Final audiophile tweak - Convolution with Ambiophonics

I have now compiled all the video into a playlist and will be gradually adding more video of different settings, cables and file format. My samples here are mostly Amazon MP3 format. The Mad World is CD (16/44,1). But I think all of them would probably sound the same due to Youtube lossy format. All my files are either upsample or downsampled to 96000kHz in JRiver. This includes DSD files.Mytek settings - No upsampling and slow filter. Volume in digital mode.

https://www.youtube.com/playlist?list=PLdMpXpnMH1M8rI-qlCRLZsPALIe8Ja8zA

To those still having two hearts about trying Ambiophioncs, I hope the email below would allay your fear.

I was copied an email by Ralph of Ambiophonics from a member of https://www.njaudiosociety.com/ .  There was also another link of someone comparing Michael Fremer of Stereophile system in the AV forum which I will post later.

Meanwhile here is the email...

Dear Ralph,

It was a pleasure meeting you, your lovely wife, and your employee, Dave. I had one of the most magical experiences of my life!!

Thank you so much, Ralph, for your open invitation into your beautiful home and for paying attention to me, taking me through the listening experience, yesterday, Saturday, 05/21/2016, allowing me to experience your Ambiophonic Virtual Sonic Reality system. Even though the members of the NJ Audio Society spoke amazingly of your revolutionary sound system, I was not prepared for what I had experienced in your unique listening room. It was no doubt the listening experience of my life!

Thank you again, Ralph. Truly!

Warmest regards,
Steve Wilson

In a recent post in PSAudio Pauls Post, one experienced person commented on Ambiophonics but from his discription he setup the speakers wrongly. This misunderstanding is too common to the extent the actual demo would give negative or no effect at all with Ambiophonics. 

To address this confusion, I made this YouTube  where you don't need to do anything other than moving your sitting position OR the speakers. 

Enjoy.....

https://youtu.be/8GHxVX6Cbys

Same setup but now driven by Class D Amp. The is a slight change in the tonal characteristics which I believe due to changing the Xover settings. I am more concerned with the HF for now. Which one would be your preferance?

Class D Amplifier




or

Class AB Amplifier

I have now implemented an additional of four speakers for convolution. My initial plan was to implement 25 convolution speakers using the Impulse Response of Sydney Opera House. The idea is to recreate the actual space of the concert hall in my room.

Having said that, the proper implementation  require some knowledge to work with pro sound software and equipment which is out of depth for me at the moment. So what I did was to create 4 "walls" using a single Impulse response of Sydney House recorded at 45 degrees to the left and 23 degree above the listener head sitting at the sweetspot.

I used the convolution setting in JRiver to send the signal to 4 different speakers place at 45 degrees and another one at 130 degrees. Each pair volume is controlled for a proper mixture of convolution. This zone two which is linked to the main speakers' zone. Since the convolution level is very low and the fidelity is not important, I use A960 Line5 digital Amp. At 25W per channel it is more than enough to drive small micro speakers.

Here is the chart of my setup.

{See https://www.hifi4sale.net/t28550p100-final-audiophile-tweak-convolution-with-ambiophonics#150547 }

The best sound production is live unamplified sound in a concert hall. This is a distance dream no matter how much money you spend to get the world's best system. Musical enjoyment is not so much about fidelity - like - getting higher resolutions but more of the ambiance the music is played. That's why many prefers to sing in the bathroom because the natural reverb create a sense of auditory satisfaction. Studies shows that strong ambiance reverberation increases higher electrical activities in us.

Even in a concert hall, the sound may not be precise and clinical. There probably no pin point accuracy but we tend to like the long decay of the concert hall. The actual preference is still subjective as some may like clarity while others may have other criteria.

To create this ambiance in my room, I now use additional speakers to act like the wall of the concert hall. [url=The proper method is here]http://www.ambiophonics.org/HomeTheater.html[/url] . This is a precise recreation of actual concert hall sound using different impulse response of the respective concert hall. Unfortunately, that requires more than a million dollars for the 60 odds Sound Labs and MBL Radial speakers. There is surely a cheaper alternatives without the need to resort to complex pro sound solution.

