timschneeb / DDCToolbox

Create and edit DDC headset correction files
GNU General Public License v3.0
137 stars 13 forks source link

Parametric EQ settings from AutoEQ gives different results #4

Open svenkerst opened 4 years ago

svenkerst commented 4 years ago

When importing a ParametricEQ.txt, the plotted graph is quite different from that of AutoEQ itself.

image

image

timschneeb commented 4 years ago

Seems right to me, the thin green line and the plot in DDCToolbox look almost identical.

Could you point out which frequency range looks out of place?

svenkerst commented 4 years ago

I have the impression that the line is sort of tilted downwards from around 500 Hz onwards. For example, the peaks at 1826 and 7073 have positive amplitudes in the AutoEQ graph, but not when imported into the DDCToolbox. Maybe it makes sense, but I just don't understand why that is.

timschneeb commented 4 years ago

Yes, you're right. There is some discrepancy between both graphs. I'll have a closer look at how AutoEQ calculates their graphs

svenkerst commented 4 years ago

Great, thank you! One more thing: shouldn't changing the filter type also change the graph accordingly?

timschneeb commented 4 years ago

It does, but you need to click in another table cell to commit your change (the GUI library I'm using handles widgets inside table cells a bit strangely)

Example: screencap

timschneeb commented 4 years ago

You're using the latest stable build right? I just checked and your issue seems to be fixed on the latest version directly compiled from sources. I was rewriting large chunks of the code which handles/stores filters a while ago so that might have fixed it.

Use the latest nightly release instead: https://nightly.timschneeberger.me/ddctoolbox-win/ddctoolbox-win64_1.3.85.zip

Screenshot

By the way, this version can download profiles directly from AutoEQ (File > Download from AutoEQ)

svenkerst commented 4 years ago

Thank you for your help. I was indeed using the latest stable build. The latest nightly build gives more similar results, but still they're a bit different (negative instead of positive 7073 Hz peak for example). Do you know what's the reason behind this?

Another problem I now have with the latest nightly build is that it crashes when I import a VDC file.

svenkerst commented 4 years ago

Converting the VDC file to a project file fixed it for me.

timschneeb commented 4 years ago

Could you upload the VDC file that crashes the app? Also, you can only import 'classic'-VDCs (VDCs that only contain Peaking filters) I'll also look into the other issue

svenkerst commented 4 years ago

Thank you very much! Here's the VDC file that's causing the toolbox to crash Example VDC.zip

timschneeb commented 4 years ago

I couldn't reproduce the crash you described (on version 1.3.85). However, I noticed some VDCs are falsely recognized as not compatible. That issue should be fixed with the latest nightly build: https://nightly.timschneeberger.me/ddctoolbox-win/ddctoolbox-win64_1.3.90.zip

rlw6534 commented 2 years ago

There still seems to be a problem with autoeq downloads and imports. As you can see the gain levels are significantly different. The VDC files also sound very different than the convolver files (especially the SE846), so something is wrong:

image

Screen Shot 2022-03-04 at 11 03 52 AM

Note all gains are negative.

and

image

Screen Shot 2022-03-04 at 11 07 44 AM

Note the treble gains are much higher.

nift4 commented 2 years ago

Manually adding Low/High Shelf helped for me: Before (top) vs after (bottom) Screenshot_20221010_204142

After vs AutoEQ graph Screenshot_20221010_204211