Theo Verelst Audio Activity Page 2

theover@tiscali.nl

Text: March 2009


   This page is to give an updated idea of my involvements in audio related subjects, or at least directly related subjects.

The original, older page about this subject is here.


Lately I´ve been into watching Blu-Ray and HD-DVD (the actual ones, with the hologram and such, in the actual red box, not progressively coded traditional dvds. I watch those too at times, but that´s not the same experience). For a list of the 60 official HD (not dvd) titles Ive watched/heared on this system, apart from my own recordings and audio CDs, see here, the system passed the test of all these materials with flying colors, I think, probably I more often see and hear limitations of the production, or more likely the brainsuckers or pixelmess.gmbh ´master´ ´processors´ which make the bluray disc from the original film or into some regional production. Brrr, I love the DTS HD Master idea, and would prefer lets say 12 or 16 bits per component frames and more often also the choice for 24 bit surround (at 96kHz if I ask really polite ?). But at least a fair number of titles is already there, and more than a few have at least good quality, some even at 24 frames/sec and with good neutral audio, lets say like in Dirty Harry or the enforcer, those sound great on the system and are good example, especially when played loud... Don´t try that too hard on your small surround speakers, you might blow them up! This system is used to the better volumes, it is used at times even 24/7 already for many years on a row (with improvements but only few speaker and electronics repairs), but it is of course not a fair comparison with the modest sized systems which aunty and dad-at-home probably prefer in their square meters of home-cinema or TV space!

But then again, this system could also be used in a Film-House, at (medium small) live performances (as it has been, even a few times in the open air) and for studio use (regularly) and Public Adress, so it is truly multi-purpose.

I use my large audio system and the medium sized monitors, all which I built myself, the pre amps, filters, amps, and speaker boxes in this kind of way a lot (at least 1000s of hours already, I often use it more than 8 hours a day, including for TV sound), and it hasn´t bored or annoyed me much yet, ever, and believe me I´m a critical audio man already for over 3 decades:



Here's a link to an editable list of all CD songs I play(ed) regularly on my systems.


TTB:
    (1 19 inch height unit) Analog, digitally controlled mixer (NO AD/DA CONVERSION in the signal path!) with ethernet control (in buildup, 1 stereo channel and ethernet control and windows and linux control program with sliders functional)
   (4 HU) Self contained Digital Synthesizer doing a polyphonic (2/4 voices) analog synthesizer simulation with a fast responding Analog Devices Blackfin DSP and CPLD and 24 bit stereo inputs and outputs + high quality pre-amplifier + high quality active analog filter bank for audio band seperation filtering for the end-amps
   (2 HU) Main quadro amplifier (see below)
   (2 HU) Previous main amplifier acting as sub woofer amp in bridge mode.

All these are self made (seriously).


The audio path for watching bluray is as follows (see also here):



This is the global setup:


I´ve used damping all around the room, mainly of three different materials, softboard, black toothed carbon-like foam plates, and heavy multicolored foam plastic 4 cm thick damping plates. The back wall is partly covered by a thick matras, in the line of sight of the sub woofer and the left channel main speaker.

The floor is sound isolated (not for the neighbours, there aren´t any below) with softboard, under hardboard plates with (normally thick) carpet on top.

The ceiling is partly covered by black foam plates, using strings I attached to hooks high in the side walls.

Probably about half of the area of the side walls, 75 % of the front and back wall, and about a quarter of the ceiling are covered with some form of damping, some of the front and back walls with multiple layers and a few piles of damping plates.

That makes the room sound nicely damped, but certainly not acoustically dead. Bass boosts are quite damped, not many standing waves are clearly dominating, though I still have some issues with the resonating windows and doors.

Absolutely great cinema sound is possible with good TrueHD or DTS-HD uncompressed audio surround tracks to the point of sounding actually better (and when needed louder) than some cinema´s I´ve heard. Seriously.



The current main amplifier, a assymmetric quadro amp, meaning a heavy stereo main amp (over 100W sine RMS / CH with 400 Watt cont. rated toroid tranformer and 2x 10,000 uF capacitors) and a secondary amp with less power, which I use for the mid high speakers to be seperately amplified (2x50W guessed).

