REW Beta Release REW API beta releases

[majorwong]

Hi,

John Mulcahy

4 more suggestions for you to add in REW:
1.Would you think of adding an unit of exportation of group delay,which can be imported in other simulation tools?
2.Measure the SPL and Imp,adding a sweep signal range from 20kHz~20Hz(from high to low) might increase stability of the test result just like soundcheck.To some extent it may reduce the 'hush' sound annoying during the test.
3.A brown theme like sunset should be considered to put into TO-DO-LIST.
4.Volume or adding mass method should show the optimal fs shift percentage(20％~50％) after 2 curves overlay,which can get rid of the manual calculation.

Thanks for your reply.​

 

[newbiejl]

Hi, I couldn't figure out the body (REST API) to set "Mic cal files" where "separateCalFileForEachInput" =true.
Have anyone tried this?
Thanks!

Here's my PUT/audio/input-cal
{
"currentInputSelection": "Microphone (E.A.R.S Gain: 24dB ) MICROPHONE (Master Volume)",
"separateCalFileForEachInput": true,
"inputDeviceIsCWeighted": false,
"calDataAllInputs": {},
"perInputCalData": {
"L":{
"calFilePath": "C:\\Users\\xxxx\\Downloads\\L_RAW_xxxx.txt"
},
"R":{
"calFilePath": "C:\\Users\\xxxx\\Downloads\\R_RAW_xxxx.txt"
}
}
}

======================================================================
But while setting up manually on the REW interface, GET input-cal will show:
{
"currentInputSelection": "Microphone (E.A.R.S Gain: 24dB ) MICROPHONE (Master Volume)",
"separateCalFileForEachInput": true,
"inputDeviceIsCWeighted": false,
"perInputCalData": {
"L": {
"calFilePath": "C:\\Users\\xxxx\\Downloads\\L_RAW_xxxx.txt",
"dBFSAt94dBSPL": -20.38971
},
"R": {
"calFilePath": "C:\\Users\\xxxx\\Downloads\\R_RAW_xxxx.txt",
"dBFSAt94dBSPL": -20.38971
}
}
}

Seems like the current API doesn't support "separateCalFileForEachInput": true
Any thoughts?

 

[John Mulcahy]

Seems like the current API doesn't support "separateCalFileForEachInput": true

It is supported but there is a bug which is preventing the paths being saved properly (unless the files have a sensitivitydBFS entry). I have fixed it for the next build. I also noticed the fullScaleSineVrms field was mislabelled fullScalSineVrms, that has also been fixed in the next build.

 

[John Mulcahy]

Would you think of adding an unit of exportation of group delay,which can be imported in other simulation tools?

That wouldn't be a unit, it would need a separate export for the group delay graph. I'll add it to the todo list.

Measure the SPL and Imp,adding a sweep signal range from 20kHz~20Hz(from high to low) might increase stability of the test result

Why would a reversed sweep increase test result stability? Reversing the sweep would mean harmonic distortion could no longer be extracted.

 

[serko70]

Reversing the sweep would mean harmonic distortion could no longer be extracted

But it would be more inline with the impulse graph

 

[newbiejl]

It is supported but there is a bug which is preventing the paths being saved properly (unless the files have a sensitivitydBFS entry). I have fixed it for the next build. I also noticed the fullScaleSineVrms field was mislabelled fullScalSineVrms, that has also been fixed in the next build.

Thanks! Just to double confirm, can you check my assumption below is correct? API document doesn't tell much about this

 

[John Mulcahy]

Yes, that's correct. You can omit structures that aren't relevant.

 

[John Mulcahy]

Builds updated today (beta 58) with these changes:

• Added: Show busy indicator while updating FSAF signal
• Added: Ctrl+click can be used to bring up the right click menus on macOS
• Fixed: Filter text export for LSQ and HSQ shelf filters had incorrect Q values
• Fixed: Selecting dBr as the Y axis on the SPL & Phase graph for a saved RTA measurement was not handled correctly
• Fixed: FSAF measurement could give a "file too short" warning on changing the signal duration when not using a file
• Fixed: API InputCalData structure fullScaleSineVrms field was mislabelled fullScalSineVrms
• Fixed: API per-input cal file paths posted to /audio/input-cal were not saved properly in the cal data

 

[Chroma]

Moving measurements around the screen with mouse right-click don't work anymore on Beta 58, this is on Mac.

