Odd microphone response

[Quadman]

I have a similar issue as well and rather than start a new thread I thought I'd post here. I will move the post if needed. My issue, I believe, is centered around the PC's sound cards or possibly the microphone (a new minidsp umik 1). I am new to REW but for years have used Audiomatica's Clio system, first clio 7 and now clio 8.52. I was hoping REW would provide a more portable reliable measuring system. First off I am running REW 5.19, I have tried measurements with REW on two PC's both with different soundcards. 1-my Dell Latitude 5400 w10 pro laptop with a realtek sound card (Waves MaxxAudio Pro) 2. a tower win 7 pro PC with a SB creative media Audigy sound card. I have downloaded and used the mic's calibration file. I have a 2 Channel system only, with simple Pass B1 buffered preamp. I use headphone out/RCA adaptor to run to preamp input and the UMIK-1 USB as input.

My issues are around frequency response with the win 10 laptop and realtek (Waves MaxAudiopro) soundcard with all EQ turned off I get literally no bass below 120 htz on speakers that measure flat to 30 Htz (25 htz if I unplug the bass ports) and my high frequencies above 2K are shelved down (sloping down). Jpeg below RTA, REW PN as source, Mic at Listening Position (LP) on speaker axis.

Now the win 7 pro PC with sound blaster creative media Audigy sound card with all enhancements/EQ turned off. In this case my bass reads a bit better still not correct but the high frequencies are not there, over 10db below main signal level. These are raal ribbon tweeters, flat to 20+K. Same RTA PN signal, mic at listening position and on axis with speaker as above measurement.

My comparision is to Audiomatica's Clio 8.52 a dedicated system with its own sound card and self calibration features in the software. This has to run on a Windows XP (preferably SP1) machine and I have Audiomatica's own calibrated microphone I use with this system. As mentioned it used audiomatica's supplied soundcard and driver. Below graph shows a much better bass area and highs that are mostly flat to 20K. Again mic was in same position as above 2 measurements, at LP and on speaker axis.

3 measurements, 3 different sound cards with clio the most trustworthy and 3 pretty different results. My first thought is my sound cards I use with REW are not giving flat frequency responses, is there a way to measure an internal sound card's frequency response on a laptop? Both REW PC's show reduced bass with the win 10 PC showing almost no bass below 120htz, both REW PC's also show reduced highs with the win 7 PC really dropping off over 10db below main signal.

Any idea if it is sound card, or mic related?

 

[John Mulcahy]

A soundcard can be measured by connecting its output to its input, but it would be unusual for a soundcard to have a response that isn't flat bar the very highest and lowest frequencies. More likely is some processing affecting the output. For example, various effects may be active by default (though turning off the enhancements in the sound properties should disable those) or the speaker output may be configured in the OS as being laptop speakers and so receive a bass limited signal.

 

[Quadman]

John I checked all the above and the speaker output was listed as large external, but what about the high frequencies even with laptop speaker setting that should not have been rolled off. I can easily check the Win 7 PC sound card (freq resp) but checking a laptop is a bit more complicated as there is only a headphone/mic in/out jack, how to connect those two together? I reached out to MiniDSP and opened a ticket to get their take on the situation as well. I purchased the Mic directly from them. It would be nice if I could use the UMIK with my clio system but it does not seem to recognize the UMIK.

 

[Antonio Di Motta]

From what I see on the answer with the Creative sound card the microphone calibration file is missing, and the sound card should also be calibrated in loop mode, although in this case as a USB microphone is used this additional calibration still introduces an error due to the ADC converter response that would not be used for USB microphone response. For this reason I always use pre-amplified analog microphones that also use the ADC converter as well as the DAC for audio output.

 

[Quadman]

From what I see on the answer with the Creative sound card the microphone calibration file is missing, and the sound card should also be calibrated in loop mode, although in this case as a USB microphone is used this additional calibration still introduces an error due to the ADC converter response that would not be used for USB microphone response. For this reason I always use pre-amplified analog microphones that also use the ADC converter as well as the DAC for audio output.

Antonio Thank you for the reply, I replaced the creative media jpeg with the microphone calibration file included, the original shot was just all SPL which does not show calibration files. I will have to study the loopback method of sound card calibration for a laptop with only one headphone/mic jack, but will still have the issue you pointed out. May I ask what mic and setup you use? It seems a lot of people use REW with the UMIK or EMM mic and get satisfactory results and I no information with REW guides on sound card recommendations.