The signal path of the main amp is free from electrolytic capacitors, and the output is DC-coupled to the speakers. All resistors are precision metal film resistors, and the specs of the chips are quite good (0.005% (that´s a 1/200th of a percent) distortion at 1kHz @ 50 W, of course broad fequency range, 6V/uS IIRC slew rate, quite over 100dB signal to noise ratio, it´s quiet at quite a percentage of full power, counting from the 24 bit DA converter, I´ve sometimes slept next to it while all 10 channels were still on..).

In total the two amplifier units with 6 channels in total have 930 Watts (!) of manufacturer rated, continuous toroid transformer power.

The main amp parts (in this case just about all from Conrad):




This is the front pannel of the digital audio synthesizer I designed and prototyped (from general empty 19 onch rack enclosure, a DSP board from Analog Devices, a Xilinx CPLD board, electronics, mechanical parts, and of course software):


See this page for the Open Source design, and a few example sounds.

The front 3 group audio system seperation filter PCB has found a place in the digital synthesizer enclosure:


The prototype works very satisfactory already for years, though if this were a commercial product, I´d have to screw it in place better, like the rest of the machinery. I used the Burr-Brown / Texas Instruments OPA 627 which I got at cheapest at about 20 dollar each (per channel!) from a US electronics supplier in the filters and preamp, and also quality parts for the rest of the signal path (metal film precision resistors, good quality audio capacitors), and a nice small toroid transformer with of course supply stabelizer chips.


The rear speakers are bi-amped monitors, self contained, 4 (!) amplifiers built into the right speaker unit, and very high quality preamp and electronic filters, and they are used as indirect side-rear speakers, as shown here pointed at the (thick, hard type concrete, moderate tick wallpaper furnished) wall (see here).

The rear of the to-be analog mixer with ethernet control:

On the left the industrial grade ethernet connector fitted in the single unit 19 inch rack, on the right the mains and the fuse and an isolated MIDI connector (not used yet), and then there are 15 stereo jack connectors waiting to be used, because I have a lot of volume control chips, and the FPGA I used already has a lot of pins which can be used to drive those chips. The little unix board (a Fox Board) in the enclosure already runs connected with the FPGA and one of the volume control chips, and works to drive it over ethernet, not yet with the knobs and the big display (which does work however) and it even runs a scripting language which allows user scripts to be ftp-ed over the ethernet (using a standard local network with computers) in easy to use Tcl. In fact I´ve tested the board even to connect up to the internet by itself, without the help of a computer.

This is a closup from the test pcb containing 4 stereo A grade volume control chips which will act as a mixer:



The signals can be patched by small connectors, like in RC models. These are the other parts I tested before they were screwed in the enclosure:




The front of the unit with all this built in:



The display isn´t connected with the linux board yet, the knobs have only been tested, the boards internally are firmly screwed in place, just like the display and knobs, so all is prepared for the internal wiring and electronics and the programming of the CPLD (for the display, tests work), the FPGA for connecting with the unix board and the chips (works in practice, but only for 2 channels, I didn´t try more yet, except in another setup with the same chips, when I made a well working 10 channel mixer), and of course the linux board, which runs good and connects with computer programs over ethernet.

With a mobile phone with internet I could talk to the machine on the picture, to change the volume of 2 channels using a web-browser, which is great for sitting in the garden or being somehow not near a computer or volume knob. I used the same internet phone to chose songs to play on the server computer (where this page is served by, too) from 2000 wav files (which I never share with anyone) using the Linux mplayer program.

A while ago I worked on a Surround Reverb (4 channel) which I have working on a dual core Blackfin DSP devboard:


Based on the very good on board 24 bit AD and DA converters, and a number (>10) of cross-coupled long delay lines to simulate real big spaces, which has worked good and is an interesting research subject on my surround system. I´ve not (yet) continued with it, and it was not an open source project, sorry.


The above DSP board can be coupled with this (or another) FPGA board over a bus connection I made work, except I didn´t use this yet in audio, but in a fast graphics connection, see here. Considering the 20MHz working bus and the already quite noticable digital power of such FPGA it would be a usefull audio project, too.



