dbx type II Tape Nose Reduction - DSP plugin

Archived from groups: rec.audio.pro (More info?)

While you're waiting for the info, start working on an RMS level sensor. I
don't know the attack and release constants, however.

Archived from groups: rec.audio.pro (More info?)

Hi TM,

Please post when you get something working. I have a few old tapes I've
transferred to hard disk that were done with DBX. It would be nice to
be able to play back. Even if not perfect, the ability to play them
back somewhat accurately would be useful.

You probably already know all of this, but I was told way back when
that DBX was a companding system that used a 2:1 ratio for compression
during encode and expansion during playback. I was also told there was
a hi frequency boost during encode that was inverted during playback. I
don't know what that curve looks like. I'm guessing other people here
can give more details that will at least get you going in the right
direction.

Good luck, and please keep me posted.

Dean


TechnoMonster wrote:
> (cross posted in rec.audio.misc )
> Hello,
>
> I am preparing to work on a software DSP plugin that will decode dbx
> Type II encoded tapes. So many people have tapes that were encoded
> with dbx type II, and there are so few hardware units floating around,
> and so I figured, why not?
>
> What I need are detailed specifications on dbx type II. Perhaps if I
> had the U.S. Patent numbers for the Model 222 or 224, I could use one
> of the many patent search services to likely get what I need.
>
> If anyone has detailed specifications and a schematic, that would be
> great! I plan to make this decoder available as a Winamp plugin for
> free.
>
> I had sent email to DBX Pro Support, and they tried to help, but just
> did not retain that information when that consumer division went out of
> business over 10 years ago.
>
> Thanks! TechnoMonster

Archived from groups: rec.audio.pro (More info?)

Hi TM,

Please post when you get something working. I have a few old tapes I've
transferred to hard disk that were done with DBX. It would be nice to
be able to play back. Even if not perfect, the ability to play them
back somewhat accurately would be useful.

You probably already know all of this, but I was told way back when
that DBX was a companding system that used a 2:1 ratio for compression
during encode and expansion during playback. I was also told there was
a hi frequency boost during encode that was inverted during playback. I
don't know what that curve looks like. I'm guessing other people here
can give more details that will at least get you going in the right
direction.

Good luck, and please keep me posted.

Dean


TechnoMonster wrote:
> (cross posted in rec.audio.misc )
> Hello,
>
> I am preparing to work on a software DSP plugin that will decode dbx
> Type II encoded tapes. So many people have tapes that were encoded
> with dbx type II, and there are so few hardware units floating around,
> and so I figured, why not?
>
> What I need are detailed specifications on dbx type II. Perhaps if I
> had the U.S. Patent numbers for the Model 222 or 224, I could use one
> of the many patent search services to likely get what I need.
>
> If anyone has detailed specifications and a schematic, that would be
> great! I plan to make this decoder available as a Winamp plugin for
> free.
>
> I had sent email to DBX Pro Support, and they tried to help, but just
> did not retain that information when that consumer division went out of
> business over 10 years ago.
>
> Thanks! TechnoMonster

Archived from groups: rec.audio.pro (More info?)

Marc Wielage <mfw@musictrax.com> wrote:
>Hey, if you'd like me to loan you a dbx II decoder to play with, I'll be glad
>to loan one to you, provided you pay for the shipping (both ways).

You _need_ to do this if you are developing a plug-in, because although
you can model the dbx system as a simple compander (a first-order device
whose current output is related only to the current input and not to any
previous inputs), the reality of the actual box isn't so nice, and you
may want to model some of the higher-order effects to help undo them.
--scott


--
"C'est un Nagra. C'est suisse, et tres, tres precis."

Archived from groups: rec.audio.pro (More info?)

> Bill, one further complication is that when recording in both dbx
> systems (I and II), the signal was split at each input and pre-emphasis
> was applied to the two branches. Then an rms detector in the one branch
> was used to set the gain of the other branch. However, the pre-emphasis
> in the two branches wasn't the same--the high frequencies in the
> control branch were boosted more than they were in the main signal
> branch of the circuit.

> The dbx II system had less pre-emphasis in the control branch than dbx
> I so that in playback, there would be less gain error if the recording
> system had high-frequency response errors (e.g. due to head
> misalignment, less-than-optimal bias, or the use of slow-speed
> recording).

The first paragraph is correct, but overly complex. The second is only
partly correct.

These are the components of dbx processing when recording:

1. Treble boost is applied to the input signal.

2. The treble-boosted signal travels to both a VCA and an
RMS level detector.

3. The RMS level sensor includes a bandbass filter (20-20K for type I,
30-15K for type II) and additional treble boost.

The bandpass filter is supposed to reduce sensisitivity to level
variations caused by the recorder attenuating the frequency extremes.
(As it's applied only to the level sense, it has no effect on the
system's
frequency response -- unless you're recording single tones.)
The extra treble boost is intended to make the level sensor more
sensitive to transients.

4. The voltage from the level sensor controls the VCA.

It must be pointed out that dbx is a single-band system. There is no way it
can alter the spectral balance of any common musical signal, because errors
show up as gain variations across the full audible spectrum. For dbx to
exaggerate the recorder's response errors, you'd have to record a single
instrument whose notes fell in the range of a non-flat region of the
recorder's response curve.

I found it amusing that magazines tested dbx-equipped decks with single
frequencies, which is meaningless. This is even true (though to a lesser
extent) with multiband systems. The "correct" way to test is with broadband
noise that has a spectrum similar to acoustic music.

By the way, I once dubbed "The King James Version" LP to a Nakamichi 700 II
with both Dolby B and dbx II. Much to my surprise -- considering the large
amount of treble boost the dbx applied -- with dbx I was able to push the
record levels (as shown on the Nak's meters) much higher into the red before
the sound became dull or distorted. I still don't understand that.