REW Beta Release REW API beta releases

[Mike0815]

One more wish:
When using HDMI as output device and AVR is shut off REW hangs and needs to be killed because it gets irresponsive. Could output device disappearance be handled more graceful? (switching to default maybe?)

 

[John Mulcahy]

Wouldn't it make sense for the reference level to also be calculated from an average over that same span?

I don't think so. Local features of the responses near the chosen merge frequency would shift the average, taking away your ability to choose a point in the response that has the correct level, whatever that may be. The responses may also be curved near the merge, so again the average would be incorrect.

 

[John Mulcahy]

When using HDMI as output device and AVR is shut off REW hangs and needs to be killed because it gets irresponsive.

That's a fault of the driver, not REW. When devices disappear REW does revert to default.

 

[Bowkey]

I don't think so. Local features of the responses near the chosen merge frequency would shift the average, taking away your ability to choose a point in the response that has the correct level, whatever that may be. The responses may also be curved near the merge, so again the average would be incorrect.

Hi John,

When using merging to stitch a far-field measurement to a near-field measurement, where the levels don't match, the current implementation makes a lot of sense.

However, I'm stitching two measurements where I want the levels of both to remain the same. Is there a way to do this without having to pick the exact merge frequency based on its level?

I think the level-averaging approach could work in both use cases, as long as the average level of plot A (in the blend range) is used as a reference to shift plot B, based on the offset of its average in that same range?

Maybe a simple 'don't shift levels' parameter could work as well?

 

[jeppekristoffer]

Is it only me or is the API version of REW slower and more demanding on the pc than the regular "non beta version"? Especially when selecting multiple measurements in the ALL SPL tab?
Is there anything I can do about it and "increase the speed"?
Rew seems to "think" alot when selecting/deselecting(highlighting) multiple measuremente
When I say multiple I mean more than 100.

 

[John Mulcahy]

Is it only me or is the API version of REW slower and more demanding on the pc than the regular "non beta version"? Especially when selecting multiple measurements in the ALL SPL tab?
Is there anything I can do about it and "increase the speed"?
Rew seems to "think" alot when selecting/deselecting(highlighting) multiple measuremente
When I say multiple I mean more than 100.

There is no equivalence. In the 5.31 builds all measurements are selected all the time, each new measurement is selected as it is generated or loaded. You are only changing whether or not the corresponding trace is plotted. In the 5.40 builds traces are generated when measurements are selected. If you want 5.31-style behaviour select all and use the graph legend checkboxes to turn traces on or off.

 

[jeppekristoffer]

There is no equivalence. In the 5.31 builds all measurements are selected all the time, each new measurement is selected as it is generated or loaded. You are only changing whether or not the corresponding trace is plotted. In the 5.40 builds traces are generated when measurements are selected. If you want 5.31-style behaviour select all and use the graph legend checkboxes to turn traces on or off.

Ok. Thank you i understand now.
Great. Thanks again.

 

[jeppekristoffer]

@John Mulcahy By the way. Thanks alot for the addition off the possibility to add the same eq-filter to multiple measurements and the new "recolour scheme". Super great.

 

[John Mulcahy]

Builds updated today (beta 40) with these changes:

• Added: Ctrl+TAB and Ctrl+Shift+TAB can be used to navigate through the measurements right click panel fields and then the list of measurements
• Added: CamillaDSP equaliser prompts for a channel number when saving yaml files
• Added: CamillaDSP target match allows shelf filters
• Changed: Force "type" properties to start of list in CamillaDSP yaml files
• Changed: When the measurement panel right click menus are shown the name field is selected, ready for editing
• Changed: Restore windows to screen they were last on
• Fixed: "Measure with these filters" caused an exception if there had not been any previous measurements
• Fixed: Possible exception when switching from single to multiple inputs
• Fixed: Window resizing limits were based on the display used on startup rather than the current display for the window
• Fixed: API did not accept the mergeBlend parameter

 

[HD MOVIE SOURCE]

That is not a bug. Per the notes:

• The result of arithmetic on measurements that have compatible impulse responses is smoothed using the measurement A smoothing, unsmoothed data is used during the calculations. Other measurements use whatever smoothing they already had applied during the calculations and the result is treated as unsmoothed (or 1/48 octave smoothed if data is 96 PPO).

If you want to ignore the measurement impulse responses you could make magnitude-only copies and do arithmetic on those copies.