 

[Antonio Di Motta]

I use two microphones, the first is based on the modification of the capsule of the Panasonic WM-61A built and calibrated by an Italian company. The second is, a Sonarworks XREF20 with calibration files, we say that it is very similar to the Behringer ECM8000 or Dayton Audio EMM-6. There are practically no response differences between the two, they are both analog and need a supply voltage and a preamplifier to work. Sonarworks has a balanced output while Panasonic has an unbalanced output.

 

[Antonio Di Motta]

I prefer the XREF20 sonawork because it has a lower noise level due to the balanced output type and works with a standard phantom voltage of + 48V 10mA and requires a preamplifier with a gain of about 10db to use all the dynamics up to at a sound pressure level of 130db SPL. A good product for the price paid of around 60Euro.

 

[Quadman]

I prefer the XREF20 sonawork because it has a lower noise level due to the balanced output type and works with a standard phantom voltage of + 48V 10mA and requires a preamplifier with a gain of about 10db to use all the dynamics up to at a sound pressure level of 130db SPL. A good product for the price paid of around 60Euro.

What mic preamplifier do you use?

 

[Antonio Di Motta]

I design all my circuits and the preamp was also made by me, so I can't help unless you are an expert in electronics and you can build a circuit yourself. To be clear, I send photos of my latest creation built entirely in my laboratory.

 

[Quadman]

I design all my circuits and the preamp was also made by me, so I can't help unless you are an expert in electronics and you can build a circuit yourself. To be clear, I send photos of my latest creation built entirely in my laboratory.

Nice, I have built preamps, linear PSU's and done a few repairs to amplifiers. For something like this I think I would go the off the shelf preamp. Tho I do own 2 other analog mics, the clio one with an RCA output and one from Neutrik with balanced out that is used with their Minilyzer analyzer. Custom making one could probably allow for most any mic to be used.

 

[Quadman]

Ok I emailed miniDSP about the responses I posted above and they suggested I use a USB dac as my sound card. Well I have a T+Adac8dsd that I use with roon and HQPlayer so I connected that to my win 10 pro laptop via USB and plugged the UMIK-1 into another USB port on the laptop. I selected the T+A dac as default in sound settings and took some measurements. I took 4, 1 sweep at the listening position with the mic on tweeter axis and 3 measurements from 1 meter away (1 sweep, and 2 RTA).

Below is a sweep from the LP with the preferences box open showing the T+A as the output device and during measurements the T+A indicated a 48K sample rate.

Now the jpeg of the 4 measurements. The red sweep line is taken from the Listening position with the mic on tweeter axis, The Orange sweep line is taken from 1 meter away with the mic on tweeter axis, I lowered output level to more closely match the LP measurement. The green RTA line is taken from 1 meter away with the mic in the same position as the orange sweep and the blue RTA line is taken from 1 meter away but I lowered the mic so it was now at woofer level and aimed just above the woofer dust cap. As reference the same laptop was used with its internal sound card in my first posted graph above called (W10 at LP FR on Axis).

In the above graphs the T+A dac was used as my sound card, this should have eliminated any effect of the internal sound card on the measurements. The general curve on the responses is very similar to my first graph published above on Monday May 25th. I get a few Hz more bass with the T+A but at 90 Hz the bass goes away and my high frequencies are similar to the first graph. In either case these graphs are a long way off of what I get from Clio 8.52.

Could I be missing some setting in REW that cuts the bass off and affects the high frequencies? Does this look like a microphone issue? And yes I used the mic's calibration file for these measurements.

 

[Antonio Di Motta]

This cut at low frequencies is very strange, I would think that the fault may be in the DAC but having tried both the internal card and an external DAC without finding differences we should exclude a problem on the low frequencies of both cards. I would try to check the setting at the mixer level of Windows 10, maybe there is some filtering active or the use of an audio output set for the rear channels that have active the high pass filter.
At the REW level I do not discount the activation of the high pass filter and therefore the REW settings are correct.
Try to see the window mixer settings and exclude any inserted equalizers or environmental effects presets.

 

[Quadman]

Its definitely not the dac. I checked my sound card options and every enhancement was turned off, no where is there an option for cutting off low frequencies. Since I used the dac as a sound card I had no option to select large speaker as output everything was greyed. The more I measure the more I think it is microphone related. Went to the win 7 machine today double checked all mixer options and had everything turned off. when I measure the high frequencies are 8-9 db below the main midrange level. whereas clio has high frequencies and midrange levels much closer together (within 3 db). So I wait to hear back from minidsp as to their thoughts on my measurements with a dac.