Arithmetic vs Diffractal - first comparison

This is a comparison of Arithmetic Diffuse Signature and Elite vs RPG Diffractals. With four units with dimension 120x60 cm. One RPG Diffractal is reversed in order to follow the barker sequence. RPG Diffractals diffuse well down to 450-500 Hz, while the Arithmetic Singnature and Elite diffuses all the way down to 100 Hz according to meausuerements. The Signature diffuses up to 8 kHz area and the Elite show measurement up to 10 kHz. Diffractals diffuse as high as 20 kHz. The Diffractals are 23 cm deep, while the Arithmetic Signature are only 2.5 cm in depth. The depth of Arithmetic Elite will depend on the stand offs used for the high frequency units.


The comparison is not meant to me a scientific study by any means. This is simply a comparison with listening and some measurements presented. Comparing diffusers like this is not a easy task because of the time it takes to switch. Considering the auditoria memory is short, this obviously is not ideal. Areas like speaker design, frequency response, the overall acoustics, music material, taste, number of diffuser, etc. will also come into play. Meaning that the result here isn't objective, but primarily a subjective experience in this specific setup.


Space
The room that was used is far from ideal to compare diffusers. Normally the speakers here are placed on the long side of the room with diffusion at far side wall reflections and front wall (energy arriving from a close rear wall).
VA Sagarmatha stereo2 (Medium).jpg


However, in order to compare to diffuser it's better to place them on the rear wall. But in this case the sloping ceiling behind listening position creates specular reflections, hinders energy towards the diffusers and there's one side wall with windows and frames that also prevents energy. Hopefully we can present more comparisons like later in an improved environment.

Setup
The speakers that were used here are a prototype horn speaker from Vera Audio together with a dual 18" subwoofer. It's active speaker design with external DSP and Vera Audio P400/1000 power amps. The speakers have a constant directivity of 80x50°, meaning a narrow directivity that will focus more energy towards the diffusers. No EQ of the room ("room correction") was applied.

Vera Audio Sagarmatha and subwoofer proto (Medium).jpg


Placement on the short wall wasn't given much attention, but speakers like these tend to measure very even almost anywhere they are placed and the frequency response was considered good enough and subwoofer (seen in the middle) was time aligned with crossover at 110 Hz.

On the rear side wall two dividing "walls" were setup in order to resemble a flat reflective surface better. This increased a dip in the 160 Hz area, but considering it had low Q, it shouldn't matter much.
IMG20230124170144 (Medium).jpg


Frequency response of each speaker is shown below with 1/12 Oct. smoothing with the Diffractals. The measurement were done with the listening chair in place and it's worth mentioning that the chair effects the frequency area between approximately 500 Hz and 900 Hz. Removing the chair flattens out this area nicely, which should be closer to the response when a person is sitting in the chair.
Left freq_Diffractal.jpg Right freq_Diffractal.jpg

Comparison method
The first comparison was done by simply placing the Arithmetic Signature directly in front of the Diffractals on top of the pillows that are filled porous material. This was the fastets way to switch between them and the Arithmetics were simply placed on the right side in the picture when listening with the Diffractals. Later Arithmetic Elite were compared to the Diffractals with same placement and different placements (explained later) were also tested.
Distance from ears to the diffusers was minimum 2.60m in this first comparison.


Arithmetic Signature placed in front of Diffractals. Some absorption material is behind the Diffractals.
IMG20230124165036 (Medium).jpg


Listening was conducted by both short durations of 15 seconds and longer durations of 30 to 60 seconds. The swiching happened many times and over a period of a few days and with various music material of both good and low quality.

While some will claim that listening impression are useless without an ABX test, we think by listening over several days with various music materials and many switchings greatly minimizes the chance of placebo. Also, the difference while not being huge was clearly different, thus a blind test didn't seem necessary here to prove a difference. And for preferences, an ABX blind test with short duration may not be the better method as it's important to listen to various music material over a bit of time.


Measurements
Frequency response didn't really change much with either type. What's more interesting is looking at the ETC of each speaker.
Here's the ETC of the left speaker with Diffractals:

Left ETC_Diffractal_VA Sagarmatha horn speakers.jpg


Same channel with Arithmetic Signatures:
Left ETC_Arithmetic Signature in front of Diffractal_VA Sagarmatha horn speakers.jpg


Right speaker with Diffractals:
Right ETC_Diffractal_VA Sagarmatha horn speakers.jpg


Right channel with Arithmetic Signatures:
Right ETC_Arithmetic Signature in front of Diffractal_VA Sagarmatha horn speakers.jpg



The high specular reflection we're seeing right before 8 ms is from the sloping ceiling. Ideally it should have been attenuated with absorption for this comparison, but this wasn't done. There are also some other specular energy with a little high gain that's arriving from the same ceiling.
The area we need to compare here is what's starting right before 14 ms and up to about 18/19 ms. With both the left and right speaker we can see that with Arithmetics there's a spike at around 14 ms that has risen compared to Diffractals. We also see that the area later is a but more supressed with the Arithmetics. One the left channel the comb filtering is better packed and less sparsed with the Diffractals but with the left channel it's the opposite.

