Petr Simanek had trouble posting this and asked me to send it to MD-L.
Please direct any private replies (MD-L is configured to clobber Reply-To:
headers) to him at [EMAIL PROTECTED]
If you respond on the list, please be careful to fix your attribution so that
you credit Petr, not me, with any text of his that you quote. ("Experiments
performed by Petr Simanek" was his own third-person reference to himself.)
For the record, I disagree with him when he says,
S> Please note that you should not exceed the -6 dB recorded VOLUME.
One should try to get the volume of a digital recording as near as possible
to 0.0 dB without exceeding it. The exception comes when two or more tracks
form a suite and the relative volumes of one track to another are important:
then one should get the peak amplitude of the entire suite as near as
possible to 0.0 dB without exceeding it. DWT
S> HOW TO GET BETTER SOUND QUALITY OF YOUR MD RECORDINGS
S> (This is the result of experiments performed by Petr Simanek)
S> The following procedure applies for SONY JE520 and all newer SONY decks
S> with digital recording level adjustment (including the newest ones with
S> the Scale Factor Edit function). This applies for both digital and
S> analog recordings at such recorders.
S> For better understanding, please note the following:
S> When I write VOLUME, I mean the peak level shown by the LED indicator
S> bars during recording or playback.
S> When I write REC GAIN, I mean the recording gain set by the REC LEVEL
S> control and displayed as a number (e.g. -2.6 dB)
S> At the conventional CD (44.1 kHz, 16 bit resolution linear PCM) the
S> distortion is least at high volume near under the 0 dB threshold. At the
S> less volume the distortion increases due to quantization noise
S> (theoretically it doubles by every -6 dB).
S> When you listen to the audiophile test CD by 1 kHz at 0 or -6 dB you'll
S> hear no distortion. But when you listen to the same tone recorded at -60
S> or -80 dB, you WILL hear the significant distortion. Theoretically,
S> above the 0 dB threshold, the distortion will increase much more
S> dramatically, but no CD is recorded at that volume.
S> Unlike the CD, minidisc works with more resolution of 20 or 24 bits and
S> uses the floating point calculations. This results in the less
S> distortion at small levels.
S> A small example: If you have a test CD with 0/-20/-40/-60/-80 dB test
S> tones, play the 1 kHz / -40 dB tone and record it to the MD with the
S> -40 dB gain. This will result in a 1 kHz / -80 dB recording. Now play
S> the recorded MD and the test CD at -80dB simultaneously and you will
S> hear the difference. The MD has the less distortion and noise. At -60 dB
S> the result will be similar.
S> When you record to the MD a simple tone, e.g. 1 kHz from a tone
S> generator or from a test CD, the sound quality (purity of tone) is
S> almost constant at higher volume up to 0 dB. Of course, everything above
S> the 0 dB threshold will be very distorted due to clipping.
S> But when you record a complex musical signal up to 0 dB, it will be more
S> distorted than the music recorded only up to the -6 dB or -10 dB volume.
S> I think it is because the ATRAC has only half number of bits to allocate
S> at -6 dB than at 0 dB and it works easier and more accurate. There was a
S> topic "do you hear ATRAC?" on MD-L and I can say: Yes, I can hear ATRAC
S> on my equipment (MD:Sony MDS-JB930, CD: Sony CXP-JB630, Amplifier:
S> Harman/Kardon HK630, Speakers: 2-way systems with AUDAX High End
S> speakers, Headphones: Koss Porta Pro, Speaker Cables: Monitor Cobra
S> Silver 6mm bi-wiring).
S> Of course I know that only the peaks reach up to 0 dB but I think if you
S> decrease the volume of all recorded music by about -6 dB, you'll get the
S> clearer, less distorted sound.
S> AND NOW THE MOST IMPORTANT:
S> At the older MD decks with analog REC LEVEL control (no digital REC
S> LEVEL control possibility - e.g. SONY 303, 500 and 510) you can set the
S> analog recording gain anyway you want. But you cannot do that at SONY
S> 520 and all newer SONY decks with digital recording level adjustment.
S> The following applies for both digital and analog recordings at such
S> decks, because the analog REC LEVEL works after the A/D conversion.
S> When you change the recording gain digitally, the recorder has to
S> recalculate the bits that represent the instantaneous amplitude of audio
S> signal except some special cases described below.
S> When you set the gain to -6.0 dB /exactly 20*(log 0,5) = -6.0206 dB/,
S> the recorder don't need recalculate the bits . Because this value means
S> exactly one half of the original volume, the recorder only shifts all
S> the bits in the byte toward right without further recalculations. This
S> results in no additional distortion. Of course you'll lose one bit of
S> total 20-24 bit resolution (in the other words you'll lose 6 dB of total
S> S/N ratio), but it doesn't matter when you record from the 16 bit CD and
S> it is almost inaudible under normal circumstances.
S> Here is a small example:
S> 001001011101100110011100 is exactly one half of
S> 010010111011001100111000
S> MSB------------------LSB
S> This applies in the similar way also at -12.0 dB, -18.0 dB or +6.0 dB,
S> but I think you'll never need to use that values.
S> In most cases of recordings from CD I recommend you to use the -6.0 dB
S> gain. You should set it exactly to -6.0 dB, don't use the -5.9 dB or
S> -6.1 dB or other similar values, which cause distortion due to
S> recalculation.
S> You can try it. You'll need a test CD (supplied sometimes by audiophile
S> or High Fidelity magazines) with the 1 kHz test tone (sine wave) at a
S> very small level (-60 or -80 dB) recorded on it. You can also use a test
S> tone generator instead of the test CD.
S> Your CD should be connected to MD deck via digital cable (optical or
S> coaxial).
S> You'll need also a good headphones connected to MD deck's phones output
S> or a good amplifier connected to MD deck's analog output and a good
S> speaker system.
S> Insert any MD, press RECORD and set the REC GAIN to -6 dB. Now play the
S> test tone 1 kHz at -80dB (or -60 dB) from the test CD and listen
S> carefully. Now turn the AMS knob to change slightly the REC GAIN. The
S> tone you listen becomes more distorted. You'll see that any other REC
S> GAIN value than the listed below causes small, however audible
S> distortion.
S> Allowed values: +12.0 dB +6.0 dB 0.0 dB -6.0 dB -12.0 dB -18 dB
S> My recommended value: -6.0 dB
S> Please note that you should not exceed the -6 dB recorded VOLUME.
S> In some High Fidelity magazines it is written that the MDS-JA50 ES or
S> MDS-JA555 ES always sound slightly better than MDS-JB930. It would be
S> interesting to compare these recordings to the JB930 or 940 recording
S> with the -6 dB GAIN, but I have no access to that ES machines. Is there
S> anyone who can do it?
S> Best regards to all people on MD-L,
S> Petr Simanek, [EMAIL PROTECTED]
-----------------------------------------------------------------
To stop getting this list send a message containing just the word
"unsubscribe" to [EMAIL PROTECTED]
