KVR Audio

Christian Budde · Post by **Christian Budde** » Wed Jul 12, 2006 6:02 pm

diverdee wrote:I ran a comparison gliss eq demo vs. posifohpit a few weeks ago.
Difficult to judge fairly coclusively, given the different modes etc.

The difference in performance between posihfopit and "Electri-Q" is huge. That's basically a reason why they don't share the same name.

Aleksey Vaneev wrote:...and if I'm not mistaken, ElectriQ is 2 times less efficient than similar GlissEQ setup (at Dyn 0 of course).

I don't agree here. In fact it's very difficult to judge. Not only that both have different modes, also the CPU performance may differ on different machines (AMD, Pentium,...). On my machine (ok it's more or less optimized to run fast here) a nearly equal comparison results the Electri-Q is about 2-times faster. But as Aleksey mentioned the GUI may cause some problems here. So far my test setup was the follow:

GlissEQ2 (latest version from the homepage): Peak EQ 200Hz, Gain 10dB BW=1Okt, Dyn0, Analyser: Blocksize 8192, Spd=8, Slope=0, Full, Normal Setting, Mode 1

Electri-Q (latest version, 1.5.3): Peak EQ 200Hz, Gain 10dB, BW=1, Digital EQ (=no oversampling), Analyser: active (internal Blocksize 8192)

Measurement: Time for one single call of ProcessReplacing() with 44100 samples stereo data (priority set to realtime)

Results (AMD64 3000+): GlissEQ2: 18ms (~1.8% CPU Usage) Electri-Q: 10ms (~1.0% CPU Usage)

If I close the GUIs or disable the analyser (I don't know how in GlissEQ2) then the CPU performance of Electri-Q decreases heavily.

Kind regards,

Christian

jmbriley · Post by **jmbriley** » Wed Jul 12, 2006 6:31 pm

I got a new title for this thread. Christian vs.Aleksey

martian · Post by **martian** » Wed Jul 12, 2006 6:31 pm

sounddesigner wrote:After doing a more thorough and "double blinded" test with similar settings on both eq's it appears to be not much of a difference between gliss eq and electri-q so i deleted much of my last post. out of curiousity i plan to test them both later once i rest up my ears.

if you compare digital mode vs dyn 0 then yes theres not much difference. but assuming you use their respective modes, electriq and gliss sound very different.

i havent properly tried electri-qs transient filters yet though, but given you have no control over it im imagining its going to be quite a different concept.

Aleksey Vaneev · Post by **Aleksey Vaneev** » Wed Jul 12, 2006 9:08 pm

Christian Budde, running ONE band is a meaningless test. Enable all 5 bands in both EQs. And of course, close the UI. Projects usually do not have dozens of UIs open, so a plain EQing performance is important.

martian · Post by **martian** » Wed Jul 12, 2006 10:39 pm

well they are both (gliss 2.7 and electri-q 1.53) equal here with 6 instances, each 5 bands active, digital and dyn0 modes, no oversampling, guis closed.

each 6 instances use 3% cpu here, on one core of an intel dualcore system.

Christian Budde · Post by **Christian Budde** » Wed Jul 12, 2006 10:51 pm

martian wrote:well they are both (gliss 2.7 and electri-q 1.53) equal here with 6 instances, each 5 bands active, digital and dyn0 modes, no oversampling, guis closed.

each 6 instances use 1.5% cpu here. intel dualcore, both cores active.

That's what I was guessing on normal machines (i.e. not the machines of the developers). I think it should be quite equal because we're both using assembler optimized parts. I don't mind if my plug would be slightly slower, but I just don't agree, that my plugin should be two times slower. Although the plugin is optimized for speed it is also optimized for flexibility, that's why I guess it's maybe not the fastest, but very fast.

Anyway in case of Gliss vs. Electri-Q I think both plugins do have advantages/disadvantages and I think CPU performance is an advantage of both plugins, no matter which of both is faster. Having measured several plugins on the market, I can say that there are several, which run a lot slower.

Aleksey Vaneev · Post by **Aleksey Vaneev** » Thu Jul 13, 2006 12:01 am

Christian Budde, it is not that I care much about comparison in this case. But since somebody said GlissEQ is more CPU expensive I had to check it out, because as far as I know, for double precision processing GlissEQ is a very optimized equalizer (in Dyn 0 mode).

