Martagon stereo configurations & applications
Hi (DaveCan),
The 'Martagon' is implemented (inspired by the Paragon of JBL) as a low china closet ('dressoir') by combining two full range enclosures into a left-right 'stereo' set (with the amplifiers in the middle) as located in private homes/rooms.
The directivity is 30 x 100 degrees (vertical x horizontal); this gives a horizontally wide stereo image when sitting comfortably on a (low) couch at a distance from 2 meter on for full enjoyment (obviously, the room should not be too smal f.i. > 40 m**2)
Several sizes 'Super size', 'Full size', ( and 'Compact' and 'Sub-compact' in the future) are available to accommodate available physical space and taste.
If the enclosures woul be tillted by 90 degrees as you suggested, you would have wide dispersion in the vertical plane while narrowing in the horizontal plane of course). I would not advise this for single in-home stereo application.
Having said this, thanks to the high sensitivity, Martagon's could be stacked for larger venues for live music and such. I will post some configuration examples.
The 'Supersize' measures 1.2 x 0.6 x 0.6 meters, the 'FullSize' 0.9 x 0.6 x 0.6 (and the future 'Compact' 0.85 x 0.6 x 0.5).
Hi (DaveCan),
The 'Martagon' is implemented (inspired by the Paragon of JBL) as a low china closet ('dressoir') by combining two full range enclosures into a left-right 'stereo' set (with the amplifiers in the middle) as located in private homes/rooms.
The directivity is 30 x 100 degrees (vertical x horizontal); this gives a horizontally wide stereo image when sitting comfortably on a (low) couch at a distance from 2 meter on for full enjoyment (obviously, the room should not be too smal f.i. > 40 m**2)
Several sizes 'Super size', 'Full size', ( and 'Compact' and 'Sub-compact' in the future) are available to accommodate available physical space and taste.
If the enclosures woul be tillted by 90 degrees as you suggested, you would have wide dispersion in the vertical plane while narrowing in the horizontal plane of course). I would not advise this for single in-home stereo application.
Having said this, thanks to the high sensitivity, Martagon's could be stacked for larger venues for live music and such. I will post some configuration examples.
The 'Supersize' measures 1.2 x 0.6 x 0.6 meters, the 'FullSize' 0.9 x 0.6 x 0.6 (and the future 'Compact' 0.85 x 0.6 x 0.5).
An externally hosted image should be here but it was not working when we last tested it.
Martagon articulated sound
Well, how do they sound?
- 1) Thanks to the point-source approach and the very wide bandwidth, stereo imaging is perfect if tested through official recordings (tracks 5 - 11 of CD1 'Checkpoint Audio' Eelco Grimm)
- 2) Martagon has been designed by optimization of the 'Impulse Response' (Arta). Because front-loaded horns in fact are acoustics transformers, horns show minimal 'ringing' upon step response (bass-kick-drum, percussion, cimbals but also sharp spikes as vocal pitch). So there is minimized 'smearing-over-time': as a result, drum bass-kick and bass(guitar) completely 'separate-out' and are perceived independently. Percussion is like playing the drums yourself, so with crazingly high dynamics and the feeling that there is unlimited overhead. Vocals sound is if your ear is the microphone: sounds like-in-heaven on quality recordings (Diana Kroll, ...) but can be revealing/disappointing on poor recorings that are compressed or contain 'repaired' bits. Dance beat sound totally 'dry' and 'fast' with sharp beginning & end (and no bouncing).
The below picture shows the frequency transfer functions and impulse response for 2 sealed bass enclosure (50 ltr and 100 ltr) a vented box of 150 ltr, a 6th order bandpass of 250 ltr and an optimized front-loaded horn of 300 ltr. The optimally damped Impulse Response of the horn pumps out its energy over the shortest possible period of time at an elevated SPL level (integral over time of a 6 order bandpass is comparable but this Impulse response 'oscillates' 3 times longer, smearing-out short pulses over time totally).
By-The-Way: the mid section of Martagon is also front-loaded horn and offers the same shortest possible impulse responses; direct radiating speakers always show artificial 'smearing-over-time'; unfortunately we are used to it..... until you here THIS: LF, MF and HF horns firing 'in-sync'!...
