For a feeling of how reverberation/upmix with the The FIReverb Suite actually may sound, this page gives mp3-examples
of stereo to stereo reverb/remix and stereo to 5-channel reverb/upmix from anechoically recorded material, using a binaural
listening simulation. In addition, stereo and 5-channel binaural listening simulation is explained and exemplified so in some
ways it is also a demonstration of CATT-Acoustic auralization.

Skip the explanation/background info, go directly to demonstration.

How the demonstration is made

The multi-channel listening is presented as a binaural simulation. You should imagine what you hear as if you were sitting in a listening room with the speakers surrounding you, like in the figure to the right, and not as if you were listening over headphones to a stereo recording (i.e. you should pretend that the headphones aren't there). The simulated listening room is well designed with appropriate diffusing and absorbing elements and the reverberation times are: 125 250 500 1k 2k 4k Hz 0.40 0.30 0.25 0.25 0.25 0.25 s The loudspeakers used have the directivities of a studio monitor. You will hear examples of how each separate speaker sounds.
The final binaural files are converted to mp3 using the best encoding setting (giving around 1:5 compression)
Both long and short samples are avilable for selection  download times: 35 sec (1.2 MB) or 12 sec (380 kB) per sample.

The anechoically recorded music:
 

Performance by the flute/guitar duo Duo Dulce consisting of Mikael Ekholm, flute, and Fredrik Runnström, guitar. Both started their musical studies at the School of Music, Malmö, in 1995. Mikael Ekholm has studied  under Anders Ljungar-Chapelon och Sven-Olof Johansson in Malmö with emphasis on solo and orchestral music playing. He has the past summers given a number of chamber-music concerts in the south of Sweden, mostly in a flute/guitar duo setting but also with piano. Mikael is also about to complete a Master of Science degree with a thesis at the Dept. of Applied Acoustics, Chalmers University of Technology, Gothenburg (topic: comparison between scale- and computer- modeling auralization). Fredrik Runnström has studied under professor Göran Söllscher and Gunnar Spjuth. Fredrik performed chamber music in 1997 at the inauguration of Rikskonserter's new venue in Stockholm. During 1997 and 1998 he has also appeared as a chamber music performer in the Swedish National Radio channel 2. He now studies at the Royal Academy of Music in London.	During the summer of 1999 Duo Dulce gave five church concert performances in the south of Sweden with a 50-min program of flute and guitar music. Web page (in Swedish): Duo Dulce
Music played here is a 35 (or 12) sec excerpt from C.P.E. Bach, The Hamburger Sonata.
Recording was made by CATT August 1999 in the  anechoic room at the Dept. of Applied Acoustics, Chalmers University of Technology, Gothenburg, Sweden. At this time the complete 50-min program was recorded to be used as musical examples for convolution with room responses.
Equipment used included: Beyer MCE93 cardioid condenser microphones Art Dual MP tube pre-amplifier Sony PRO DAT PCM-R500 at 44.1 kHz with 20-bit processing

Subsequent processing:

• All-digital editing/transfer using standard sound editors
• No de-noising, limiting, compression or other processing of the recorded material,
  i.e. used "as is"
• Stereo to stereo reverb/re-mix (2x2) and stereo to 5-channel reverb/upmix (2x5)
  FIR-filters were created
  by PureVerb™ . The room and 5-channel setup used were:
• Room dimensions W x L x H  =   12 x 18 x 8 m³
• 5-channel microphone distance 2 m (one reverberation radius - the 'medium'
  distance preset - 'close' distance gives less strong reverberation)
• 5-channel microphones all cardioids configurated as:




• Reverberation times:

125	250	500	1k	2k	4k	8k	16k	Hz
1.8	1.7	1.6	1.6	1.5	1.4	1.0	0.5	sec

• Directivity index (controls the direct sound to reverberation ratio):

