DigitalAudioSoft - EQ Suite

sonolive · Post by **sonolive** » Sun Sep 24, 2006 3:21 am

On 2006-09-24 03:16, astroman wrote:
well, then let's have a look at <a href=http://www.gyraf.dk/gy_pd/pultec/pultec.htm>the real McCoy</a>
... now that they mention a 56k modem line separation transformer 'works and sounds' quite well, I will for sure build it - I have a small pile of that crap in the closet and always hoped it would be useful for something...

VERY VERY INTERSTING RESSOURCE INDEED, didn't know it , thx !

regarding the re-use of CWA's atoms

since there are basic 'abstract' math routines, a filter (for example) can be implemented in a number of ways - as shown by Adern.

I assume that the basic filter type by CWA is equivalent to an RC (resistor/capacitor) ladder, while a Pulteq type eq is LC (inductivity/capacitor) based.

The latter could be implemented by either: adding a math model of a true coil,
by applying a model that uses the OP amp 'subsitute' coil design,
or by altering the parameters of an existing eq to approach the originals response.

while the last point may read 'cheating' it is in fact quite useful - you might turn just one single knob on the surface, but with (the help of) an internal table it sets half a dozen parameters even according to complex rules.

cheers, Tom

FOR SURE, that's the point !

cheers
olive

Post by **hubird** » Sun Sep 24, 2006 8:58 am

I checked the test (high band) of Martin, and the result is 100 (full 100) percent phase cancellation.
It's amazing, full 100% cancellation

I can't judge anything technical here, but, as a (simple) consumer, having removed all the dust, I see this picture:

One. The phase cancellation test of Martin is a fact, there is no doubt about his measurements, right??
Zero signal from both phase inverted channels.

Two. The factor of 'specific combination of internal parts' is ruled out by testing more than one 'preset', right??
Zero signal etc.
Must say I didn't check this point, but I guess I can trust Martin in this.
EDIT: I checked also the second test, with two identical bands on the CW eq.
The picture of Martin shows a Q on the Poltq that shows the wrong position, should turned fully to the left, clockwise.
Only a very small high click left on the master out of my DynamicMixer.

Three.

On 2006-09-24 01:59, sonolive wrote:
[How could it be different ? We all (DP or SDK developpers) work on CW plateform developpement with CW atoms for programming SHARC DSP !
So IT MUST BE CW ATOMS Combinations !!! but all the interest of our (and others developpers) is THIS ! What combination for what sound !

So there's no discord about all developers using basic (S)DK CWA elements (as expected), right??

It really is 'the combination and way of applying the different building elements' that makes the difference, right?

Hm, then the 100% cancellation, even on one band, is rather strange.
What's then the reason to invest in the eq if the results with both eq's are the same?
Soundwise spoken that is

Would then the use of more bands at the same time make the difference between the stock eq and the Polteq?
I have some doubts, after the single band test, who will blame me for that

[ This Message was edited by: hubird on 2006-09-24 10:47 ]

next to nothing · Post by **next to nothing** » Sun Sep 24, 2006 9:38 am

well i guess the only alternative is to make a project that incorporats 3 peq's and a band splitter, hook it up like u did with the others examples and see if u get full cancellation using all bands simultanious.

Post by **hubird** » Sun Sep 24, 2006 9:46 am

yes, and I suspect the cancellation will still be there, tho not 100% anymore.
I leave this test to someone else

MCCY · Post by **MCCY** » Sun Sep 24, 2006 9:51 am

astroman · Post by **astroman** » Sun Sep 24, 2006 11:05 am

Olive, maybe we can agree also on a few other facts...

(oops a couple of posts have been made...)

you may have noticed that I made my own comparison with 'amusingly' strange results - or embarassing conclusions

, whatever you prefer.

just like on your system, the 2 Eqs did by no way cancel each other out.
Well, there's always something to learn - and in this particular case it's (at least to a degree) a rather important fact about how to structure a project.
At least that's the bottom line for me.

We should really leave the Polteq's internals out of the discussion.
Noone is interested to discredit your devices or DAS as a company - Martin has also pointed this out.
I always considered his comments more joking about people 'hearing by watching' than in your direction... anyway

A more economic ('straight to the point') and visually pleasing user interface IS a significant amount of work, and the device may contain a few sound sweeteners and internal optimizations...

Yet it cannot be denied that Martin is right in all those cases he examined.

I've repeated his setup with a fresh(!) project and all his predictions came true.
Originally I had simply added the devices to a more or less crowded 'current' project.
The more I tweaked and modified, the more it got messed...
Obviously the devices had been spread about any free 'DSP slot' available and the infamous sample delays became omnipresent.

