Importing a pulse sweep from Acourate into REW?

Horacio Lewinski

New Member
Thread Starter
Joined
May 17, 2018
Messages
57
Hello.

I've used REW in the past, but over the last couple years I turned my stereo into 3-way active running digital crossovers inside HQPlayer. Those filters are generated on Acourate.

So in attempting to use REW again to work on my room acoustics I'm faced with the inability to run a sweep to be convolved on HQPlayer. I understand there is a way to make HQPlayer run the seep and REW to record the result, and I understand I can also take the sweep ran on Acourate, save it as 24 bit wav file, and import it into REW. I'm trying the latter, unsuccessfully.

The hints and tips thread says "Impulse response files in .wav or .aiff format can be imported by dragging them onto the REW window". However when I do this nothing happens.
I also tried going to File>Import pulse and selecting the wav file. It seems to take it but nothing happens. I don't know how to tell REW to graph it. When one takes a measurement with REW the graph is displayed automatically.

I must be missing something.

I guess my question is twofold:
a) What is the best approach to measure a system like this with REW?
b) If the approach I've been trying is good, what am I doing wrong?

Thank you!
 

John Mulcahy

REW Author
Joined
Apr 3, 2017
Messages
7,212
Please attach an example of a WAV file you are trying to import. Which version of REW are you running?
 

Horacio Lewinski

New Member
Thread Starter
Joined
May 17, 2018
Messages
57
Hello.

REW version is 5.18.
Please find the requested file. The website didn't allow uploading a wav file so I changed the extension to txt. Please change back to wav.

Thanks in advance!
 

Attachments

  • Pulse48LmonoforREW.txt
    112 bytes · Views: 9

John Mulcahy

REW Author
Joined
Apr 3, 2017
Messages
7,212
There isn't any audio data in that file, just header information (hence it being only 112 bytes long).
 

Horacio Lewinski

New Member
Thread Starter
Joined
May 17, 2018
Messages
57
I re-ran the creation of the wav file on Acourate. I must have made a mistake before.

It is now working. Thank you!!

Now, onto interpretation :)
Acourate runs a logsweep on the left and then another on the right, and generates two files. This is for working with each speaker. But if I want to evaluate my room needs for treatment (especially bass treatment/traps, say under 200Hz) should I analyze each L & R sweep separately or add them up onto a new wav and import that into REW for evaluation?
 

John Mulcahy

REW Author
Joined
Apr 3, 2017
Messages
7,212
Import each speaker result separately. You can sum them in REW afterwards if required, using the trace arithmetic features in the All SPL graph.
 

Horacio Lewinski

New Member
Thread Starter
Joined
May 17, 2018
Messages
57
Thank you.

I was able to test and adjust some things, but I'm thinking it will be best to run the sweep through HQPlayer so it gets convolved with the DSP filters and take the measurement with REW. How can I do that?

Regards
Horacio
 

John Mulcahy

REW Author
Joined
Apr 3, 2017
Messages
7,212
Easiest is to generate a sweep with a timing reference and save it as a WAV file (as described in "Generating the sweep signal" here), tick the "Wait for timing reference" box on the Measure dialog, click Start measuring and then play back the WAV file.
 

Horacio Lewinski

New Member
Thread Starter
Joined
May 17, 2018
Messages
57
I was able to generate the WAV file and played it through HQPlayer to make sure it sounded like a sweep and it does. Good.

Then onto measuring: in Preferences I use the Hilo as input, go to Measurement and tick Wait for timing reference, and when I press Start measuring I get an error message: "The soundcard did not provide any input data. Please check that it is connected". But the card is connected and in fact is seen by REW in the Preferences/Input dropdown menu.

What could this be caused by?
 

Horacio Lewinski

New Member
Thread Starter
Joined
May 17, 2018
Messages
57
John,

I attempted to simplify and got a little further and was able to measure, but the results look wrong.

Now I have my audio computer play the wav sweep from HQPlayer, so it's convolved with the crossover and room correction filters like the music does.