Note: rear-loaded LF horns: ALWAYS trouble with interphering frequency components coming from the direct radiating speaker and delayed from the rear-horn.....
Well, how do they sound?
- 1) Thanks to the point-source approach and the very wide bandwidth, stereo imaging is perfect if tested through official recordings (tracks 5 - 11 of CD1 'Checkpoint Audio' Eelco Grimm)
- 2) Martagon has been designed by optimization of the 'Impulse Response' (Arta). Because front-loaded horns in fact are acoustics transformers, horns show minimal 'ringing' upon step response (bass-kick-drum, percussion, cimbals but also sharp spikes as vocal pitch). So there is minimized 'smearing-over-time': as a result, drum bass-kick and bass(guitar) completely 'separate-out' and are perceived independently. Percussion is like playing the drums yourself, so with crazingly high dynamics and the feeling that there is unlimited overhead. Vocals sound is if your ear is the microphone: sounds like-in-heaven on quality recordings (Diana Kroll, ...) but can be revealing/disappointing on poor recorings that are compressed or contain 'repaired' bits. Dance beat sound totally 'dry' and 'fast' with sharp beginning & end (and no bouncing).
The below picture shows the frequency transfer functions and impulse response for 2 sealed bass enclosure (50 ltr and 100 ltr) a vented box of 150 ltr, a 6th order bandpass of 250 ltr and an optimized front-loaded horn of 300 ltr. The optimally damped Impulse Response of the horn pumps out its energy over the shortest possible period of time at an elevated SPL level (integral over time of a 6 order bandpass is comparable but this Impulse response 'oscillates' 3 times longer, smearing-out short pulses over time totally).
By-The-Way: the mid section of Martagon is also front-loaded horn and offers the same shortest possible impulse responses; direct radiating speakers always show artificial 'smearing-over-time'; unfortunately we are used to it..... until you here THIS: LF, MF and HF horns firing 'in-sync'!...
Note: rear-loaded LF horns: ALWAYS trouble with interphering frequency components coming from the direct radiating speaker and delayed from the rear-horn.....
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
Thank you for your explanation, I understand now about the dispersion. I really like them and they look cool! I love stuff that isn't the same old take on a box with sliced angles on it to make them look modern, and with all the usual port(s) etc that apply.. The audio forums are full of stuff like that, and many to be fair are great.. Your take on the Paragon is really neato and top notch!
Will plans be shared at some point and do you have pics of the build process?
Will plans be shared at some point and do you have pics of the build process?
Last edited:
Thanks Wim for the photo and graphs. Those are very impressive!
A couple of questions - does the bass pass thru the mid horn, too? It's hard to see if the back of the mid horn is open or closed.
And there seems to be a smaller, partial horn inside the mid horn. What is the purpose of this part?
Thanks
A couple of questions - does the bass pass thru the mid horn, too? It's hard to see if the back of the mid horn is open or closed.
And there seems to be a smaller, partial horn inside the mid horn. What is the purpose of this part?
Thanks
Interesting. As big as they look and seem, the Supersize is just slightly smaller than my Altec VOTT 838 bass cabinet on its side. Not so bad.
Martagon
Hi All,
to respond to your questions:
- 1) Mid/High in center: why not at top?
Concept is inspired by Paragon with its wide spheric central difractor.
With the mid/high in the middle, you can turn/direct the HF horn towards the curved pane
and gain some extra dispersion both vertically and horizontally
(of course, this is not compatible with source-point concept anymore...).
Putting the Mid/High at the top: no problem, but then this one degree of freedom is gone.
- 2) Acces LF driver:
SuperSize: via back-sub-panel
FullSize: via side panel
- 3) Asymetric shape of Mid-horn
The folde-LF-horn is folded as a garden hose (so no 180 degree turns).
Due to this, the mouth of the LF is not in the center of the cabinet and
therefore the physical integration of the Mid-horn has yielded this asymmetric shape
- 4) Partial horn in mid-horn: diffusor
Mid horn is one constant dimension; to prevent standing waves, a diffuser is used.
