Yeah, I had sort of feeling your ARTA measurements are not calibrated the right way.
With 16bit measurement resolution your traces must be waaaay higher - even for a clean 24bit measurement plots are too low !
- did you use averaging ?
In general its no good idea to measure a 24 bit device with a a 16 bit soundcard - so I'm a little bit lost about the validity of plots shown - possibly have to check out myself
Did not get your point ? - do you say the transformer in your plots (multi-tone measurement stock versus transformer ) is driven at a different level – though plots show roughly same levels?
How much was the difference in drive level then?
Confused...
Michael
Last edited:
Have to correct myself.
We are roughly ~ 80dB below signal level up from 200 Hz with that transformer – by no way 24 bit !
Possibly down at ~ 100-110dB up from 1kHz - if I extrapolate distortion linearly.
Michael
Last edited:
The reason I used 48KHz/16 bits is so that Windoze won't resample and recalculate everything.
I'd use 24 bit mode if the silly soundcard drivers would allow it. They do in ASIO, but not MME or DS. As I don't have the pay version of ARTA, I can't use the ASIO function. I do use ASIO in HOLMImpulse
All drive levels for all ARTA tests were the same. -6dBFS from the computer. It's the outputs that needed attenuation to match the sound card inputs. Just the built in pad and/or volume knob were used to attenuate.
Linear averaging was used on all plots - 20 passes. Kaiser5 window. FFT 65536
I'd use 24 bit mode if the silly soundcard drivers would allow it. They do in ASIO, but not MME or DS. As I don't have the pay version of ARTA, I can't use the ASIO function. I do use ASIO in HOLMImpulse
All drive levels for all ARTA tests were the same. -6dBFS from the computer. It's the outputs that needed attenuation to match the sound card inputs. Just the built in pad and/or volume knob were used to attenuate.
Linear averaging was used on all plots - 20 passes. Kaiser5 window. FFT 65536
I am the proud owner of a brand new in the box DCX-2496. For some time I have been using a DEQ-2496 Real Time Analyzer followed by a 3-way Rane analog crossover that feeds a Rane MA-6 , six pack amplifier. The speakers are Linkwitz style open baffle panels. Please do not tell the single end tube guys.
I am pulling the Rane crossover out and putting in the digital crossover. I want to possibly keep the DEQ-2496 in the loop, it is dialed to the equalization required the open baffle speakers. I also like being able to plug in the Behringer microphone and check out the speaker and room performance.
Do any of you long time users have any thoughts, ideas or recommendations? And a related question , will the DCX-2496 do the needed open baffle equalization without the RTA in the loop?
Thank You
DT
All just for fun!
I am pulling the Rane crossover out and putting in the digital crossover. I want to possibly keep the DEQ-2496 in the loop, it is dialed to the equalization required the open baffle speakers. I also like being able to plug in the Behringer microphone and check out the speaker and room performance.
Do any of you long time users have any thoughts, ideas or recommendations? And a related question , will the DCX-2496 do the needed open baffle equalization without the RTA in the loop?
Thank You
DT
All just for fun!
Hey DT we won't tell the tube fans you're here. 
The DCX can do almost all of what the DEQ does - and more. Tho I'd rather do room EQ with the DEQ - it's just easier and more powerful. I do think that the DCX will be fine for your open baffle rig, it worked for me on several I built.
I use a DEQ in front of the DCX for a bit of room EQ and also just for the pretty RTA graphics. I found that the better and better I got at using the DCX and building crossovers, the less I needed the DEQ. But it was a wonderful tool for getting there.
Bottom line, they can work well together. Just come digital (AES - balanced) out of the DEQ into the DCX. Easy.
The DCX can do almost all of what the DEQ does - and more. Tho I'd rather do room EQ with the DEQ - it's just easier and more powerful. I do think that the DCX will be fine for your open baffle rig, it worked for me on several I built.
I use a DEQ in front of the DCX for a bit of room EQ and also just for the pretty RTA graphics. I found that the better and better I got at using the DCX and building crossovers, the less I needed the DEQ. But it was a wonderful tool for getting there.
Bottom line, they can work well together. Just come digital (AES - balanced) out of the DEQ into the DCX. Easy.
Caution! In the DCX the microphone only works for measuring the delay between speakers. It does NOT measure the frequency response of room or speaker.
Hi Panomaniac,
did the multi-tone measurement with my unmodified soundcard – here is what I found:
1.) ARTA shifts down 0dBFS to the –20dBFS line in multi-tone measurement – bug or feature ?
2.) As expected – If you change your output signal other than with a pure voltage divider there is no guarantee that the observed differences in performance are from the insertion of a transformer.
To show how the output stage (2 opamps after the DAC in case of my soundcard) produces different results below two measurements
– first one with digital signal OUT at –3dB
– second with digital signal OUT at –13dB
Output digitally attenuated by 3dB :
Output digitally attenuated by 13dB :
For both measurements input gain of the soundcard was adjusted to provide roughly same resolution ( ~ -10dB record level).