I now installed REW on a laptop and connected the Tascam US-122MkII to it thru USB. The mic is connected to the Tascam. Sweep goes from 18 to 24000Hz at 48kHz. I tell REW to start measuring, tell HQPlayer to play the sweep, and I hear the external timing reference and the sweep starts, and REW records, BUT:
  1. REW goes onto generating the graph before the sweep ends...so how come it knows what the top end measures?
  2. REW indicates clipping is happening. At this point I'm only reducing volume on HQPlayer, but that lowers the volume while the noise floor stays the same...can't be good.
  3. The attached measurement looks wrong: way too much low end volume. On Acourate I set a target curve declining with frequency, but nowhere near this. And this should sound awful and it doesn't (although I do believe I have too long decay times in the low end). Maybe it's what it is and I need to correct, but in light of so many anomalies I'm second guessing it.
  4. The waterfall chart is wrong. Doesn't show all the frequencies, and has about 2 seconds decay...seems too much, even for my perception of long decay times.
upload_2018-5-27_17-56-21.png
 

Attachments

  • 2018-05-25 response 20-20000.jpg
    2018-05-25 response 20-20000.jpg
    49 KB · Views: 9
  • 2018-05-25 response 20-200.jpg
    2018-05-25 response 20-200.jpg
    47.2 KB · Views: 7
  • 2018-05-25 response 20-20000.jpg
    2018-05-25 response 20-20000.jpg
    49 KB · Views: 8
  • 2018-05-25 waterfall.jpg
    2018-05-25 waterfall.jpg
    52.6 KB · Views: 8

Horacio Lewinski

New Member
Thread Starter
Joined
May 17, 2018
Messages
57
Just in case I'm also attaching the wav sweep file, should I have made a mistake when creating it.

Thanks John a lot with this!!
 

Attachments

  • MeasSweep_48000_18_24000_LR_refL.txt
    6.2 MB · Views: 6

John Mulcahy

REW Author
Joined
Apr 3, 2017
Messages
7,212
Please don't leave the audio devices as "Default", the default device changes as devices are connected/disconnected. Select the actual device you want to use and the actual input/output you want to use on it. Make sure Mon Mix on the US-122 is turned all the way to Computer. Make sure the sweep settings on the Measure dialog are exactly the same as the settings you used when generating the sweep file. If you have clipping turn down the mic gain on the US-122.
 

Horacio Lewinski

New Member
Thread Starter
Joined
May 17, 2018
Messages
57
Thank you.
On US-122, Mon Mix is/was 100% to Computer, Signal 100% to Mic, mic gain I guess is Phones/Line Out and tried different settings with no effect. I did go back and save a new measurement wav to make sure the settings were exactly the same as I'm using now - that solved the issue of REW completing the process before the sweep was over. And that seemed to have fixed the issues.

For some reason I pasted a wrong image of my Preferences page. Here's what I'm using on the laptop:
upload_2018-5-27_21-39-4.png


Also my Measurement setup page:
upload_2018-5-27_21-40-32.png


For some reason adjusting Phones/Line Out on the US-122 doesn't change the clipping situation. Maybe it's because of the increased volume at lower frequencies?

Here are the response and waterfall. Now looking more like I was expecting. Anything looking odd in these? Otherwise I can assume the measurement setup is OK and start focusing on working on fixing the room and system.
 

Attachments

  • 2018-05-27 response 15-20000.jpg
    2018-05-27 response 15-20000.jpg
    47.5 KB · Views: 7
  • 2018-05-27 response 15-200.jpg
    2018-05-27 response 15-200.jpg
    30.7 KB · Views: 6
  • 2018-05-27 waterfall.jpg
    2018-05-27 waterfall.jpg
    66.1 KB · Views: 7

John Mulcahy

REW Author
Joined
Apr 3, 2017
Messages
7,212
The knob marked Mic at one end and Line at the other is the input gain, all the way to Mic is maximum gain. Turn it back towards Line to reduce the gain and fix the clipping. Otherwise looks better but post an mdat file of a measurement once the input gain is adjusted.
 
Top Bottom