I suspect that your results, which were also deviating from Martin's, have the very same source.
It doesn't help at all to put the most precise measurement gear at the end of the processing chain, when the 'system loader' spoils things right at the beginning

That is my only point here - it's not about your work or creativity or whatever - and it is something that EVERYONE can benefit from.
Since I've experienced similiar effects (with sample accuracy) a couple of times already (completely independant from DAS stuff), I am of course (please forgive me...) not unhappy that you seem to experience the same.
It tells me that it's not my humble old Pulsar One, which I suspected ...

To be honest, Martin explicetely pointed this out (to use a fresh project), as he had the very same effects on his 14/15 DSP card, when trying to phase cancel things in a 'full' project - so I'm confident my Pulsar One is ok.

Of course it's always a good idea to plan and structure a project from the ground up with focus on 'first things first', but there are enough situation where schedule (or just lazyness) suggest something else - and one adds the 'quick and dirty' modifications.

When accuracy is a concern it's a good idea to start from a clean base and verify from time to time that the system acts as it's supposed to - but of course only in THOSE cases, where it makes sense

If a track in a 'live' recording is 1-3 samples late doesn't matter, but a layered drum sound may not be that amusing, if sounds different the next day.

cheers and all the best, Tom

digitalaudiosoft · Post by **digitalaudiosoft** » Sun Sep 24, 2006 11:11 am

On 2006-09-24 10:51, MCCYRANO wrote:

In the moment the Algorithms (of both EQs) are distributed on different DSPs there occure problems with timing (delay of some samples) which leads to wrong phaseses which leads to impossibility of cancelling one with the other.

of course they are working on same dsp.as i say ,and ,you prove that ,only a newbie with sdk do this kind of error !
it's amazing ! now you want to modify phase to try to have right but you are wrong mister sdk newbie !!!

eric
________________
http://www.digitalaudiosoft.com

[ This Message was edited by: digitalaudiosoft on 2006-09-24 12:19 ]

Shroomz~> · Post by **Shroomz~>** » Sun Sep 24, 2006 11:45 am

On 2006-09-24 12:11, digitalaudiosoft wrote:
of course they are working on same dsp.as i say ,and ,you prove that ,only a newbie with sdk do this kind of error !
it's amazing ! now you want to modify phase to try to have right but you are wrong mister sdk newbie !!!

Eric, tip of the day:- try to stay cool & not insult people on a direct & personal level.

astroman · Post by **astroman** » Sun Sep 24, 2006 11:45 am

sorry Eric, you are wrong
Martin isn't talking about the EQ itself, but about the device/environment it's embedded in - the mixer/project routing

it should be mentioned that he worked out the procedure systematically in a very professional way - not only twice checking his results, but also revealing the 'flaws' of the test procedure.
the sideeffect of his findings regarding the Scope environment is rather significant imho

cheers, tom

[ This Message was edited by: astroman on 2006-09-24 13:05 ]

eliam · Post by **eliam** » Sun Sep 24, 2006 11:54 am

And tell us, hum...hum... mister seasoned developer, what's wrong with mister newb sdk developer's reasoning?

Why don't you simply adress the simple questions raised instead of attacking people's personal credibily and integrity (and compromising yours at the same time)?

spacef · Post by **spacef** » Sun Sep 24, 2006 11:59 am

i don't have the devices you test, but something that happens on digital systems (and in scope for sure) is cancellation by distortion... it can happen that a very high level with distortion just cancels any sound. it's not acoustical, but a digital phenomenon. it happenned to me several times during development of various devices...
may be you guys want to make sure it is not something like this.
I'm off that thread though, i didn't really follow, and i have no time to look into this precisely (sorry), and may be i'm wrong.

[ This Message was edited by: spacef on 2006-09-24 13:12 ]

digitalaudiosoft · Post by **digitalaudiosoft** » Sun Sep 24, 2006 12:00 pm

On 2006-09-24 12:54, eliam wrote:
And tell us, hum...hum... mister seasoned developer, what's wrong with mister newb sdk developer's reasoning?

Why don't you simply adress the simple questions raised instead of attacking people's personal credibily and integrity (and compromising yours at the same time)?

eliam , of course...

