REW AND AUDIO ARCHITECT

[NATHYDREAD]

Hi,
I just bought a Crown CDI 4 300 that includes DSP run by audio architect. Which Default equaliser should I point in REW 's preferences for this soft ? I tried mini DSP 2x4HD but doesn't no work well...
Thanks,
David

 

[Wayne A. Pflughaupt]

With an equalizer that REW doesn’t support, it’s best to just use REW’s pink noise feature and adjust filters in real time.

Regards,
Wayne

 

[Matthew J Poes]

Hi,
I just bought a Crown CDI 4 300 that includes DSP run by audio architect. Which Default equaliser should I point in REW 's preferences for this soft ? I tried mini DSP 2x4HD but doesn't no work well...
Thanks,
David

I actually will disagree with Wayne on this. The Crown dsp is standard IIR filters so use the generic. You can also manually add filters to REW to use some of Harmans other filter options like shelf filters and such. It's great software, enjoy! This is what I use.

 

[Wayne A. Pflughaupt]

Different equalizers calculate filter bandwidth differently – that’s why REW has selections for accurately supporting different equalizers. So there’s a good chance the generic setting won’t work any better for you than the one for the miniDSP. Can’t hurt to try though. With an outboard equalizer you could do a loopback to get response characteristics and then model a comparable filter with REW's different equalizer settings (like I did below), to see if one of them came close to matching. Tough to do with a built-in DSP though...

Behringer DSP 1124
1/3-octave (20/60)

Behringer FBQ2496
1/3-octave (.333)

Velodyne SMS-1
1/3/octave (4.3Q)​

Regards,
Wayne

 

[NATHYDREAD]

thanks you two...
I will try the generic (haven't seen it acyually before you pointed it)...
geee ! haven't seen this one coming before buying the amp ! looks like I will have to learn digital equalisation...
let s hope it is easier than passive equalising !

could using rephase be a way around this pb ? as the unit is FIR capable ...

Or could mister REW create it ? audio architect is certainly more used than mini DSP, isn't it ?

 

[John Mulcahy]

The USM-810 is a Crown unit, could try selecting that.

 

[NATHYDREAD]

many thanks sir, will certainly try tomorrow

 

[Matthew J Poes]

The USM-810 is a Crown unit, could try selecting that.

That's good to know.

Since the Harman software doesn't have 1/3 octave equalization and instead relies on standard PEQ filters wouldnt the generic work just as well. You set amplitude, frequency, and Q seperately as usual. Are there issues with how they define Q?

 

[Matthew J Poes]

thanks you two...
I will try the generic (haven't seen it acyually before you pointed it)...
geee ! haven't seen this one coming before buying the amp ! looks like I will have to learn digital equalisation...
let s hope it is easier than passive equalising !

could using rephase be a way around this pb ? as the unit is FIR capable ...

Or could mister REW create it ? audio architect is certainly more used than mini DSP, isn't it ?

I'm sure it is far more widely used, but only professionally. I don't know how wide spread REW is in the commercial venues department.

I would love to know if you find that your amp can take an FIR filter directly. I had thought that in the past those filters were locked ans only Hamran could develop them internally. I didn't think it was possible to upload a filter and convolve it.

 

[John Mulcahy]

Are there issues with how they define Q?

Yes, there are several ways to interpret a Q or a bandwidth figure depending on the equaliser selected.

 

[Wayne A. Pflughaupt]

...As clearly shown in the pictures in my previous post. All three are 1/3-octave filters.

Regards,
Wayne

 

[Matthew J Poes]

Yes, there are several ways to interpret a Q or a bandwidth figure depending on the equaliser selected.

If I share a screen shot would you be able to tell if their implementation of Q matched your generic model well enough?

 

[Matthew J Poes]

...As clearly shown in the pictures in my previous post. All three are 1/3-octave filters.

Regards,
Wayne

Yes that's why I asked about PEQ. If the filters are fully definable what would differ from the generic? Frequency is frequency so the issue would be gain and q. Gain is usually a reference to dbs and Q has a standard defintion. I had thought bandwidth was the typical problem child with everyone defining this slightly differently.

I'll post a screen shot tonight and maybe you and John can see if you agree with my assessment that generic matches well enough. I've been doing it that way for quite a while and it always seemed to work right.

 

[Wayne A. Pflughaupt]

You're luckier than I was, Matthew! If I recall, I used the generic setting with my Yamaha YDP2006 equalizers and it didn't work well. I resorted to the RTA, which is why I typically recommend that when people ask about equalizers REW doesn't support.

Getting ahead of the curve (no pun intended), here's what the generic EQ setting gets for a 4.3 Q, 60 Hz filter cut 12 dB.

Regards,
Wayne

 

[Matthew J Poes]

You're luckier than I was, Matthew! If I recall, I used the generic setting with my Yamaha YDP2006 equalizers and it didn't work well. I resorted to the RTA, which is why I typically recommend that when people ask about equalizers REW doesn't support.

