Equalizer APO/Peace with Foobar2000 Sample Rate Wrong on E30

[ngs428]

Hello,
I have a WIndows 10 PC, and am using the latest version of Equalizer APO with Peace, Foobar2000 and Topping E30 DAC and L30 amp .

For Equalizer APO to work with Foobar2000 I had to go into Menu>File>Preferences>Playback>Output>Device and change to "Default - Speakers (2-Topping USB DAC)" I was on "WASAPI (event) (2-Topping USB DAC)". Seems to be working fine, but the sampling rate shown on my Topping E30 screen is always 44.1.

Is there a different setting I should be using in Foobar2000 to pass through the correct sample rate?

 

[Vasr]

It is a bit of a catch-22 with Windows. You cannot use an APO in direct mode (WASAPI, ASIO, etc). If you use shared mode, then the Windows sound engine resamples everything to the rate selected for the device in the audio devices list in the control panel. This is where you will need to make the change not foobar or EAPO both of which will send without resampling.

It is probably set at 44.1khz at the moment for the output device in audio devices. You can set it to the highest sample rate you want to play and everything below it will be up-sampled to it and everything above will be downsampled to it. There is no way to avoid resampling for content with rates other than what you have picked for shared mode in the Control Panel. The latter mode is necessary if you want to attach the EAPO in the chain.

 

[ngs428]

It is a bit of a catch-22 with Windows. You cannot use an APO in direct mode (WASAPI, ASIO, etc). If you use shared mode, then the Windows sound engine resamples everything to the rate selected for the device in the audio devices list in the control panel. This is where you will need to make the change not foobar or EAPO both of which will send without resampling.

It is probably set at 44.1khz at the moment for the output device in audio devices. You can set it to the highest sample rate you want to play and everything below it will be up-sampled to it and everything above will be downsampled to it. There is no way to avoid resampling for content with rates other than what you have picked for shared mode in the Control Panel. The latter mode is necessary if you want to attach the EAPO in the chain.

Thanks for the reply. Would there be any suitable equalizer add-ins in foobar which would allow me to use direct mode?

I took a look at musicbee and it has a built in equalizer which works in direct mode (still passes the correct sample rate). Although the eq looks very basic.

 

[Ron Texas]

Thanks for the reply. Would there be any suitable equalizer add-ins in foobar which would allow me to use direct mode?

I took a look at musicbee and it has a built in equalizer which works in direct mode (still passes the correct sample rate). Although the eq looks very basic.

The "default speakers topping" setting is shared mode WASAPI. I don't know how direct it is. You could set the bit rate on the windows sound properties page for the DAC at 96k and the DAC will always show 96. That way you are not down sampling high res files. Also, you could add the SoX component and specify it as the default resampler in Foobar (without putting it in the active DSP stack.

If you must use WASAPI there are some VST plugins which work, but are hard to use. Also, there is Mathaudio. Be careful to set the volume low the first time you use it. Finally, you could generate FIR filters with Re-Phase and use the convolver component. That still requires resampling to a single frequency.

Remember, once you are applying EQ, the output is not bit perfect anyway.

 

[ngs428]

The "default speakers topping" setting is shared mode WASAPI. I don't know how direct it is. You could set the bit rate on the windows sound properties page for the DAC at 96k and the DAC will always show 96. That way you are not down sampling high res files. Also, you could add the SoX component and specify it as the default resampler in Foobar (without putting it in the active DSP stack.

If you must use WASAPI there are some VST plugins which work, but are hard to use. Also, there is Mathaudio. Be careful to set the volume low the first time you use it. Finally, you could generate FIR filters with Re-Phase and use the convolver component. That still requires resampling to a single frequency.

Remember, once you are applying EQ, the output is not bit perfect anyway.

Thanks for the reply Ron. Oddly enough, I was googling this topic and found a reply from you and Patrick1958 on Mathaudio and will be giving it a try.
https://www.audiosciencereview.com/forum/index.php?threads/parametric-eq-on-foobar.10019/

In Mathaudio I see the EQ gain only goes down to -6dB, I need -10dB for my 6XX. Do I just reduce all the gain values for each band 4 dB lower than stated?



Also, I do have SoX installed, so I could venture down that path too..

 

[jeffbook]

Thanks for the reply Ron. Oddly enough, I was googling this topic and found a reply from you and Patrick1958 on Mathaudio and will be giving it a try.
https://www.audiosciencereview.com/forum/index.php?threads/parametric-eq-on-foobar.10019/

In Mathaudio I see the EQ gain only goes down to -6dB, I need -10dB for my 6XX. Do I just reduce all the gain values for each band 4 dB lower than stated?



Also, I do have SoX installed, so I could venture down that path too..

In my experience with the latest version of MathAudio Headphone EQ (v. 2.7.7), I have noted that the maximum dB level for any parametric EQ input is shown as 0 dB. All other EQ filters have a peak reading of less than 0 dB. This lowers the overall output level prior to any gain reduction selected by the output slider. So if, for instance, the maximum gain of any particular parametric EQ filter, say 6 dB for example,is combined with the Headphone EQ gain slider set to -6 dB, the overall gain output from this plug in would likely be -12 dB.

