Quality of sound through speakers- something off (Page 2 of 2)

Ha!  I have experimented with the Hafler effect quite a lot ... added amplification for example etc.

Lanny

Thank you again Qexl for this exciting feedback.

Who has the solution to offer us several lives to try all the artistically possible combinations upstream and downstream of Pianoteq / Organteq??
Modartt could make a little effort, all the same ...

Bruno

Hello All,

As being another old guy in this forum, I used the Hafler circuit in the early 1980's to very pleasant effect.  Here is what you do to achieve the (David) Hafler Effect.  You create a rear channel by feeding a separate rear-mounted speaker from BOTH positive (red) terminals of your amplifier.   Those audio signals that are "in phase" rarely get passed because there is no voltage differential in the signal between left- and right channels.  However, the amplifier passes all signals that are not identical ... to the rear speaker.  I used to own an old Bose 1801 amplifier that was designed in the 1970s to work with the infamous power-hungry Bose 901s of the day.  Why? That's because the Bose 1801 had separate volume controls for each channel; this made it very easy to "null out", and control the overall gain of those signals that were out of phase between each channel. (Reverberation is usually out of phase and diffused with respect to the intended audio signal.)

As an aside:  I do not wish to start a war about speaker cables in this forum, but I do want to share a true anecdote that refers to the Hafler Effect.  A few years ago, I attended a hi-fi show where Audioquest was touting their then-newest line of speaker cable.  When the demonstrator was asked what the differences were between various levels of their own cables, the Audioquest engineer/marketing/sales person spoke only in vague terms that was basically double-talk and made no engineering sense at all. 

I suggested that Audioquest try hooking one model of cable to the positive terminal of the left channel, and a different model of Audioquest cable to the positive terminal of the right channel, send a mono source through the amplifier, and then listen for ANY changes signal at all that came through.  Supposedly, this Hafler Effect would demonstrate the superiority of the better model of Audioquest cable by allowing one to hear, only the highest audible frequencies or whatever voltage differences had been laid bare.  The Audioquest representative became very silent and curt, and he refused my request for such an experiment.  I was effectively escorted out of the demonstration as being some kind of cable-crazy troublemaker! (Aside, I think the major difference between Model A and Model B of the same brand was ... $300USD per meter pair, and not much else!)

My two cents once again.

Cheers,

Joe

Hello Lanny,

You are absolutely correct, sir.  I completely concur, having forgotten to acknowledge Class D amplifier topology.  My answer lay in older types of amplifiers that the so-called "negative" (usually black) speaker wire connectors are both actually connections to ground.  My apologies.

Cheers,

Joe

EDIT:  The same warning goes for conventional amplifiers operating in "bridged-to-mono" mode.  This will not work.  End Edit.

Hey PaptainClanet,

wild! That's ealy rininteresting

Love it - wishing you well with the build - exciting stuff and great ideas.

Sincerely, looking forward to the results and hope it gives you many years of fantastic enjoyment. Keep up posted for sure, I'm pretty chuffed seeing people using Piantoeq for this kind of setup.

Adjustment of the speaker response curve with the free APO equalizer for PC:  (feedback from the test)