Listening impression with Diffractals vs Arithmetic Signature
It can be difficult to describe differences we hear and we might be using different language to convey what we mean. But I would describe the difference the following way:
With the Diffractals there's a hint of warmth added to the sound and the higher frequencies seem to sound a tad more washed out or more diffused. This created a more comfortable result and perhaps a bit more spacious sound field. With the Arithmetic Signature there was an immediate experience of more energy which also perhaps gave slightly better clarity and intelligbility for the vocal. I believe this difference is related to the higher specular gain at around 14 ms which gives something closer to a Haas kicker. But the signature was definetly colder with the Arithmetics compared to the Diffractals. For most music material I preferred using the Diffractals with the added warmth and which sounded smoother. It's possible that besides the higher gain specular reflection, the difference is also a result that the Diffractals with tiny wells diffuse the higher frequencies better. Wells with depth creating more temporal diffusion is another possiblity.

See next page for further testing with Arithmetic Elites.
 
Last edited:
Next up was comparing Arithmetic Elite to Diffractals. The Elites have two added high frequency planks in front the of the Signature. However, these are not covering the whole unit but only smaller parts.

Also here we started out simply by placing the Elites in front of the Diffractals. Swicthing between the units took less than a minute.
Elitejpg (Medium).jpg

Elite close up (Medium).jpg


Measurements
Take note that these new measurements can't be compared to the previous ones in post 1. The reason is that the mic isn't in the exact position. We can only compare the measurements presented together where the mic was in the exact same position.

Left channel with Diffractals:
Left ETC_Diffractal_VA Sagarmatha horn speakers2_vs Elite.jpg


Left channel with Arithmetic Elites:
Left ETC_Arithmetic Elite in front of Diffractal_VA Sagarmatha horn speakers2.jpg


Righ channel with Diffractals:
Right ETC_Diffractal_VA Sagarmatha horn speakers2_vs Elite.jpg


Right channels with Arithmetic Elites:
Right ETC_Arithmetic Elite in front of Diffractal_VA Sagarmatha horn speakers2.jpg


We're seing a similar result as before. At 14 ms or slightly earlier a reflection with the Arithmetics has higher gain, especially with the left speaker.

Listening impression with Diffractals vs Arithmetic Elites
It's possible that the gap in regards to smoothness was minimized some by changing to the Elites, but difficult to say for sure and would really have to compare the Signature to the Elite to confirm this. The difference with the high frequency planks certainly wasn't big. The experience with a warmer and more comfortable listening experience with the Diffractals and a bit more energy and colder presentation with Arithmetics Elites was still present.

If an average of the frequency respons of both channels are used with 1/3 smoothing, the level from the lows to 15 kHz has a fall with 6 dB.
324177721_3441847706133775_3736752790816423120_n.jpg


The Bruel & Kjaear response has a drop here of 7.5 dB and the Harman preference curve has 9 dB. However, these horn speakers have a constant directivty much lower in frequency compared to speakers used in those tests and also have a narrower directivity. They tend to sound darker than typical speakers. Despite of that, we also tried to add 1 dB in the lows to get closer to the B&K response, but it didn't change the experience.

See next post for a comparison with absorption.
 
Last edited:
What about absorption on the rear wall?
We also tested absorption by placing three 20 cm absorbers and one 10 cm deep absorber in front of the Diffractals. These also have a frame on the rear that adds some air gap.
Absorption (Medium).jpg


Not suprisingly, the reflections attenuates further with absorbers.


Left channel with Diffractals:
Left_ETC_Diffractal.jpg


Left channels with absorbers:
Left_ETC_absorbers.jpg


Right speaker with Diffractals:
Right_ETC_Diffractal.jpg


Right speaker with absorbers:
Right_ETC_Absorbers.jpg


Listening impression with absorbers vs diffusers
While the measurement above is only showing Diffractals, the comparison was also conducted some with Arithmetic Elite vs absorption. In both comparisons, it's evident that absorption isn't psycho-acoustical preferable with this distance vs diffusion. Initially it sounds strange where the brain is missing something that should be there. We are used to listen to a wave of sound from the rear wall and it's strange and unatural when it's not there. Another major difference is spaciousness and envelopment. This is largerly gone with absorption. The result is clearly a less engaging and enjoyable listening experience.

Something to point out here is that with absorption we loose some level, especially above 1 kHz. Maybe this should have been compensated for with the DSP, but it's doubtful it will make much of the difference due to a lack of natural rear wave and the great benefit of a larger sound field that diffusers create.

