KVR Audio

autloc · Post by **autloc** » Sat Nov 05, 2005 6:50 pm

Nuffink - Finally checked out some processed impulse responses in Cubase -- look pretty symmetric to me.

bmanic · Post by **bmanic** » Sat Nov 05, 2005 7:29 pm

nuffink wrote: Whaddya say chaps? Anyone else think that a grand for a tone control deserves a few alternative views, free from the kind of smack down protection that the audiophiles are affording each other?

".. and it keeps going, and going and going *Insert M-16 burst sample here* .. ". Trivia: from what movie?

@David McClain, thanks for the update. The new graph mode will be very useful.

- bManic

gd_vm · Post by **gd_vm** » Sat Nov 05, 2005 7:42 pm

umm...I think its Hot Shots part duex......."war, its faaaaantastic"

rogeriomatos · Post by **rogeriomatos** » Sat Nov 05, 2005 8:05 pm

Aleksey Vaneev wrote:dbmcclain, if your EQ should give a symmetrical reponse (forward and backward IIR filtering) then I do not know why with a single dirac spike I do not get a symmetrical response as a result.

Moreover (and do not get me wrong - that's my opinion only), forward&backward IIR filtering using window overlap is not the best solution. Building a true linear-phase kernel (by evaluating IIR filter itself) and then using convolution to apply it to the sound should theoretically be more precise - but of course, you won't get any 'interesting' effects of windowed forward&backward IIR filtering.

Hey Aleksey, do you use/hear it? To me, 3 band free , is better than you eq plugs so, why worry about simetrical details if it sound wonderfull?

rogeriomatos · Post by **rogeriomatos** » Sat Nov 05, 2005 8:07 pm

And again, thanks dbmcclain!

bmanic · Post by **bmanic** » Sat Nov 05, 2005 8:08 pm

gd_vm wrote:umm...I think its Hot Shots part duex......."war, its faaaaantastic"

Damn, you are quick. Yes indeed!

- bManic

Aleksey Vaneev · Post by **Aleksey Vaneev** » Sat Nov 05, 2005 8:17 pm

rogeriomatos, as I've said, no problems here. If it sounds OK to you then use it. The final mix you get is what ultimately matters.

dbmcclain · Post by **dbmcclain** » Sat Nov 05, 2005 8:56 pm

Hi Aleksy,

Yes, you are correct that the "proper" thing to do for highest precsion would be true FIR filtering. But as I pointed out earlier the overhead would be unsustainable by most mortal computers... 7200 taps for every incoming sample!

We do it with 6 and are able to maintain our IMD to -120 dB from peak levels. That seems more than good enough, given loudness masking that will occur in anyone's ear.

Cheers,

- DM

dbmcclain · Post by **dbmcclain** » Sat Nov 05, 2005 9:00 pm

umm... so sorry... I didn't mean to misspell your name Aleksey. Cheers!

Christian Budde · Post by **Christian Budde** » Sat Nov 05, 2005 9:37 pm

Although i don't want to post again into this post, i must admit, that i was wrong. I took the version 1.60 and test it again. Most of the things i mentioned in my first post are gone. Here are som details:

Notch filters can't hardly be linear phase by design, but this looks good so thumbs up for this. At least 1kHz is 1kHz.

Bell curve is true up to nyquist, very good. Some phase-non-linearities remain, but that's due to endless design i guess..

Low frequencies are also ok. That's always a problem of FIR systems. If you do not use a too high Q, everything is fine. Otherwise the phase will get more non-linear and the gain will not be reached.
But that's due to design and they can't bend it, as i mentioned before.

So as a conclusion, this looks very good (only talking about quality 5!). For this the CPU time is ok, but i think it can be reduced with some assembler optimisations.

Now I got the point. Sorry about my scam.

Regards from Rio,

Christian

Frippertronix · Post by **Frippertronix** » Sat Nov 05, 2005 10:02 pm

Yes, David does his homework, and he has decades of signal processing expertise, which is probably as much or more than any developer on KVR can claim. Improvements are always possible, but no one can say he hasn't shown open mindedness about that!

And can we lay off about the high-end price tag? Do you guys huff and puff about every high priced piece of music equipment (of which there are hundreds)? Who cares if it's "just digital" if it does a high end job? If you want to develop something similar and charge 1/10th as much, I guess you can, but you could probably have done that already.

Thanks again, David, for the continuing improvements!

Dave

zck · Post by **zck** » Sat Nov 05, 2005 10:09 pm

bmanic wrote:
nuffink wrote: Whaddya say chaps? Anyone else think that a grand for a tone control deserves a few alternative views, free from the kind of smack down protection that the audiophiles are affording each other?

".. and it keeps going, and going and going *Insert M-16 burst sample here* .. ". Trivia: from what movie?

- bManic

yea, opinions of all sorts are good. Endless antagonizing is not.

Many thanks for the new 3 bands !

fritzman · Post by **fritzman** » Sat Nov 05, 2005 10:09 pm

Hi Christian,

thanx for your super fair posting and even with measurements! Lotsa character to do this. Big thumbs up!!!

All the best, FRitz

P.S. I still love Electri-Q in analog mode for tracking. Low CPU usage and great coloured (in the best sense) sound.

greendoor · Post by **greendoor** » Sun Nov 06, 2005 12:05 am

Small cosmetic feature with 1.60 (and probably earlier versions too): If a band is unloaded (none), the eq and graph are not updated with the change. It requires another band to be loaded, e.g. All Pass, to get the eq and graph to revert to flat.

Just a small detail in the interest of perfecting this fabulous eq.

aMUSEd · Post by **aMUSEd** » Sun Nov 06, 2005 12:18 am

I can't keep up - last time I looked there was a 1 band with gui and a 4 band without - where do you get this 3 band one? (with gui??) - can't see it on the site and the thread's getting too long to read every page.

10-Band PLParEQ is Here!