An impression of a scene from ´The Usual Suspects´ from bluray on the 720HD DLP beamer I use with a 106 inch diagonal (2.3m eff. wide) pro grade 1.5 reflecting 16:9 screen I use, and the lights dimmed:

(See also here)


I lately always use the Lexicon DA converter with TOS link from the bluray player at 96kS/s aith 24 bits:



Which beats the built-in DTS converters from the players, in spite of the extra sample-conversion,



(or TrueHD ).



The Lexicon Omega 24 bit AD/DA converter, 4 ch in 2 ch out, with Usb and Tos IO, which unfortunately can´t handle pullng the knobs off very well, so it needs to heat up and reset a number of times before it works good...
The sounds is the best, though, very even, detailed, nice low end and transients, and recording is quite nice with it, almost transparent, and excellent microphone pre amplifiers (check out the very good specs, or try one!).

At the moment they cost about 160 euros at Thomann (in germany) where I bought this one years ago, that´s a real bargain.



Less high-tech points

I´ve recently put the sub and left main speaker on solid ground by putting 9 nails through the carped and underlying damping, so that they rest solidly on the floor concrete :

Major sound improvement was the result!

Remember, this is a big system, it takes quite some power to run, especially since I make no use of bass reflex effects, or standing waves in the room: everything is damped to have neutral response! Every tone, all the low and transient waves must be created by pure amplifier and speaker power, not by resonating energy. So when home-cinema sound levels are reached, quite some power is pushing the airwaves and moving speaker cones back and forth, and because action equals counter action, the speakers may tend to woble a bit. A bit however is enough to audibly modulate higher frequencies, which is prevented by the nails resting on the solid concrete.



The fuse blows when I put on the main amp without slowstart addition about 1 in three times, because of the possible current the transformer can draw! So I made the above 150 Watt halogen light with shunt switch and 1.5 sqmm wiring to slow start the main amp, which has worked decent for quite some time now, and no more fuse problems anymore.


I don´t use these a lot, even though they´re quality electronics parts, I prefer non-bipolar capacitors as tweeter filter.  Some time ago I replaced a electrolitic capacitor from the feedback path of the main amp by replacing it by an array of some polyXXX ordinary high quality caps, which can be seen in the picture above of the amp. Maybe not needed for everyone, but liberating for my hearing standards.

These are passive audio tools, a volume control in between two machines with an extension cord, and a passive 4 channel stereo mixer (meaning no amplification or other electronics which could introduce signal distortion):




Phase linearization of the mid high speaker (while keeping damping and airtightness under control):




The audio equipment  I built (all!) myself in the midle, the right main speaker, and the screen during buildup-time:

also visible a over half a meter high (infact it was about 2 meters high when I moved it) stack of heavy damping materials left to the speaker. I later added small pieces of (other) damping between the speaker and the screen, to prevent 'blowing' sound behind and in front of the wide screen, that can make quite a difference for the mid and high frequencies.

I think it is a great improvement to have gone from a wooden floor and partially wooden walls:

to the above concrete floor with nails support and damping.

I´ve experimented with Audio Technica AT2020s for HD film and music recording, for instance for singing I´ve tried this, which appears to work decent:

It´s fun to try the live sound of the damped space with the above amplification and this mic (or two of them), also by  taking a signal from another room, without AD/DA conversion to try out realism levels.

A midi file I used to make a song with using a S90 synthesizer and this mic:

  CD quality  Another one bytes the dust   rendered by me (See local diary page 53) .

I performed a couple of times at the jam session of the new Amsterdam Bim Huis, which was fun. I recorded a conservatory graduation performance (I think), in HD.

On Vimeo I put films pieces with HD cam sound, and a 3D graphics example I rendered myself with music from open source software played by me.

   


Click on the pictures to go to the film fragment pages. The audio except for the graphics was made in 320kbps mpeg-2 stereo with the HD cam.

Recently a HD video in small (web) form made it to the dutch section lecture retrospect pages, here is a Wiki server I run which contains three versions, amoung which a 720p quality one.