125	250	500	1k	2k	4k	8k	16k	Hz
0.0	0.0	0.0	0.0	0.5	1.0	2.0	3.0	dB

• Stereo to stereo and stereo to 5-channel convolution/upmix were done by MultiVolver™
• Binaural listening simulation was created by CATT-Acoustic™ (left/right binaural filters
  are summations of signals from all five loudspeakers at the listening position in the
  5-channel listening room, a.k.a. auralization).
• Binaural listening simulation was equalized for headphones Beyer Dynamic DT-990
  (using other headphones will give a difference in coloration but is not crucial for the demo)
• Stereo to binaural (2x2) convolution/remix and 5-channel to binaural (5x2) convolution/
  downmix were made using MultiVolver™.
• MP3 encoding from 16-bit WAV was made by the FraunhoferIIS-A MPEG Layer-3 Encoder v3.1
• The complete chain for a binaurally simulated 5-channel listening to a stereo to 5-channel
  reverb/upmix can be depicted as:

Demonstration

Material	Processing	Sketch	Comment/mp3-file	Guidance to "state of mind"
Pink noise	Binaural listening simulation		A burst of pink noise cycles C L R  Ls  Rs.  L = Left R = Right C = Center Ls = Left surround Rs = Right surround Download/play: FIReverb_1_lspsetup.mp3 (short, 148 kB)	I  imagine sitting in my 5-ch listening room, and bursts of pink noise emit from one loudspeaker at a time.
Stereo flute/guitar duo	Anechoic		Exactly as recorded. Download/play: FIReverb_2_stereo_anechoic_s.mp3 (short, 386 kB) FIReverb_2_stereo_anechoic_l.mp3 (long, 1106 kB)	I am listening to an anechoic stereo signal via headphones: The sound is very clear but completely inside my head.
Stereo flute/guitar duo	Binaural listening simulation		How the anechoic stereo recording of the duo sounds when replayed trough speakers L and R with only the natural acoustics of the listening room added.  Download/play: FIReverb_3_stereo_replay_s.mp3 (short, 386 kB) FIReverb_3_stereo_replay_l.mp3 (long, 1129 kB)	I imagine sitting in my 5-ch listening room listening to stereo-speaker replay of  an anechoically recorded stereo signal. I must now think that I do no longer listen via head- phones. I notice that it is more relaxing to listen and that the sound is no longer inside my head (1).
Stereo flute/guitar duo	Binaural listening simulation		How the anechoic stereo recording of the duo sounds after a stereo remix with reverb as heard over only the left and right speakers in the listening room. The two read squares indicate placement of two virtual sound sources used for the right and left anechoic stereo signals. Note: the amount of reverb is likely to be too high for many tastes but this is a demo that attempts to clearly illustrate the principles rather than be subtle. Download/play: FIReverb_4_stereo_remix_s.mp3 (short, 383 kB) FIReverb_4_stereo_remix_l.mp3 (long, 1185 kB)	I imagine sitting in my 5-ch listening room listening to stereo-speaker replay of a stereo recording of the duo playing in a concert hall.
Stereo flute/guitar duo	Binaural listening  simulation		How the anechoic stereo recording of the duo sounds after a 5-channel upmix with reverb as heard over the five speakers in the listening room (2). With a real listening the difference from the previous example is more obvious. Also see Note on the previous example. Download/play: FIReverb_5_stereo_upmix_s.mp3 (short, 383 kB) FIReverb_5_stereo_upmix_l.mp3 (long, 1185 kB)	I imagine sitting in my 5-ch listening room listening to 5-speaker replay of a 5-ch recording of the duo playing in a concert hall.
Stereo flute/guitar duo	Binaural listening  simulation		How only the surround channels of the 5-channel upmix with reverb sound, as heard over the rear speakers in the listening room (3). If the stereo recording already has some reverberation, the surround filters can be used separately if they in PureVerb™ are made to match the hall in question.  Download/play: FIReverb_6_stereo_surr_s.mp3 (short, 387 kB) FIReverb_6_stereo_surr_l.mp3 (long, 1185 kB)	I imagine sitting in my 5-ch listening room listening to only the two surround speakers of a 5-ch recording of the duo playing in a concert hall. I notice that the surrounds have an obvious "room sound" as created by the rear-facing virtual mics used in PureVerb™.

 [ top ] [ info ] [ home ]