Left channel frequency response Diffractals vs absorbers. 1/12 oct. smoothing.
Left channel freq overlay.jpg


Righ channel frequency response Diffractals vs absorbers. 1/12 oct. smoothing.
Right channel freq overlay.jpg


See next post with diffusers vs a reflective wall.
 

Attachments

  • Right channel freq overlay.jpg
    Right channel freq overlay.jpg
    392.8 KB · Views: 68
Last edited:
Arithmetic vs reflective wall
Testing this was not only important in regards a purely reflective wall, but also in order to test Arithmetic Diffuse Elite against a wall instead of in front of Diffractals. The Arithmetic design is after all based on diffusion both from the panel and the reflective surface behind the panel. The holes in the panels also means the unit can be placed in front of an absorber/bass trap and even in front of woofers in a DBA setup. We have earlier measured the unit in front of subwoofer and saw that the response didn't change below 300 Hz as a side note.
307072841_1148058065799768_8541071749169738115_n (Medium).jpg 307123325_1504391500003408_8084328764173666324_n (Liten).jpg

Here's the reflective rear wall. Distance to listening position a little over 3m.
reflective wall (Medium).jpg


Arithmetic Elite placed against the wall:
Arithmetic Elite up against wall (Medium).jpg


Listening experience reflective wall vs Arithmetic Diffuse Elite
In some ways a reflective wall with a distance of 3 m is actually better than a totally absorptive wall because you have more energy and ambiance. But after a while when listening to various material you start to hear discrete high gain reflections the alters the tonality. Very typically listening fatigue kicks in much earlier. Adding the Arithmetic Elites was greatly preferable with a smoother presentation and much more spacious sound stage.

Measurements
Let's move on to measurements of both to see the difference. Here it's interesting to look at frequency response too considering the Arithmetic are suppose to work down to 100 Hz area.

Frequency response left channel with reflective wall:
Left freq_reflective rear wall_VA Sagarmatha horn speakers.jpg


Left channel with Aritmetic Elites:
Left freq_Arithmetic Elite on rear wall_VA Sagarmatha horn speakers.jpg


Right channel with a reflective wall:
Right freq_reflective rear wall_VA Sagarmatha horn speakers.jpg


Right channel with Arithmetic Elites:
Right freq_Arithmetic Elite on rear wall_VA Sagarmatha horn speakers.jpg



Comment:
In terms of frequency response we don't see much difference. This is interesting considering the Arithmetic are suppose to work to 100 Hz. An explanation for this could be due to the fact the measurements of the Arithmetic are conducted in a large room with a diffuse sound field and something doesn't exist in a small room like here. Another cause is that we aren' covering enough surface area. However, with absorption of similar dimension we see the response changes more.

See next post for ETCs and further comparisons.
 
Last edited:
ETCs with reflective wall vs Arithmetic Elites

Left channel with a reflective rear wall
Left ETC_reflecitve rear wall_VA Sagarmatha horn speakers.jpg



Left channel with Arithmetic Elites:
Left ETC_Arithmetic Elite on rear wall_VA Sagarmatha horn speakers.jpg


Right channel with reflective wall:
Right ETC_reflective rear wall_VA Sagarmatha horn speakers.jpg


Right channel with Arithmetic Elites:
Right ETC_Arithmetic Elite on rear wall_VA Sagarmatha horn speakers.jpg


We can see that Arithmetics brings down some high gain reflections. Some are increased slightly in level, though they are quite low in level.

Diffractal vs Arithmetic Elite flushed to a reflective wall
A short comparison was also done with both types flush to the rear wall. However, here we had to remove the units that were not tested from the room. This took time and AB comparison was much more difficult to do. But all in all, the previous comparisons between the units seemed to remain.
Arithmetic Elite up against wall on books (Medium).jpg Diffractals up against wall on books (Medium).jpg

A comparison by turning the setup 180° was also conducted. Here the comparison was also conducted by simply placing the Arithmetic Elites in front of the RPG Diffractals. Listening impression were the same here as before.
setup turned 180  deg (Medium).jpg Elites 180 deg (Medium).jpg


Conclusion
In this specific setup all the diffusers worked very well. They were highly preferable over either a reflective rear wall or an absorptive rear wall. The RPG Diffractals were preferred over the Arithmetics. Whether a different environment and/or a different frequency response could have changed this remains a question. Diffusion (disregarding absorption placed behind for lows) lower in frequency than at 450/500 Hz doesn't seem to be important for a real wall with good distance, if we can trust the measurements of the Arithemetic by NWAA labs to be accurate in this area. This is quite contrary to early arriving reflections where treatment down to the Schroeder frequency is very important.
A great benefit with the Arithmetic is how shallow the units are compared to QRD type of diffusers.
 
Last edited:
Top