More on this to follow (see chapter 5.3.4.2).
- 5) Build process: see below
- 6) Drawings shared at some point... see below
AT LAST (for today)
Please find attached 5 capita selecta of a formal design report I created some time ago; about my intentions: I will come back to this shortly
Hi All,
to respond to your questions:
- 1) Mid/High in center: why not at top?
Concept is inspired by Paragon with its wide spheric central difractor.
With the mid/high in the middle, you can turn/direct the HF horn towards the curved pane
and gain some extra dispersion both vertically and horizontally
(of course, this is not compatible with source-point concept anymore...).
Putting the Mid/High at the top: no problem, but then this one degree of freedom is gone.
- 2) Acces LF driver:
SuperSize: via back-sub-panel
FullSize: via side panel
- 3) Asymetric shape of Mid-horn
The folde-LF-horn is folded as a garden hose (so no 180 degree turns).
Due to this, the mouth of the LF is not in the center of the cabinet and
therefore the physical integration of the Mid-horn has yielded this asymmetric shape
- 4) Partial horn in mid-horn: diffusor
Mid horn is one constant dimension; to prevent standing waves, a diffuser is used.
More on this to follow (see chapter 5.3.4.2).
- 5) Build process: see below
- 6) Drawings shared at some point... see below
AT LAST (for today)
Please find attached 5 capita selecta of a formal design report I created some time ago; about my intentions: I will come back to this shortly
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
Looking at 1.1 in document 4 and time alignment. Could it be true that your delays are related to the physical path lengths only but that there is no compensation for the delays created by the digital filter themselves which become greater as the frequency decreases. The final step response does not seem to really align. Bass is still behind mid/hf (mid/hf seem indeed aligned.)?
//
//
Martagon new build-picture
Hi All, I notice that the 'build'pic' of the 'SuperSize' arrived heavily distorted, so I re-send it in landscape now>
More on your questions to follow later...
Martagon Pictures, Martagon Images, Martagon Photos, Martagon Videos - Image - TinyPic - Gratis foto's delen en afbeeldingen & video's hosten
Hi All, I notice that the 'build'pic' of the 'SuperSize' arrived heavily distorted, so I re-send it in landscape now>
More on your questions to follow later...
Martagon Pictures, Martagon Images, Martagon Photos, Martagon Videos - Image - TinyPic - Gratis foto's delen en afbeeldingen & video's hosten
Martagon
Hi All: responses to your outstanding question.
- 3 Octave rule?
No. Pumping out acoustic energy equally distributed over low/mid/high is based on the assumption of 1 / f noise; the slope of 'x' / f then dictates if it is 'y-octave' per section.
The concept here is on music and optimized stereo-imaging.
The base-frequency of 'lead'-'voice' in music (vocal, brass, violin, cello...) is roughly from 150 to 3000 Hz.
It is definitivey not optimal to split this CRUCIAL range because of phase en amplitude ripple in the transition (sub-range). Capturing the most of 'lead-voice' energy by one section/mid-speaker is the adopted principle here. On top, crucial is to have linear phase in the transition frequency ranges between low and mid speaker and mid - high. More on this later.
Also, if speakers can handle the acoustic power, it is better to have 6 or 12 dB attenuation in the stop-band because higher order filters will induce extra 'ringing' and exactly the pro of horns is that they deliver minimized ringing themselves by nature.
- Time alignment? Bass still behind mid/high?
If so, it is minimal: look at the perfect linear phase in the transition range between low - mid Capita Selecta 5, chapter 1.2.2
BACKGROUNDS
Impulse responses of band-limited systems do show Frequency Modulation over their time course: see Capita Selecta 2 page 5 and 6.
If we look at the 2.1.3 picture, at the very left (beginning) the sharpness of the upgoing slope is determined by the low-pass (at 25 kHz in the example). At the very right, the last slow component, is determined by the high-pass at 28 Hz in this example; in fact the IR is a very rapidly changing FM chirp. This all obeys fundamental physical law for 'causal' systems.
If we spilt the spectrum on 3 parts, the 3 corresponding impulse responses will more clearly reveal 'FM'- chirping because the upper- and lower frequencies are closer to each other.