Two months ago, the question of improving the audio system, downstream of pianoteq, was raised in the case of the speakers;
I thought then that listening with headphones improved Pianoteq's rendering much more effectively (and much more economically) than listening on loudspeakers. As a result, I had never seriously tried to improve listening other than using headphones. I am also lucky to be able to play on acoustic piano, much more realistic than any virtual piano model to date regardless of the technology used. The main advantage of the virtual piano for the amateur that I am, is the possibility of playing at any time, without waking my entourage, while having the possibility of shaping the sound with these formidable possibilities of synthesis which go very beyond those of an acoustic instrument and which make pianoteq an entirely original instrument.
But Organteq arrived, with in this precise case the impossibility of reconstructing the instrument (and the building which is part of it) at home. The headphones then proved to be insufficient to physically feel the vibrations of this magnificent instrument. The search for a solution allowing to project in my room in a faithful way the most serious sounds proved essential. To my surprise, the first tests on loudspeakers quickly showed me that Pianoteq could also benefit by establishing another relationship with the instrument, by offering a more woody sound than hitherto no listening on headphones had was able to provide (despite various tests and proposals for adjusting the internal equalizer of Pianoteq)
I then tried various combinations of speakers, in the context of tests and amateur equipment, before arriving at an exotic configuration, certainly acoustically objectionable, and which also has real disadvantages, but which provides a sound that 'it is possible to bring it a little closer to that obtained with a good headset, with the additional freedom of movement of the head and a more tactile feeling of the sound obtained, in the "player" position on the keyboard.
After having quickly perceived how many notes were artificially reinforced or muffled, and not being able to make a treatment of the walls, floors and ceilings to balance the absorption or reflection of sounds, I tried to make an equalization as precise as possible of the different frequencies between 14hz and 23khz (without too much illusion on the capacities of direct perception of the highest frequencies for the almost boomer that I am). The first difficulty was to find an interactive equalizer with enough adjustable frequencies (and response curve modalities for each of them).
Wishing to be able to quickly use Pianoteq or Organteq on demand without the inertia and complexity provided by a DAW, I preferred a solution compatible with the standalone versions of these instruments. I finally (to date) adopted the free APO equalizer which has the advantage of allowing to set up an almost infinite number of filters of a wide variety of types, without however (with a powerful PC) generate an audible delay in sound.
Having 5 different types of speakers and 3 types of subs, after listening to different combinations of these speakers and subs, I finally selected only the best available (Ruark Equinox for the speakers and Martin-Logan Grotto-i for the subs), other much more economical speakers (Sony of a mini-chain for example) however arrive at an almost as good result (in the register in which they were provided);
The equalizer adjustment was made iteratively after reproduction (the 1st time without equalization) of a set of frequencies from 13.7 Hz to 23.6 kHz (A-2 to F#9) incremented in steps of 1/4 tone - equal temperament, with A 440 Hz for the reference of the frequencies of a given tone.
After each reproduction, followed by a re-adjustment, the recording was made on 2 SE Electronics SE-8 cardioid microphones placed at the position of the head of a player on the keyboard, with a separation between the 2 microphones as a head to better separate the right and left tracks.
The adjustment of the extreme bass and treble has proven to be very profitable with this method, but on the other hand much less for the mid frequencies between 500 Hz and 9 Khz, giving with a "mechanical" equalization of the differences observed at the microphones, a much too loud and over-acute, hence the establishment of a counter-adjustment "by ear" on these frequencies.
The problem comes (I suppose?) From an equalization on the basis of feedback from the recording done after remixing the 2 L and R channels in mono to have a medium level. In fact it probably should be? equalize the L channel and R channel separately for medium frequencies.
One of the beneficial effects of this equalization and of the positioning of the chosen speakers is the very weak diffusion of sound far from the keyboard (including for very low frequencies) which has the advantage of not disturbing the surroundings (even night) while enjoying a very acceptable sound level given the proximity of the speakers.

Link: https://lh3.googleusercontent.com/RPj3s...8-h1598-no

The result (still largely perfectible) obtained after 3 iterations of sinusoid generation, reproduction by loudspeaker, recording at the microphone, re-adjustment of the equalizers is currently as follows: (available on the links below, which I invite you To download)
(1) Original source recording (without HP) of sinusoids from 13.7hz to 23.6khz quarter tone by quarter tone, based on the reference of La 440hz (thus between La -2 and Fa # 9)
with a level of -8db (so as not to approach the level of possible distortion beyond 0db at the DAC level, capable of correct reproduction from 10hz to 40khz)
Link: http://dl.free.fr/foIKdAezK
(2) Reproduction of these sinusoids on HP recorded with my 2 microphones at the location of the player's head on the keyboard
        (with a separation between the 2 microphones simulating the head) The level remains moderate (at the level of the APO equalization, a "preventive" attenuation of 12db + sound level of the PC volume at 24% of the volume, the volume level on the Yamaha amplifier would represent on a time dial a position at 50mn on a 60-minute dial. (1 / 3-)
The recording of each track separated from the next one by 1/4 of a tone is done during the regular duration of 1.5 seconds, allowing to present with Audacity the level curve, with a horizontal axis of times according to the frequency as if if the latter was logarithmic (also expressed in db on the vertical axis), thus allowing to find "graphically" a traditional frequency response curve.
(below reproduced in .flac format, to be opened with Audacity to visualize the response curve as described above)