In my previous post, I explained how I used JRiver zones and cheap speakers to create the wall. Now, I realized that I could get a much more simpler and cheaper solution with a very basic HT system. Currently, I added and additional of 4 speakers using Panasonic SH-XC333 (I now realise there were cheaper Sony option at RM380). This gives me a total of 8 speakers to create the ambiance of St. Cecilia concert hall using a single impulse response filter. Ideally, I should be using the 20 over IR but that is currently not within my reach. The role of the surround a.k.a  virtual wall speakers are just to act like a wall of the concert hall surface and therefore it need not be loud.

With the use of the convolution function with JRiver for the additional 8 speakers ( it will be soon increased to 16) . I have now able to recreate the reverberation like here.  How much and when you want them is all depend on you with a push of a button. Initially, I preferred the convolution for western classical but now after getting accustomed to virtual large hall ambiance I now prefer most of them with the surround speakers on.

This video below sshows what different ambiance could do to the sound.

https://www.youtube.com/watch?v=dWNV_JFolLg

I was pleasantly surprised to see XTC function included in Weiss DAC501/502.

XTC is crosstalk cancellation DSP like Ambiophonics. The advantage of Ambiophonics is it uses very little DSP and available FOC while the advantage of Weiss XTC is you need not move the speakers closer. I have not heard a Weiss 501 yet but it looks they actually embracing XTC more readily than other manufacturers.(edit 28/3/2017: -Looks like Weiss XTX is BACCH based).

Earlier attempt by Weiss with the collaboration of PSI Audio was the speaker unit with similar configuration as my setup. Unfortunately, I believe it did not materialize.

Some similarities between Weiss and Ambiophonics.

Weiss XTC

"The Ambiophonics method combines an exploitation of seldom applied, but well documented, psychoacoustic principles with the basic rules of good musical performance space design to create believable concert-hall sound fields in dedicated home listening rooms. Ambiophonics moves the listener into the same space as the performers, by accommodating to individual external ear and head characteristics, minimizing interaural correlation at the listening positions, abandoning the traditional stereo loudspeaker equilateral triangle, recreating early reflections and reverberant fields via computer, and eliminating front and rear loudspeaker crosstalk. The Ambiophonic sound reproducing technique is a worthy successor to both stereophonic or surround-sound listening configurations, in that it can consistently generate a "You Are There" sound field even from standard LPs, DVDs or CDs that the ear-brain system will accept as real."

Ambiophonics pricinciple

"Ambiophonic programs and components eliminate localization distortion, stereo loudspeaker crosstalk, produce front/rear sound stages (180°/360°), and deliver the high fidelity one ascribes to normal everyday 2-eared hearing. Loudspeaker binaural (in contrast to the stereo loudspeaker triangle) generates an unprecedented sense of depth, clarity and presence. Also, in the absence of stereo crosstalk, comb filtering and central low-bass energy doubling are eliminated making truly flat acoustic response possible.

I have now reached the end of my six years journey understanding Ambiophonics. It took me almost 5 years to get the tonal balance correct with the typical audiophile recordings typically with little ambiance information to reap the benefit of Ambiophonics. However, once you understood the effects, it was a quick upgrades within the last six months reaching the ultimate satisfaction. 

Now, I have upgrade the NUC3050 with i7 PC. Although, audio requires very little processing power the addition of convolution processing requires  faster CPU which can handle multiple convolution channels. The current setup is with 7 DACs and 8 amplifiers. The numbers of the DAC will be increased by another 6 to 10 depending when the law of diminishing returns sets in. That is when extra addition of convolution no longer detectable by blind testing and within the limits of i7. 

For now this is the setup as below. All with the reliable JRiver. You have to open the image separately for the picture to be shown in full. 

Managed to reconfigure JRiver to deal with more zones. I am not familiar with other media players except for iTunes and WMP. In my opinion, JRiver is a very versatile media player with capabilities that can only be implemented with pro tools.

Now running 10 amplifiers and 22 surround for convolution. Below is the sample with 8 speakers.

My complete progress can be seen in my profile link.

