long.
You can see that the DSP crossover and REW/calibrated microphone was the ticket to real learning and success--and much deeper understanding. I'm still learning, but not quite at the rate I was over the past 10 years or so. It's still fun.
Chris
Last edited:
Chris, to clarify, in your previous post #77, you said "No, You still have other choices to make. One is to get the driver/horns closer together to get the center-to-center distance between tweeter, midrange, and bass bins as close as possible." You are inferring physical distance so the DSP does not have to correct as much in the digital domain correct?
Also, can you expand on using PEQ and shelving filters for crossover duty?
Also, can you expand on using PEQ and shelving filters for crossover duty?
Think of the extreme case: the drivers are perfectly coaxial (like in a multiple entry horn--MEH). There is no lobing, so the need to ensure minimal driver overlap vanishes--and all drivers' output experiences "unity summation" (just like the name of the patent US6411718) within the horn's aperture. All the drivers couple to each other acoustically and behave as one driver externally.
This effect was made clear to me as I examined the Danley SH-50 and measured its overall output (including polar output) and then measured the individual ways (tweeter, midranges, woofers), and saw the degree of overlap that exists between drivers in this design. That's when the light bulb went off for me, and when I finally understood what Tom Danley was saying in his patent description. MEHs do not need to minimize driver overlap because they experience acoustic summation internally before their output leaves the horn aperture (mouth), and their polar output is constrained by the horn aperture to be consistent throughout the crossover region.
Drivers on a flat baffle will also couple acoustically if their centerlines are all within 1/4 wavelength of the center frequency of crossover. Same effect.
When I moved the K-402 horn downward toward the KPT-KHJ-LF bass bin centerline (about 8 inches from its factory position), not only did the polars get better in the crossover region, but the entire loudspeaker suddenly sounded more "together". Apparently any horn-loaded loudspeaker that can get all its drivers within 1/4 wavelength at their individual crossover frequencies will behave like MEHs. This is one reason why I think crossing over to a tweeter above ~1.2 kHz is a mistake (i.e., the centerlines of the drivers can't exist within 1/4 wavelength at crossover frequency due to physical size constraints)--because it introduces polar lobing between the tweeter and the next lower frequency drivers--which needs to be avoided.
Chris
Last edited:
Read into the thread that I linked to, above (the K-forum thread on Subconscious Auditory Effects of Quasi-Linear Phase Loudspeakers) starting at page 3. The sequence of steps to do it can be found here: https://community.klipsch.com/index...ar-phase-loudspeakers/page/4/#comment-2388972