Link: http://dl.free.fr/qudUOuhX0
(3) Reproduction on HP, with the same APO equalization, and recording at the same position, with the same microphones of several pieces intended to show the result:
Link: http://dl.free.fr/g1KtuBuEc  (Result of this test to listen with headphones of course.)

Content:
Piece 1: From start to 11mn00s: Bach-Vivaldi / Concerto for 4 Pianos / MultiPiano Ensemble
 YouTube link of the original recording: https://youtu.be/7OwQOb6bd1M
Piece 2: From 11mn00s to 21mn00s: Saturn, the Bringer of Old Age - The Planets - Holst - Charles Dutoit
YouTube link of the original recording: https://youtu.be/6sbnsLmwlbc
The interest of this piece (besides very beautiful images of Saturn by NASA) is the presence of various instruments and especially
at the end (from 19mn12s for around twenty seconds, then around 20mn17s until the end the use of organ pipes between 16 and 20hz to double the double bass). The sound level having been slightly too important for this during this recording, I repeated just after this end of transition to a level passed from 55% to 40% of volume on the PC to make disappear the distortion, and the noise of beat of the sub-bass membrane linked to the excessive level compared to the "reduced" capacity of clearance of the membranes of these speakers at such low frequencies.
The transition to lower level of these very low frequencies (eliminating the noise of membrane movement) is presented here between: 21mn00s and 23mn02s
Piece 3: From 23mn02s to 31mn45s: JS Bach / Organist Andre Issoir, Prélude and Fugue in La min BWV 543 (Organ of the Abbey of St Cyprien in Périgord)
Original file, link: http://dl.free.fr/vPM1Shrgj
Piece 4: From 31mn45s to 34mn50s: Bruno test (Some chords) on real Bosendorfer 200 (96khz / 24 bit sound recording on 3 microphones: 2 SE8 + 1 Rode 1000 at the tail of the piano)
Original file, link: http://dl.free.fr/ea5Q8N8vl
Piece 5: From 34mn50s to 37mn40: Bruno test (Some chords) on Pianoteq 6.7 (.fxp Bosendorfer model) with Piano Casio GP500 keyboard
Original file, .flac generated by Pianoteq: link: http://dl.free.fr/egV59R6hg
Piece 6: From 37mn40 to the end: Short Bruno test on Organteq 1.05 (default configuration) with Piano Casio GP500 keyboard
Original file, .flac generated by Organteq: link: http://dl.free.fr/voOOqmRy6

In conclusion, in these times of confinement during which some of us sometimes have a little more time,
I can only invite to test the APO equalizer for those of you who wish to improve the sound of Pianoteq or Organteq with their speakers. On the other hand, it takes a lot of patience before obtaining a truly satisfactory result.
Links: https://sourceforge.net/p/equalizerapo/...reference/    https://sourceforge.net/projects/equali...les/1.2.1/ 
The processing of the audio chain downstream of Pianoteq / Organteq can subsequently make it possible to focus with the settings available on Pianoteq / Organteq only on the intrinsic part of the virtual instrument.

Bruno

Excellent work bm - always so interesting to read of your well documented experiments.

And it's really great to know you're finding enjoyable ways to work your space MeDorian

Everyone will have their own perfect fit. For me it's enough to go the near-field stereo way for various reasons but always fascinated by the efforts of others to tease their dpiano towards more realism in the room.

Keep at it - I can only speak highly of the usability of Pianoteq.

Look forward to that - best of luck with it