The below picture shows the authentic Impulse(Step) Response for the DSP processed SuperSize LF unit; so physical horn lenght and filtering summed-up.
The 't = 0' is given by the short chirp at the left (from the HF unit; at the vertical yellow line), measured at the mouth of the system (so '0' meter).
The LF response shows a 5 (6) phasic behaviour: recall that any audio system (other than closed box) acts as a 3-order high-pass at the low-frequency side, and as a 2nd order at the high-frequency side.
The zoom is 1:8 (because of long duration)
The second picture is zoom 1:2 to show both the first part of the LF and the full response of the MID. Now, alignment means that the full MID responses is shifted to the very first negative component of the LF; in the TOTAL response, this first negative component of the LF is 'obscured': I guess this caused the confusion.
At last: Back-loaded horns: How do you think that phase between direct-radiated frequencies and delayed components coming from the rear-horn will look for back-loaded horns...? It is totally wild....
- 10 SPL level?
Indeed, isolated measurement on the LF unit were done at '0' meter. Other measurement were done in the outside air at 2 meter typically; the transducer cabinets were tilted by 45 degrees to somewhat reduce the influence of low frequencies bouncing back from floor.
All measurement were done via Arta and are NOT calibrated.
However..... the Fostex claims a 105 db at 1 Watt and 1 meter: I had to attenuate the HF Fostex by -2 dB to have the same SPL level as the MID and to elevate the LOW units by 2 dB. This brought me to the provisional conclusion that the LF has sensitivity of 101 dB (28 - 150 Hz) and the MID of 103 dB.
This is in agreement with the number of dB sensitivity you can gain with decent front-horn loading.
The ration of mouth/throat for the SuperSize horn for instance is 9.8; a factor '8' would give 3 + 3 + 3 dB.
FURTHER....
It is know that 2 speakers give you + 3 dB gain. But, if the speakers couple acoustically, that is if the distance between them is (substantially) smaller than the wave length of the sound, then the gain is (3 + 3) dB.
In other words: if you place 2 Martagons 1 or 2 meters apart (which is typical), for the range from 28 to 100 on they summ up a nice 6 dB (if LF is identical on both left and right which is often the case in contemporary music) ).
This argument acoustic coupling is even more true if you would 'stack' several Martagons: not only the 6 dB gain holds, but also the mouth-size increases, so the 'hybrid' characters shifts towards more horn than transmission line, so in fact the middle frequency of the physical horn shifts downwards (to be compensated via the DSP).
- Plywood
Indeed, I used layering starting from 3 mm plywood.
HOWEVER: this project was also an aesthetic challenge; there is NO reason why the curved panes could not be replaced by 2 or 3 'straight' panels; the wavelength of the target frequencies is that long they do do not 'see' minor deviations from a smooth curvature.
Hi All: responses to your outstanding question.
- 3 Octave rule?
No. Pumping out acoustic energy equally distributed over low/mid/high is based on the assumption of 1 / f noise; the slope of 'x' / f then dictates if it is 'y-octave' per section.
The concept here is on music and optimized stereo-imaging.
The base-frequency of 'lead'-'voice' in music (vocal, brass, violin, cello...) is roughly from 150 to 3000 Hz.
It is definitivey not optimal to split this CRUCIAL range because of phase en amplitude ripple in the transition (sub-range). Capturing the most of 'lead-voice' energy by one section/mid-speaker is the adopted principle here. On top, crucial is to have linear phase in the transition frequency ranges between low and mid speaker and mid - high. More on this later.
Also, if speakers can handle the acoustic power, it is better to have 6 or 12 dB attenuation in the stop-band because higher order filters will induce extra 'ringing' and exactly the pro of horns is that they deliver minimized ringing themselves by nature.
- Time alignment? Bass still behind mid/high?
If so, it is minimal: look at the perfect linear phase in the transition range between low - mid Capita Selecta 5, chapter 1.2.2
BACKGROUNDS
Impulse responses of band-limited systems do show Frequency Modulation over their time course: see Capita Selecta 2 page 5 and 6.