 

[Mike0815]

Hello John,

I'm measuring with pro version and 8 microphones.
Could you add an option to automatically cross correlate to the first mic and then to vector average the correlated? (immediately after measuring) Currently the options allow to average without cross correlating things first I think.
This would speed up things considerably here.

Thanks
Michael

 

[John Mulcahy]

Moving measurements around the screen with mouse right-click don't work anymore on Beta 58, this is on Mac.

I have rebuilt the macOS installer to fix that, if you download again it should be OK.

 

[Chroma]

I have rebuilt the macOS installer to fix that, if you download again it should be OK.

Thanks works again...

 

[John Mulcahy]

Could you add an option to automatically cross correlate to the first mic and then to vector average the correlated?

I can add options for that, with a couple of caveats:

• It will only work when the Save inputs individually option is selected
• The vector average will have uniform weighting for all inputs, regardless of the weightings chosen for the rms average

 

[dcibel]

Minor bug report. In the measurement setup, when entering the number of measurement repeats, if you don't hit the Enter key in the box and just click off the drop-down, the number of repeats isn't updated. Selecting the drop down will still show the value entered, and the number of repeats is then updated when you click in the box.

This has caught me a few times, that I enter a value and click off the box and don't notice that it hasn't been changed.

 

[John Mulcahy]

In the measurement setup, when entering the number of measurement repeats, if you don't hit the Enter key in the box and just click off the drop-down, the number of repeats isn't updated.

Is that on Linux? I can't reproduce that behaviour on Windows.

 

[dcibel]

Yes, on Linux. I have a similar behaviour with the "add a number to the name" context menu.

 

[John Mulcahy]

MacOS has the same behaviour, I have added a fix in the next build.

 

[JoinUs]

Is it impossible to make each measurement's thumbnail selected in the SPL tab as it is in the current beta?
Sometimes I do it without a keyboard, but at the moment, I think it only works with alt + keyboard direction keys.

Also,
Does FDW and MTW have to be smoothed unconditionally? Can't I disable this?
I set after 1k to 500ms in MTW, but when export IR and import back, over 1khz is changed.

 

[John Mulcahy]

Is it impossible to make each measurement's thumbnail selected in the SPL tab as it is in the current beta?

I don't understand that question.

Does FDW and MTW have to be smoothed unconditionally? Can't I disable this?

FDW and MTW results are log spaced at 96 PPO and so have inherent smoothing of 1/48 octave.

 

[Breeman]

Hi John,

I might doing this incorrectly but when I create a electrical filter in the EQ screen and generate measurement from it, the generated measurement does not have the phase component. See below.

I was hoping that the generated measurement would have a phase component like this one.

Below is the speaker measurement before and after applying the filter (used the arithmetic A*B function).

The phase response of the speaker is not modified "as it should under normal circumstances" when such a filter is applied.

I was expecting something similar to below with the red graph showing the original phase response and the green one representing the phase repsone with the filtter applied. Am I following the correct method in using electrical filters or is the design intent to produce filters without phase response from within EQ screen? It would be great if the EQ screen can have an option to include phase in the generated response.

 

[John Mulcahy]

Did you use Generate measurement from filters, which will produce an impulse response from the current filter settings, or Generate measurement from target shape, which only produces a magnitude response?

 

[Breeman]

Hmm, just realised that when I use the "generate minimum phase" option for the electric filter under the SPL & Phase and use the same arithmetic function, I get the correct phase response. It would still be prefereable in EQ tab, to generate filters with phase response to avoid the extra step.

 

[John Mulcahy]

See above, use Generate measurement from filters.

 

[Breeman]

Did you use Generate measurement from filters, which will produce an impulse response from the current filter settings, or Generate measurement from target shape, which only produces a magnitude response?

No, I just used the Target Settings to simulate a filter I want to apply to the speaker response, not necessarily to equalise the speaker's response or generate measurement from filters which in this use case, would be flat. It is the generate measurement from target shape that I used but as you said, it only produces a magnitude response.

 

[Breeman]

See above, use Generate measurement from filters.

Using that option produces a flat line. I am trying to use EQ more as a crossover simulator "without equalising/matching the speaker response against the target curve using PK filters for an example" to better understand its behavior.