We see that the hasch at the top end gets increased when approaching 0dB digital attenuation.
Depending on design this may or may not clear up at the levels you have measured at with the transformer variant.
Michael
Last edited:
Interesting. This is just a loop back on your soundcard, right?
Right. That is why I tried to keep the digital level always at -6dB and used the analog volume control (pot) on my soundcard for level adjustments.
Where was this attenuation done? In Arta or somewhere else?
Hey DT we won't tell the tube fans you're here.
The DCX can do almost all of what the DEQ does - and more. Tho I'd rather do room EQ with the DEQ - it's just easier and more powerful. I do think that the DCX will be fine for your open baffle rig, it worked for me on several I built.
I use a DEQ in front of the DCX for a bit of room EQ and also just for the pretty RTA graphics. I found that the better and better I got at using the DCX and building crossovers, the less I needed the DEQ. But it was a wonderful tool for getting there.
Bottom line, they can work well together. Just come digital (AES - balanced) out of the DEQ into the DCX. Easy.
I'd say that's a bit misleading. They do different things. The DCX is very limited in EQ functions, and, of course the DEQ has no crossover capability. Both can do level adjustments. Both do A/D -> processing -> D/A. Both use AKM DACs and ADCs and SHARC DSP's.... though the DEQ has two SHARC's to the DCX's one. As mentioned elsewhere in this post, the DCX does not do frequency response measurements, while the DEQ does.
I'm running the same setup that Pano mentions, with the DEQ sending AES digital to the DCX. I'm running analog into the DEQ, and I've used the Behringer mic for reading and correcting room/system response.
I had tried to make most of the corrections using just the single per-band pararmetric EQ that the DCX provided, along with various crossover slope and center combinations, but was not able to achieve anywhere near the eveness of a correction result that I got much more easily with the DEQ. If your equalization is simpler than mine, though, you may find that you can tweak a smooth response out of the system by adjusting the DCX with the DEQ set flat and used only for measurement. You could then pull the DEQ out entirely.
Let us know what you find.
P.S. To be clear, the big difference in EQ between the two is the the DCX offers a single parametric EQ for each output, whereas the DEQ offers the choice of a 31 band graphic equalizer per channel, or multiple parametric EQ's per channel, or a setup that provides graphic EQ with the ability to change Q and center frequency on each band (a paragraphic EQ).
Cheers,
Paul
Last edited:
Yes, exactly
Attenuation was done in ARTA - so the DAC was working at -3dB / -13dB - and so did all the following stages.
In your case - having used a volume pot of the soundcard is already bad - you have at least altered the signal level of the (buffer / line driver) stages after the volume pot - meaning those (buffer / line driver) stages might have delivered different performance at different signal levels (as shown to be the case with my measurements) and possibly *this* is what you might have measured ?
Michael
Last edited:
24 bit, symmetric IN and symmetric OUT - external is my favorite - but seems there are more and more internals that are stunning.
There are a lot available nowadays - actually it might be even hard to find 16bit anyway
Check out good driver support and compatibility with ARTA or whatever you use!
Would have recommended the Terratec Phase 24FW as this one is now for around EUR 100.- but have heard that there can be compatibility problems on the fire wire hardware level.
I use Mackie Onyx 400f but this one is outdated (and was at the expensive side anyway)
Michael
There are a lot available nowadays - actually it might be even hard to find 16bit anyway
Check out good driver support and compatibility with ARTA or whatever you use!
Would have recommended the Terratec Phase 24FW as this one is now for around EUR 100.- but have heard that there can be compatibility problems on the fire wire hardware level.
I use Mackie Onyx 400f but this one is outdated (and was at the expensive side anyway)
Michael
Last edited:
That's true, and a good point. However I did do soundcard loop measurements at both attenuations and didn't notice any big difference. I was very worried about the levels affecting the measurements.
It is certainly worth another check to see - thanks for pointing it out!
I do wonder if we should take this to a different thread, so as not to clutter up this one. A thread about using consumer soundcards and software for measurements and how to achieve the best results.
Add a resistor of your liking in parallel to the mute transistor
possibly not scientifically the best way but very easy - and easy to remove too
Michael
Add a resistor of your liking in parallel to the mute transistor
possibly not scientifically the best way but very easy - and easy to remove too
Michael
It would have to be a pretty small resistor, though... 500 Ohms for 3dB of attenuation, 250 Ohms for 6dB, etc.
You could also play around with the values of the 499 Ohm input and 2.32K Ohm feedback resistors in the next stage after mute, in order to reduce gain of that stage. Stock gain is about 4.6X in this stage... well -4.6X, since it's an inverting stage.
Reducing the 2.32K would reduce the gain.