BTW, I was checking in 'Analog' mode.. 'Digital' mode is only like 5% slower in Electri-Q. So, no sense to compare.

Once again, it is all about modes.

martian · Post by **martian** » Thu Jul 13, 2006 12:03 am

just for the record:

5 active bands, no oversampling, 6 stereo instances, guis closed, sx3.1.1, intel dualcore (only 1 core active).

[waves ren 3.5 dirx] 2.5%
[glisseq 2.7 dyn0] 3%
[electri-q 1.53 digital mode] 3%
[posifopit 1.36] 5%
[geq 1.12] 6%
[aireq 1.04 sse2] 8%

Zaphod (giancarlo) · Post by **Zaphod (giancarlo)** » Thu Jul 13, 2006 7:27 am

It depends on cpu... I have plugs that on some machines are fast (athlon) and on pentium are slow and viceversa. Not only for de-normalization. For example a Centrino works VERY bad with SSE2, better FPU somethimes!! Because FPU was optimized a lot.
So a run on the 'developer' configuration makes no sense (developers optimized HIS cpu behaviour)
Only 1 thing I don't understand. Why the 2 creators are inserting posts here in public. I think this damages the opinion on their product (IMHO when a developers try to defend his products it SEEMS he reduces everything to a race). I think they are 2 different products; like Christian said on a different topic, it isn't orange with orange. BUT if YOU accept this topic, you should compare the RESULT using a product or the other.
NOW. Users makes a lot of topic similar, because they have a small amount of money, so they SHOULD make a choice. A magazine puts pro e cons EVERY time. Disabling a feature for comparing oranges with oranges DOESN'T make sense here. When I buy Audi I don't make a blind test on the engine only. Every feature is IMPORTANT for the choice

Oliver.Lucas · Post by **Oliver.Lucas** » Thu Jul 13, 2006 8:07 am

Seriously guys, does anyone really care if it uses 3% or 4% CPU?
Fact is: Both developers seem to have succeeded in writing very efficienct code.

It seems to come down do personal taste in the end. I really love the speed and simplicity of Christians GUI approach. Flexible but NOT overloaded and distracting. The sound is brilliant, it really gives you amazing results!
On the other hand the dynamic EQ option in GlissEQ is nice to have, I guess. (I cannot seem to get any usable results out of the transient dynamic mode in Electri-Q, everything sound very random to me).
So maybe Christian will work on a proper dynamic function and Alexey will sort out the GUIs and both are one step further....

If you ask me they both already are (compared to the so called industry leader)

kmonkey · Post by **kmonkey** » Thu Jul 13, 2006 8:42 am

4damind · Post by **4damind** » Thu Jul 13, 2006 8:51 am

Sorry, but I find the GUI of Electri-Q horrible. Electri-Q is no competitor to GlissEQ. But GEQ7 from Kjaerhus and GlissEQ playing on the same high level.

defjamm · Post by **defjamm** » Thu Jul 13, 2006 9:13 am

4damind wrote:Sorry, but I find the GUI of Electri-Q horrible. Electri-Q is no competitor to GlissEQ. But GEQ7 from Kjaerhus and GlissEQ playing on the same high level.

kjaerhus golden eq

voxengo gliss

really high level?

fgimian · Post by **fgimian** » Thu Jul 13, 2006 9:24 am

defjamm wrote:
4damind wrote:Sorry, but I find the GUI of Electri-Q horrible. Electri-Q is no competitor to GlissEQ. But GEQ7 from Kjaerhus and GlissEQ playing on the same high level.
kjaerhus golden eq

voxengo gliss

really high level?

hey this is really fascinating, would you care to produce graphs of a few more EQs ? For instance, the Electric-Q and also aireq ?

Just a request, carry on

Fots

Stratification · Post by **Stratification** » Thu Jul 13, 2006 9:32 am

4damind wrote:Sorry, but I find the GUI of Electri-Q horrible. Electri-Q is no competitor to GlissEQ. But GEQ7 from Kjaerhus and GlissEQ playing on the same high level.

Just goes to show how much of this is subjective. Electri-Q is one of the best looking GUI's that I've seen - once you've changed it to the original blue, that is

GlissEQ vs. Electri-Q