Click here:

Due to some criticism that it was impossible to hear the difference with the sound quality of Youtube and SoundCloud, I am posting the old recording which was recorded in 24/96 format. The original file is also included there for side by side comparisons.

Edit on 18/4/2017. Apparently the playback is not in WAV. Please PM me if you want the WAV file.

Except for the original sound, the stereo and ambiophonics recordings were made with binaural microphone at my siting location. I am still in process of locating all the speakers correctly and will update the final sound with either 26.1 speakers sound or 20.1 sound. I am leaning towards the latter. This is a binaural recordings and therefore use the headphones or earphones.

Thanks for viewing.

Edited - See latest videos post 10 June 2017 onward.

Updated with the latest 20.1 sound. The progress can be heard when you compare the sound clips side by side. Although convolution of 2.1s was not meant for studio recording but it sound okay too. 

And this is how a big ensemble sounds like. 

https://youtu.be/XNyZ-_YEoKY

The work in progress continues to address the correct timing between 34 channels. It is quite a challenge to correctly determine the correct delay so that the 18 speakers correctly matches the main speakers.

Unfortunately, the correct delay cannot be determine by calculation and requires listening and make the adjustment in the JRiver.

This is Norah Jones again with all the speakers correctly adjusted for smooth decay. I still need to adjust the errors of sibilance coming from the cheap surround speakers. Unfortunately, the cheap microphones couldn't capture the low frequencies exactly.

This recording made with Roland CS-10Em microphones from the sitting location. You are hearing the actual sound coming from all the 20.1 speakers.

Sharing what I posted in CA forum. Sharing more info about my current setup. I have also added SIR2 convolver to the JRiver. As usual, theis videos must be heard with headphones. 

My current system is rather simple and the best tweak I added is the convolution channels for the stereo playback. The additional price I paid is just about $600 to get the additional twenty over speakers and DACs. I got these.....



These addition now allow me to experience far better improvement  as can be seen in the videos below.

These videos were made in my room from my listening chair using a cheap binaural microphone (Roland CS-10EM). You have to use headphones to hear the changes. Due to 18 additional speakers recreating the ambiance by adding to the main speakers, the sound played directly through speakers can be muddy and congested and therefore use headphones to hear closer to what you would hear at the listening seat. Due to the limitation of the $75 microphone, the high frequencies are too bright. It also failed to capture the low level ambiance.

Some points about these videos:-

1) No changes made to the files. All were recorded in 24/96 format and uploaded as such. However, youtube would have compressed the audio.

2) They were recorded using Zoom H1 recorder connected to Roland microphone. Level was set at 90 of 100.

3) The actual music level while making this recording at  about 214cm from the main speakers were Avg =  76dB and Peak at 87.5dB.

4) The actual preference is convolution at 50 or 80%. In this video, the first one was at 50%, second one at 80% and the third at 100% level which I use for large orchestra music with high ambiance content. 

4) This setup is my early attempts to recreate the 3D concert hall ambiance in small space. I will add more info about this later.

6) You may hear slight timing difference due to my omission when changing the default values when doing this demonstration videos.

7) You may hear the air conditioner which is getting a little noisier lately. 

In the above videos, I have deliberately added excessive convolution to show the effect of using them. The videos below is the same recording but one is without convolution as the original recording does not have much ambiance to retrieve and the second one with just the right amount which sounds pleasant to my ears. The harsh sound of high frequencies is still there and it now points to the shortcimings of $75 binaural microphone.

This is the recording of Norah Jones - Come Away With Me and 2L samplers. Here the sibilance and soundstage is somewhat better when I use a small board to separate the left and right microphones by holding them inform of my face while recording. Of course, the board itself added some distortion. The sound was recorded from 20 speakers and therefore sound through speakers may sound congested. Use headphones to hear the difference and SQ as it is difficult to judge sound quality from Youtube videos.   

https://youtu.be/i40VgJGjjwU

https://youtu.be/R8gO5Si3FqI?t=56s
 
https://youtu.be/bN-LjnZuAjE?t=23s

I made minor changes to address the sibilance and too centralized sound in the Youtube. Slightly better results. No sanctuary here....

Reproduced.

