Paganini Violin Solo DSD256 (11.2 MHz 1bit) recording by Professor Yamasaki of Waseda University

[Bunpei]

Some of you might have listened a very impressive Paganini Violin solo piece of DSD256 (11.2 MHz 1 bit) sampling at some audio shows.

 

Nikhil posted a very valuable analyzing chart for the Paganini Violin solo and his comment on another thread.

http://www.computeraudiophile.com/f13-music-downloads-and-streaming/originally-recorded-dsd256-piano-solo-sources-japanese-group-wechseldominate-19942/#post517024

However, I am afraid that there is a certain gap between the original thread title and recent topics discussed there. That's why I created a new thread dedicated for the Paganini Violin solo for avoiding some confusions and for providing some useful information on the source file.

 

The original source file is on the following web page of "1 bit Consortium Japan" for a free download for these several years.

1bit Audio Consortium Web Page

Its link is this.

http://1bitcons.acoust.ias.sci.waseda.ac.jp/data10/2006_nagano_ohba.wsd

The source file was there, but deserted for a long time, partly because most of us has no hardware capable for playing such a high sampling rate DSD source, and partly because the source file lacks any regular header part of WSD format in spite of its WSD file extension.

 

Just the end of year 2015, a special issue of Japanese electronics technology magazine, "Toragi" (Transistor Technology), titled as "Building 1 bit Recorder and Fundamentals for Delta-Sigma Modulation"

トラ技エレã‚*工房 No.5 (This page is in Japanese)

included a normally-WSD-formatted file of the source in its companion DVD media.

However, you may need to convert from WSD format to DSF or DSDIFF format by using a suitable format converter program for playing the source file on your player.

[Nikhil]

Bunpei,

 

That was a fantastic recording that really was special to listen to the first time I played it in my setup. Just as others have mentioned it really is a standout recording. Massive dynamics and presence. On second (or third) listen I realized that - if anything negative can be said about the recording - it can be a bit hard on the ears at higher listening levels. Back off the volume knob just a touch while listening to the piece.

 

With new DACs available now that go all the way to DSD256 (11.2MHz), it is a good time to revisit this recording. We've got a much better understanding of DSD now then we did a few years ago.

 

Here is the Musicscope analysis of the file once again:

 

[Bunpei]

Hi, Nikhil!

 

Thank you very much for your re-posting here!

 

Bunpei

[Bunpei]

<Availability of the Paganini Violin Solo>

In the past, Bunpei downloaded the WSD file in the web page and converted it into DSDIFF formatted file with his own complicated way. The resulting DSDIFF file was distributed privately through his personal communication network to some audiophiles. As a result, it is said to be available at ExaSound user web site.

However, there is no web site where people can download the DSDIFF format file directly.

(If you are eager to obtain it, please send a PM of this forum to Bunpei.)

 

<Recording details>

The piece is a solo play part of famous Paganini's Concerto. It was played by Kristof Barati at Stradivari Society in Chicago on October 23, 2006. Recording staff were Geoffery FUSHI, Mr. and Mrs. Oba, Ms. Nagano (and Professor Yamasaki). A special microphone manufactured by B&K for Stradivari Society and a proprietary recording device developed by Professor Yamasaki with a second-order delta-sigma modulator were used.

 

<Popularity>

Bunpei heard that the source was played on Merging NADAC in some audio shows held in U.S.

[mrhizzo]

Anyone know what program can I use to convert to DSD/any other format?

[Bunpei]

Hi, mrhizzo!

 

May I make it clear? From which format do you say?

[mrhizzo]

Is from .wsd to .diff or flac. What is possible. I really don't know how to convert this file.

 

Thank you!

[Bunpei]

Korg AudioGate software can convert an audio file of WSD format to audio files of other formats. In my case, Korg AudioGate 2 was used because its conversion function was free in the form of Twitter-ware. However, AudioGate 2 accepts no input file of 11.2 MHz sampling. They can handle files of at most 5.6 MHz sampling rate. Therefore, I made a trick. I rewrote the actual value of "Sampling Frequency" field in a WSD header (A specification of WSD file format is available from here.http://1bitcons.acoust.ias.sci.waseda.ac.jp/pdf/wsd_file_format_ver1_1.pdf It is written in Japanese.) from 11.2 MHz to 5.6 MHz by a DIY small piece of program written in perl language. Then, Korg Audio Gate2, could convert the WSD file to DSDIFF file. The sampling rate field value in the resulting DSDIFF file was restored as 11.2 MHz by another program in perl. (In general, a file format conversion or re-sampling algorithm is independent on an absolute sampling frequency even in the case DSD is involved.)

[Bunpei]

The following two scripts in perl language show just a concept of rewriting and restoring the sampling frequency field value in each format.

(Bunpei does not guarantee your results. When you apply these to your own files based on your own risk, be sure to take a backup of your original files to be processed. These scripts may damage your files.)

 

The perl script for rewriting "Sampling Frequency" field of WSD header is like this;

#!C:/Perl/bin/perl

use strict;

use warnings;

open FILE, "+<", "20061023_kristof_11M.wsd" or die "file open error: $!";

binmode FILE;

my $fs=pack("N",44100*128);

seek FILE, 0x24, 0;

print FILE $fs;

close FILE;

 

The another perl script for restoring "Sampling Frequency" field of DSDIFF header is like this;

#!C:/Perl/bin/perl

use strict;

use warnings;

open FILE, "+<", "20061023_kristof_11M.dff" or die "file open error: $!";

binmode FILE;

my $fs=pack("N",44100*256);

seek FILE, 0x3c, 0;

print FILE $fs;

close FILE;