If we look at the 2.1.3 picture, at the very left (beginning) the sharpness of the upgoing slope is determined by the low-pass (at 25 kHz in the example). At the very right, the last slow component, is determined by the high-pass at 28 Hz in this example; in fact the IR is a very rapidly changing FM chirp. This all obeys fundamental physical law for 'causal' systems.
If we spilt the spectrum on 3 parts, the 3 corresponding impulse responses will more clearly reveal 'FM'- chirping because the upper- and lower frequencies are closer to each other.
The below picture shows the authentic Impulse(Step) Response for the DSP processed SuperSize LF unit; so physical horn lenght and filtering summed-up.
The 't = 0' is given by the short chirp at the left (from the HF unit; at the vertical yellow line), measured at the mouth of the system (so '0' meter).
The LF response shows a 5 (6) phasic behaviour: recall that any audio system (other than closed box) acts as a 3-order high-pass at the low-frequency side, and as a 2nd order at the high-frequency side.
The zoom is 1:8 (because of long duration)
The second picture is zoom 1:2 to show both the first part of the LF and the full response of the MID. Now, alignment means that the full MID responses is shifted to the very first negative component of the LF; in the TOTAL response, this first negative component of the LF is 'obscured': I guess this caused the confusion.
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
At last: Back-loaded horns: How do you think that phase between direct-radiated frequencies and delayed components coming from the rear-horn will look for back-loaded horns...? It is totally wild....
- 10 SPL level?
Indeed, isolated measurement on the LF unit were done at '0' meter. Other measurement were done in the outside air at 2 meter typically; the transducer cabinets were tilted by 45 degrees to somewhat reduce the influence of low frequencies bouncing back from floor.
All measurement were done via Arta and are NOT calibrated.
However..... the Fostex claims a 105 db at 1 Watt and 1 meter: I had to attenuate the HF Fostex by -2 dB to have the same SPL level as the MID and to elevate the LOW units by 2 dB. This brought me to the provisional conclusion that the LF has sensitivity of 101 dB (28 - 150 Hz) and the MID of 103 dB.
This is in agreement with the number of dB sensitivity you can gain with decent front-horn loading.
The ration of mouth/throat for the SuperSize horn for instance is 9.8; a factor '8' would give 3 + 3 + 3 dB.
FURTHER....
It is know that 2 speakers give you + 3 dB gain. But, if the speakers couple acoustically, that is if the distance between them is (substantially) smaller than the wave length of the sound, then the gain is (3 + 3) dB.
In other words: if you place 2 Martagons 1 or 2 meters apart (which is typical), for the range from 28 to 100 on they summ up a nice 6 dB (if LF is identical on both left and right which is often the case in contemporary music) ).
This argument acoustic coupling is even more true if you would 'stack' several Martagons: not only the 6 dB gain holds, but also the mouth-size increases, so the 'hybrid' characters shifts towards more horn than transmission line, so in fact the middle frequency of the physical horn shifts downwards (to be compensated via the DSP).
- Plywood
Indeed, I used layering starting from 3 mm plywood.
HOWEVER: this project was also an aesthetic challenge; there is NO reason why the curved panes could not be replaced by 2 or 3 'straight' panels; the wavelength of the target frequencies is that long they do do not 'see' minor deviations from a smooth curvature.
Thanks for that. I wondered if you had to layer the plywood curve. Must have been a lot of work! But the appearance is worth it. 
The timing and phase issues are fascinating. When I first started using the DCX as a crossover for horn systems, I spend countless hours virtually moving drivers back and forth and looking at the impulse response. I never really came to a hard and fast rule or conclusion, just used what summed the best and gave a benign phase plot. Looks like you have also concentrated on phase.
The timing and phase issues are fascinating. When I first started using the DCX as a crossover for horn systems, I spend countless hours virtually moving drivers back and forth and looking at the impulse response. I never really came to a hard and fast rule or conclusion, just used what summed the best and gave a benign phase plot. Looks like you have also concentrated on phase.
No, either it's loaded by the horn or it's not. In free space 30 Hz? Maybe with a direct radiator in that size box but probably not at 1o1 db with 2.83 volts.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.