Okay, perfect, thank you.

 

[John Mulcahy]

Builds updated today (beta 41) with these changes:

• Fixed: Ctrl+TAB in overlay measurements list popup could skip some measurements
• Fixed: Edits to measurement name or notes might not be committed when using Ctrl+TAB to move to next or previous measurement
• Fixed: Ctrl+left click did not select multiple measurements in the main measurements list

 

[serko70]

Right click "Remove" removes the one before last in Beta 41.

 

[John Mulcahy]

Builds updated today (beta 42) with these changes:

• Added: Support for Q-SYS biquad coefficient format when saving generic equaliser coefficients to file
• Fixed: Measurements right click panel remove button could remove preceding measurement

 

[Mike0815]

Hello John,

thanks for the Camilla pipeline changes (and of course for all the other fixes). Just checked it out and can confirm it is working without any issues

 

[serko70]

Bug report:

When you remove a measurement from REW, the measurement BEFORE the deleted measurement will go back to REW's default EQ target type even if it's target type is set to "None" before the deletion of the measurement after it.

 

[John Mulcahy]

When you remove a measurement from REW, the measurement BEFORE the deleted measurement will go back to REW's default EQ target type even if it's target type is set to "None" before the deletion of the measurement after it.

Thanks, looks like that happens if the measurement being deleted is the currently selected measurement. I have fixed it for the next build.

 

[Cristianolo]

Builds updated today (beta 41) with these changes:

• Fixed: Ctrl+TAB in overlay measurements list popup could skip some measurements

I noticed that using the CTRL+TAB method after right-clicking on the measurement, there is no change to the selected measurement, only the floating panel has the data modified, following the order of the list.
Is this how this new function is supposed to work?
Wouldn't it be interesting if the selected measurement were to be skipped one by one along with the floating panel?

 

[John Mulcahy]

Is this how this new function is supposed to work?

Yes.

Wouldn't it be interesting if the selected measurement were to be skipped one by one along with the floating panel?

Maybe, happy to hear opinions on that.

 

[Breeman]

Hi John,

Is it possible to measure the response time of an the audio interface like the Steinberg UR22C,using the loopback cable, store this in REW and then use two XLR microphones simultenously to measure while referencing the stored response time as a timing ref? Would appreciate it if can be made a feature in REW.

 

[John Mulcahy]

Is it possible to measure the response time of an the audio interface like the Steinberg UR22C,using the loopback cable

If by "response time" you mean the delay between sending a signal to the interface and getting a signal back from it then no, that isn't possible due to the various buffers between data generation and the interface receiving data from the OS. The only useable references are the inputs from the interface, timing is relative to the signals on an input.

 

[Breeman]

If by "response time" you mean the delay between sending a signal to the interface and getting a signal back from it then no, that isn't possible due to the various buffers between data generation and the interface receiving data from the OS. The only useable references are the inputs from the interface, timing is relative to the signals on an input.

Thank for the quick response.

Would an interface like the Focusrite Scarlett 8i6 suffice then for this task? Would a phono cable from line output to line input represent a valid loopback setup (taking into consideration the potential slight differencies in internal electronics/buffers between line and mic inputs) for providing timing reference whilst measuring with two XLR microphones at the same time?

 

[John Mulcahy]

Would a phono cable from line output to line input represent a valid loopback setup

Yes.

 

[jeppekristoffer]

Bug: (latest build from 7th of july)
In All SPL tab when selecting measurements its no longer possible to use the TAB or SHIFT+TAB key to "scroll" between measurements in the lowest part of the screen to then "activate/deactivate" with the SPACE key. I have to pick the measurements with the mouse. Then i can "activate/deactivate" them with the SPACE key or with the mouse.

 

[John Mulcahy]

In All SPL tab when selecting measurements its no longer possible to use the TAB or SHIFT+TAB key to "scroll" between measurements in the lowest part of the screen to then "activate/deactivate" with the SPACE key.

Are you referring to the graph legend area? I can tab back and forth through that. Focus has to be on the legend first, e.g. by clicking on a trace checkbox. Seems to behave the same as a few earlier builds I have tried.

 

[jeppekristoffer]

Are you referring to the graph legend area? I can tab back and forth through that. Focus has to be on the legend first, e.g. by clicking on a trace checkbox. Seems to behave the same as a few earlier builds I have tried.

False alarm. Its working now. Yes i was referring to the grapgh legend area. Sorry.