Getting ahead of the curve (no pun intended), here's what the generic EQ setting gets for a 4.3 Q, 60 Hz filter cut 12 dB.


View attachment 6156


Regards,
Wayne

That's good to know. In the time that REW has existed with EQ calculations I've only owned equalizers that are either supported by REW or seemed to work fine with the generic setting. The only exception I can think of is the single band PEQ in the Dayton plate amplifiers but I racked that up to inaccuracy in the controls. I vaguely recall there being a known issue with Q as well.

With the Crown amps and Harman software, which I've had since something like 2010 or so, it's always been fine. I've generally had very good results.

I do get issues with shelf filters. The final calculated results for PEQ aren't right unless the shelf filter is in place at the start. I racked that up to a better order of operations.

 

[andyc56]

I do get issues with shelf filters. The final calculated results for PEQ aren't right unless the shelf filter is in place at the start. I racked that up to a better order of operations.

Were you using a miniDSP device, and if so, were you using biquad file import in this situation, or manual filter entry? I've found there are at least two conventions for specifying the "frequency" of e.g. an LF shelving filter. For an LF shelf, one convention is the LF corner, used by Behringer in the DCX2496 among others. Another convention is the "half-boost frequency": that is, if the ultimate LF boost were, say, 6 dB, the "frequency" of the LF shelf is sometimes specified as that for which the boost is half the maximum, or 3 dB in this example. In the case of miniDSP hardware, if biquad file export from REW is used for import into the miniDSP, this convention should become irrelevant.

 

[NATHYDREAD]

The USM-810 is a Crown unit, could try selecting that.

I tried it today and it did work fine ! and cherry on the cake, instead of having to fill in 8 or 10 frequency correction points I had only 4 with this equiliser ! (the crown has 8 bands input and 8 output)...
I am so happy ... i ve donated to REW ...! (should have done it long ago)

 

[NATHYDREAD]

I'm sure it is far more widely used, but only professionally. I don't know how wide spread REW is in the commercial venues department.

I would love to know if you find that your amp can take an FIR filter directly. I had thought that in the past those filters were locked ans only Hamran could develop them internally. I didn't think it was possible to upload a filter and convolve it.

Not sure you can upload it as in a miniDSP, I just thought rephase could output the numbers and i would have to enter them manually into the Crown...but the soft is complexe and i am only using 10 % of it if not less
It is a great unit by the way, cost effective, and "improved" class D (no harshness).... I ve done one tweeter so far.... and i am impatient to go on, cause that is the bet here: what i would loose in musicality i will compensate with proper equlisation and room curve...and learning and fun....
Results were good with mini DSP on the JBL 4425
Thanks for the help

 

[Matthew J Poes]

Were you using a miniDSP device, and if so, were you using biquad file import in this situation, or manual filter entry? I've found there are at least two conventions for specifying the "frequency" of e.g. an LF shelving filter. For an LF shelf, one convention is the LF corner, used by Behringer in the DCX2496 among others. Another convention is the "half-boost frequency": that is, if the ultimate LF boost were, say, 6 dB, the "frequency" of the LF shelf is sometimes specified as that for which the boost is half the maximum, or 3 dB in this example. In the case of miniDSP hardware, if biquad file export from REW is used for import into the miniDSP, this convention should become irrelevant.

Hi Andy, REW can't communicate with Harman Audio Architect nor can it take direct input of biquads.

The problem I ran into was that I used REW to come up with the PEQ settings for a flat response. After I had a flat response I wanted to add an upward tilt toward 20hz and added a shelf filter that increased at about 6db's per octave with 6db's of gain centered at 80hz. I added this after having done the PEQ to smooth the response as it was an afterthought. Once I added this I found the response was no longer smooth. The mode related anomalies returned essentially. I then removed the PEQ filters and left the shelf filter, remeasured, and recalculated the correction filters. This time, I got good results.

Here are some screen shots from Audio Architect

Example EQ, see image for Q, Frequency, and Gain settings and graphic of shape.


Shelf filter

 

[Matthew J Poes]

Not sure you can upload it as in a miniDSP, I just thought rephase could output the numbers and i would have to enter them manually into the Crown...but the soft is complexe and i am only using 10 % of it if not less
It is a great unit by the way, cost effective, and "improved" class D (no harshness).... I ve done one tweeter so far.... and i am impatient to go on, cause that is the bet here: what i would loose in musicality i will compensate with proper equlisation and room curve...and learning and fun....
Results were good with mini DSP on the JBL 4425
Thanks for the help

Yeah I've used the CDi's before, but never in my own system. They are pretty awesome amplifiers. I don't feel the DSi is acceptable for full range use, but its a great subwoofer amplifier. I've tried it and found the sound to be harsh and there is a small amount of hiss with efficient speakers. The only issue I've had with the Crown amplifiers is that the plastic parts haven't always held up. My DSi amplifier switch broke and I had to permanently turn it on. A replacement part is hard to obtain. I've had past Crown amps with fans go bad and other plastic parts failures. It's typically similar issues, even with the fans, I've noticed cracking of the parts. Not sure why, I'm guessing heat.

