I wrote:

>I would say that around 99 % of digital EQ 
>available is of the Minimum Phase type that 
>exactly mimics analogue EQ as we are used to. 

James Howarth:

>Pretty close to agree except that there are 
> plenty of Linear phase EQs, 

Nowadays maybe but there was almost zero commercially Linear Phase EQs during the first two decades of digital audio.

The geat majority of digital EQ seen anywhere is Minimum Phase because the latency, time delay, through a Linear Phase EQ, makes it useless during overdubbing for example.

Also it needs much more processing power than Minimum Phase EQ so is a resource eater and this rules it out again for most applications.

Then we are all used to analogue Minimum Phase EQs so in many respects Digital Linear Phase EQ can be seen as a novelty item that just adds a different tool in the toolbox. But people used to Minimum Phase EQ  becomes a bit confused trying out Linear Phase digital EQ as it gives a different end result when twiddling the knobs.

So for the greatest compatibility then Minimum Phase EQ is the best bet so far.

Still the amount of Linear Phase EQ in most facilities, even today, is close to zero even if you can buy it from several sources today.

>and some more sophisticated recent equalizers 
>that will do all-pass networks allowing some 
>compensation for errors in the record side, in 
>our case NAB. 

"Some" compensation, yes.

Because lacking any knowledge about what the true phase rotation was on the recording machine then this will be anybodys pure guess how much delay compensation should be used on any tape.

Having tried this out extensively in the 70/80s by me I will say that the effect is mostly in the very subtle category if it is audible at all.

Keith O. Johnson did a paper about this in the JAES in the mid 60´s and concluded that it was of some help when doing multiple tape dubs as it avoided high frequency transient overshoots that caused self erasure on the magnetic tape.

But most of the time the audible effect is near inaudible.

>And some that sound different because they are 
>using insanely long FFTs and better because of it. 

I am thinking about how the actual code is done.

64 bit double precision was what the DECCA/London backroom boys decided was the minimum needed for transparent results using double blind listening tests back in 1984.

Thus it was used in all the digital mixing desks built by DECCA back then.

Most code written uses much less precision even today more than 30 years later on.

But at least the resources used will be much less..........and the audible degradation is not noticed by most people in fact.


Best regards,

Goran Finnberg
The Mastering Room AB

E-mail: [log in to unmask]

Learn from the mistakes of others, you can never live long enough to
make them all yourself.    -   John Luther

(")_(") Smurfen:RIP