"and compromising yours at the same time..."

never mind ! i don't want to let stupid members writing sh..t post on das!

have you try to phase deleted the signal with all band activated ?
you don't understand like them and you make me tired.

so have a good night.

eric

i repeat again ...try with all band activated ,and after this,if your test is positive...without changing phase of course,you will have to do another test because it wont be the end only the begining !
________________
http://www.digitalaudiosoft.com

[ This Message was edited by: digitalaudiosoft on 2006-09-24 13:19 ]

diminu · Post by **diminu** » Sun Sep 24, 2006 12:30 pm

never mind ! i don't want to let stupid members writing sh..t post on das!

Hi, Eric
I really think, that you´ll better cool down, to avoid the impression, that´s something wrong with your product. In my opinion your postings are rather bad propaganda for the poltec.

Cheers
diminu

digitalaudiosoft · Post by **digitalaudiosoft** » Sun Sep 24, 2006 12:34 pm

On 2006-09-24 13:30, diminu wrote:
never mind ! i don't want to let stupid members writing sh..t post on das!

Hi, Eric
I really think, that you´ll better cool down, to avoid the impression, that´s something wrong with your product. In my opinion your postings are rather bad propaganda for the poltec.

Cheers
diminu

ok ...

_________________
http://www.digitalaudiosoft.com

[ This Message was edited by: digitalaudiosoft on 2006-09-24 13:35 ]

Post by **hubird** » Sun Sep 24, 2006 12:35 pm

Ok, we now have agreement on the 100% cancelation of one band

Pfew...

Btw Eric, even humble stupid me understood that Martin didn't mess around with the sample delay corrections.
He was just leveling out the delay caused by spreading the plugs over more than one DSP.
This is what Eliam says, 'and compromising yours at the same time)?'
Your personal approach blinds you, as also your 'colour' of sound remark in the other thread showed.

I would like to do the test of all bands, but that's one bridge too far for me.
To correct the DSP sample delays over several DSP's is one thing.
To translate the settings of five filters to the four of the CW eq isn't possible at all.

And what if four of them do cancelate...?
The fifth could make the difference...

Edit, missed your last 'ok', Eric, before submitting this post.

[ This Message was edited by: hubird on 2006-09-24 13:41 ]

digitalaudiosoft · Post by **digitalaudiosoft** » Sun Sep 24, 2006 12:43 pm

olive is right ,you are to stupid to talk with developpers.and i agree with him..
..and i'm stupid to answer those sh..t post too.

so now, you can say what you want...the only benefice is that everybody will see how stupid and jalous you are !

eric

astroman · Post by **astroman** » Sun Sep 24, 2006 12:58 pm

well, there are phase cancellation on 2 bands, Huub

but you don't seem to get the point of the sample delay yet - it has nothing to do with the Polteq or the number of bands.

SFP messes the signal flow of the routing, regardless if the device is 'inserted' in a mixer channel or just put on the surface.
Each EQ runs on it's own DSP (hopefully), but for the phase cancellation you need 2 devices...
if you cable them parallel, then you'd expect sample accuracy of the 2 signals
which it does in a 'fresh' project
but definetely not in a crowded one or one that is heavily tweaked - it changes over time - in my case (yesterday) it ended that not even 2 identical Eqs with identical settings cancelled each other and the phase coherency button on the mixer reversed it's function

THAT is the important fact of Martin's finding

cheers, Tom

[ This Message was edited by: astroman on 2006-09-24 13:59 ]

MCCY · Post by **MCCY** » Sun Sep 24, 2006 12:59 pm

Post by **hubird** » Sun Sep 24, 2006 1:14 pm

On 2006-09-24 13:43, digitalaudiosoft wrote:
olive is right ,you are to stupid to talk with developpers.and i agree with him..
..and i'm stupid to answer those sh..t post too.

so now, you can say what you want...the only benefice is that everybody will see how stupid and jalous you are !

eric

Post by **hubird** » Sun Sep 24, 2006 1:36 pm

On 2006-09-24 13:59, MCCYRANO wrote:
Is the following your thesis? :
'Poltec-only-sounds-different-to-PARA4EQ-when-all bands-are-activated-and-has-the-same-sound-when only-1-or-two-are-activated-(no-matter-which).' ?Is this your new hypothesis?

It is not so abnormal to use your EQ with just 2 or 3 bands turned on and for now it has been proven, that for such cases PARA4EQ has same sound. The results on single bands show that overall sound will not change most likely in other cases.

And this is exactly the situation where we are landed.
Personaly I would like to have the Polteq for the great GUI, compared with the spartanic CW eq, but not specially for it's sound character.
Till the counterpart has proven

[ This Message was edited by: hubird on 2006-09-24 14:37 ]