 

[NATHYDREAD]

Yeah I've used the CDi's before, but never in my own system. They are pretty awesome amplifiers. I don't feel the DSi is acceptable for full range use, but its a great subwoofer amplifier. I've tried it and found the sound to be harsh and there is a small amount of hiss with efficient speakers. The only issue I've had with the Crown amplifiers is that the plastic parts haven't always held up. My DSi amplifier switch broke and I had to permanently turn it on. A replacement part is hard to obtain. I've had past Crown amps with fans go bad and other plastic parts failures. It's typically similar issues, even with the fans, I've noticed cracking of the parts. Not sure why, I'm guessing heat.

Well I had the D75 class AB, and now this 2017 award winning CDI drivecore 4 300.... it is not a mark levinson but no harshness at all as I said is it the one you know ? https://www.crownaudio.com/en/product_families/cdi-drivecore-series-amps

 

[Matthew J Poes]

Well I had the D75 class AB, and now this 2017 award winning CDI drivecore 4 300.... it is not a mark levinson but no harshness at all as I said is it the one you know ? https://www.crownaudio.com/en/product_families/cdi-drivecore-series-amps

Yeah it's a good amp. The DSi that I have is an older architecture with an older class D amp topology. It isn't drivecor.

There have been some complaints over hiss from the CDI and while I didn't have that experience I also never used it in my system or with mains. I used it with subs. Other drivecor amps have measured well and Crown typically uses a variation of a single topology in their amps. Today I believe that have about three unique topologies but shared across more than 10 amp lines.

 

[NATHYDREAD]

Hi,
well talking about hiss.... indeed without the crossover i got hiss... I have made a pad first -6dB, now -10dB but still hiss, and I cannot increase the padding at this point so ...a bit puzzled here as these pads were working well with the miniDSP and a sono crown used as test amp for its power... the latter should be noisier....so

Second point , it seems i have been too optimistic regarding DSP.... But it might be a Crown issue so off topic, may be...

The thing is I can equilise fairly good, i can make the xover all right on each driver, but when i try to measure the whole speaker i got a very different answer than the sum of the parts.... see for your self...

I am puzzled as I had none of this with the mini DSP.... and beyong eache driver measures kind of allright....

Precision: each driver measurement is done at 1 meter from the driver 0 °. The speaker measurement is also at one meter but with the mic at mid hight and pointing a little above the woofer...
I tried to inverse phase ....bigger dip....so
any clue ?

 

[Matthew J Poes]

Hi,
well talking about hiss.... indeed without the crossover i got hiss... I have made a pad first -6dB, now -10dB but still hiss, and I cannot increase the padding at this point so ...a bit puzzled here as these pads were working well with the miniDSP and a sono crown used as test amp for its power... the latter should be noisier....so

Second point , it seems i have been too optimistic regarding DSP.... But it might be a Crown issue so off topic, may be...

The thing is I can equilise fairly good, i can make the xover all right on each driver, but when i try to measure the whole speaker i got a very different answer than the sum of the parts.... see for your self...

I am puzzled as I had none of this with the mini DSP.... and beyong eache driver measures kind of allright....

Precision: each driver measurement is done at 1 meter from the driver 0 °. The speaker measurement is also at one meter but with the mic at mid hight and pointing a little above the woofer...
I tried to inverse phase ....bigger dip....so
any clue ?

Are you relying on the visual appearance of the overlap to predict the summed response? If so then what you are getting is exactly what you expect. The drivers still have substantial overlap that causes interference that leads to the apparently response. This isn't how you want to design the crossover.

You should download a crossover software for this instead. An active crossover designer is ideal but other options exist. I use passive designers for active because many (like bagbys) let you also simulate active. I suggest taking a response measurement and impedance measuremeny and loading them into the software (they need to be converted to frd and zma files). The impedance isn't as crucial for active but it's often good to add a zobel filter and some padding or protection to the tweeter and you want to model that. I also suggest using the polar response in your design. Take measurements in 5 degree increments. Only looking at the axial response can lead to a bad off axis response.

I promise this isn't a DSP or crown issue. What you got is exactly what one would expect. This is all just a learning process so don't sweat it.

 

[NATHYDREAD]

thanks for the confort..)
I tries to mimick the passive crossover: as far as i understand it on the JBL 4430, it is LR 12dB/o at 1000 Hz....I tried this first and then 36dB/o.... but it s a releaf to read it s my bad filtering.... I didn't know they were such tools as crossover softwares... I can provide the impedance curve of the woofer as I have a DATs V2 thingy....got to see if it is "loadable"...

My poor capabilities in active filtering were not so much a pb with mini DSP, in other words the xover were not generating such chaos.... well let's take the learning road, then...
Many thanks, indeed!