I put up a 14 minute improvisation of "Oh when the saints go marching in" I recorded with the Lexicon converter on a S90 synthesizer,  while recording the playing in 1080i HD on youtube, in 720  HD, using only open source software (Ardour audio recording, dvgrab, ffmpeg):



I´ve repaired (indeed) a Gainbrain II compressor unit from a (former commercial) studio, and this, too:


A not very recent but still used and decent digital multitrack recorder from Akai. In fact I worked with it in the studio a bit, too, and even ´saved´ some backup files from 2 units, and imported the wavs in various audio programs (e.g. rosegarden) with good succes. I tried the gainbrain of course and even recorded a blues with it.


Meanwhile as Associate Member, I´ve been to numerous AES lectures, like this one at Philips research:



Where in 2006 I´ve brought a (smaller then the one I used to write this on) notebook by train to do a short introduction of some of the things on this page (see):



In fact I´ve demonstrated my main system at the 2008 year meeting of the dutch Aes chapter (see here):



Last year the main conference was in Amsterdam, so I visited and joined some of the technical commitee sessions in this room:







I was at the above workshop by Dennis Chambers, one of the top, probably the top Jazz-fusion drummer of the world, and asked him if he wouldn´t find it a normal idea to let the artists chose the PA system he or she prefers! He gave an elaborate answer.


I´ve presented my equipment to the general public twice on the MILE 2005 and MILE2006 musical instruments live event / tradeshows with a little stand:




This was the poster I brought the second time, when I´d made the new quadro amp in 4 days just in time for the event, where it performed great:

(click to enlarge)


A new field in audio is FPGA-based sound synthesis:



See here.

Something else I´ve worked on is a wireless, bidirectional digital remote control for the abovementioned digitally controlled analog mixer, using these TI 2.5 GHz wireless transceiver boards with the famous MPS 430 family microcontrollers:





Software-wise I´ve meanwhile worked with windows (XP, XP Pro) and Linux (Fedora 8, 8/64, 10/64 mainly) and various well known and less well known software such as Cubase with Lexicon reverb plugin, Miditizer (great sounds, albeit the sampling aspect isn´t perfect) Pinnacle studio 10.5 for HD film editing and processing, Visual Audio, Sofware Radio ,Cinelerra (linux HD video processing and audio processing in 32 bit floating point with effects), Rosegarden/Ardour, Guitar-rack and Rakarack using Jack (all on Linux), various Virtual synthesizers, Gigasampler (I conpiled it working on Fedora), Jamin (large FFT based equalizer/multiband compressor), Ingenuity, of course many LADSPA plugins, ffmpeg, finally a working Bristol synthesizer simulation suite, and my own software.



Rosegarden on a HD screen notebook (actually 1080x1920 pixels! Very good for HD film monitoring), running
XP-Pro and Fedora 8 (dual core), and using the Lexicon Omega or my self built 16 bit DA converter.

Of course I also make music, like here with a Ibanez electric guitar on the front side high impedance 24 bit Lexicon input with Linux software (fft based 1024 band equalizer/multiband compressor, guitar effect unit, virtual rack with reverb, delay and compression all in 32 bits floating point Jack signal path):


And this is an example result:

   .wav guitar rough  (CD stereo quality, 9MB)

I´ve done some ARM cpu developing :


   Analog Devices Arm7 MC with AD converter from a volume control slider and serial interface to PGA2511 volume control chip from Burr Brown/TI. Devenv:


   Keil/Gnu for a working driver program to set set the digital volume by a DC voltage from a slider, at some speed and with full accuracy.

An interesting package I experimented with (with working results, exept the Usb-UI coupling wasn´t perfect), with the hardware from my digital synthesizer is Visual Audio Designer from Analog Devices:






My name is on the Linux Audio Software page because of my additive synthesis software (under geneal synthesis packages) I made while in France some time ago.

For example, mathematical wave graphs and audio generator software:


see here.

I´ve even made an internet server with interaction-time mathematical formula rendering into mp3, see here.

See also my Linux Audio page for impulse and mastering effects experiments, as well as Gigasampler tracks.
Recently I´ve set up the first working prototype of a Cuda based audio program.



home page [ 2 3 ]       email: theover@tiscali.nl [2]
This server can be searched!