This is a very effective tweak to enhance your current system by adding the virtual wall to recreate a concert hall like listening experience. IMO, this is the best room acoustics treatment without the need to spend $$$ on diffusers and other tweaks. 
 
One of the limitations we face in our listening room is space. In a confined space, we do not have the luxury to have RT60 of a big venue like a concert hall.  In all classical recordings, the ambience is eliminated or limited to the just the front stage so that during the replay, your front speakers would reproduce the captured front stage ambience and your side walls would add a little reflection to enhance to overall listening experience. The side wall reflection in the recording is completely eliminated unless it a proper 5.1 surround recordings. Most audiophiles already mastered the correct setup and room treatment to get this ratio correct. However, this ratio will never be as accurate or complete as the real concert hall’s RT60 of about 2 seconds.
 
Previously, I used IR of St. Cecilia recorded every 15 degrees at 23 degrees elevation. The 16 IRs although made a huge difference between standard stereo and stereo with convolution sound but it was never enough as in real concert hall where the reflection comes from various angles. Now, I have found a way to enhance this effect to get closer to concert hall experience.
 
From the picture in my first post, you could see I am now using 48 IR channels to recreate the virtual hall. This is done by sending additional copies of the respective IR with a delay of 10ms incrementally. I am now using only three copies but likely to double that as long as the CPU fan remain noiseless. 

Just an update. I have successfully added 216 channels for IRs alone. Total channels JRiver processing simultaneously is 222 channels playing in perfect sync without the need for master clock. 

To those interested in convolution here is a good write up. But this is limited to recording and not recreating the ambiance.

https://www.soundonsound.com/techniques/convolution-processing-impulse-responses

After the JRiver MC24 upgrade syncing the multiple became quiet a challenge due to different implementation of JRiver zoning function. This forced me to go for pro audio software and decided to choose Audiomulch over Cubase. There is no specific reason but Audiomulch was a little easier with prior guide provided.

This my signal route in Audiomulch after signal from JRiver. I have fine tuned and added EQ which is not shown in the chart but this the basis setup that should work with any system.




The latest sound can be heard in these two videos seen below. You can also checked all the videos to hear how the sound was improved over the years.



and

Finally, took the big step to replace the cheap DAC and Mytek with one multi DAC pro equipment - Motu 24Ao. I have also replaced Audiomulch with Reaper. Now, JRiver feeds Reaper directly to run 72 channels. Reaper also made it easier to do side by side comparison with pure stereo playback. The biggest advantage of the extra channels is it helps to correct rooms natural peaks and dips besides controlling the desire RT suitable for a given genre. 

https://youtu.be/jRt0II_kr60

After a a few months of chit chat over Whatsapp, I finally found time to visit STC's den to experience first hand what ambiophonics sounds like. Spent about 3 hours or so experiencing various musical genre and also a few 'settings'.

Ir was 3 dimensional experience where music just envelopes around you in a very reliastic manner. Layer bu layer was of the music was exposed in such detail. The drum rolls on track like Black Magic Woman was presented in a manner that I could actually imagine how the drum set was configured. The soundstage was full but at the same time, the vocal was set smack in the middle and did not loose any prowesness. 

Bass however, was little shy to my preference. It was extended and not thud thumping bass. But I can do with a little bit more of a dive in the lower frequency.

I can see why (but I dont understand why) audiophiles can reject this ambiophonics idea but experience it and let your emotion or hearing decide it. I certainly enjoyed it and if I ever have my own hifi room back, I may look into ambiophonics.

Thanks STC for the demo and also the murtabak kambing hehehe.

After several years of struggle to get the tonal balance correct I have finally reached the end of the journey as the I9-9900K CPU is already at max level to process over a hundred tracks for the perfect convolution. 

More info at https://goo.gl/CF1C6B

This setup is based on Ambiophonics with domestic concert hall concept using additional speakers to reproduce the hall ambiance using impulse responses of St. Cecilia Concert hall. 30 surround speakers produce the ambiance from various angles. 

https://youtu.be/WwCgHK8hMb8

The YouTube version of the original is here or you can download the free version from 2L website. 

https://youtu.be/Ym4sH9VaHbU