While I have actually not measured this as of yet, I have noted a significant reduction in output volume when the EQ is inserted on my old HD280 Pro headphones. (6XX in signature below to be shipped from Drop on 1/1221)

 

[Ron Texas]

Thanks for the reply Ron. Oddly enough, I was googling this topic and found a reply from you and Patrick1958 on Mathaudio and will be giving it a try.
https://www.audiosciencereview.com/forum/index.php?threads/parametric-eq-on-foobar.10019/

In Mathaudio I see the EQ gain only goes down to -6dB, I need -10dB for my 6XX. Do I just reduce all the gain values for each band 4 dB lower than stated?



Also, I do have SoX installed, so I could venture down that path too..

I have tried every well known way to do parametric EQ with Foobar and have found equalizer APO works best. Differences in sound quality are mostly attributable variations in frequency response because the methods work differently. What I love about Equalizer APO is it can read the filter settings from an REW text file, so no typing that stuff in.

 

[Vincent Kars]

Would there be any suitable equalizer add-ins in foobar which would allow me to use direct mode?

Maybe this one: https://www.foobar2000.org/components/view/foo_dsp_xgeq

 

[ngs428]

I have tried every well known way to do parametric EQ with Foobar and have found equalizer APO works best. Differences in sound quality are mostly attributable variations in frequency response because the methods work differently. What I love about Equalizer APO is it can read the filter settings from an REW text file, so no typing that stuff in.

I asked Mathaudio themselves about the gain settings I was questioning and they stated:
--------------------------------------------------------------------------------------------------
Headphone EQ automatically sets the total gain, so you can ignore the
"Preamp gain" from the Oratory settings. See item 12 at
https://mathaudio.com/headphone-eq.htm for more information.

---
Kind regards,
Sergey
https://mathaudio.com
--------------------------------------------------------------------------------------------------
Also, with MathAudio you can save and load presets. Working to create one now, kind of cumbersome....

 

[ngs428]

Maybe this one: https://www.foobar2000.org/components/view/foo_dsp_xgeq

thanks for the option Vincent. If this MathAudio doesn't work I can give that one a shot!

 

[Vincent Kars]

Give MathAudio a shot first.
Even the smallest resonance peaks can be addressed.
No other EQ can do this as far as I know

 

[ngs428]

For anyone who has the HD 6XX or HD 650 headphone, here is the Oratory1990 preset for MathAudio Headphone EQ.

 

[Ron Texas]

Give MathAudio a shot first.
Even the smallest resonance peaks can be addressed.
No other EQ can do this as far as I know

My take on Mathaudio is it's one stop shopping. Hook up a microphone, measure, press a few buttons and voila, you are done. No need to learn how to use REW or transfer it's data somewhere else. If the system is music only, that's it. Things get a bit more difficult when video is thrown in. I suppose one could use the PEQ in VLC or go system wide with Equalizer APO keeping Foobar out of it with WASAPI.

Probably what led me to Equalizer APO was wanting EQ with Netflix. My smart TV sent audio directly to the DAC and I had no hardware equalizer to deal with it well. So, I watch Netflix using a browser and it works.

Truth be told, there are many ways to do this.

 

[jeffbook]

This is a follow up to my earlier post on this thread. In my earlier post on this thread ( #5), I mentioned that it appears that MathAudio Headphone EQ for Foobar reduces the overall output by setting the maximum boost applied in the EQ to a level of 0 dB. I also said that I had not made any measurements to confirm this hypothesis.

Since then I have made some measurements with the results shown in the attached files. These were generated by playing a pink noise signal in Foobar 2000 with the MathAudio Headphone EQ plug in enabled in the DSP manager of the Playback Preference. Output from my laptop went to the ASIO driver for a Topping D10 DAC, and the analog output was input into a Focusrite Scarlett 2i2 with measurements taken with ARTA.

There is a screenshot shown of an EQ I developed for my Sennheiser HD280 Pro headphones. In this EQ, there is a total range of about 11 dB between the maximum boost and cuts in the filter.

In one graph, the output of the pink noise signal with no EQ applied in the MathAudio plug in is compared to the output with the HD280 Pro filter active. The maximum boost peak touches the no EQ applied pink noise output. The remainder of the filtered response is below the output from Foobar with no EQ applied. The Headphone EQ gains slider was set to 0 dB in this test. In the second graph, the Headphone EQ slider was set to -6 dB. The filtered output maximum peak is -6 dB below the pink noise output when no filters were applied.

The practical result is that any boost EQ applied in MathAudio Headphone EQ should never clip the output form Foobar. Of course, the downside of this is that, depending on the EQ applied, a significantly reduced signal level will be sent to a headphone amp. With both my Topping L30 (sent back to Amazon due to the ESD/blown headphone matter) and the JDS Lab Atom that replaced the L30 needed to be set to the high gain level for maximum flexibility in listening levels.

 

[ngs428]

Thanks for the analysis @jeffbook

I asked mathaudio about the headphone eq gain and they stated:


> Headphone EQ automatically sets the total gain, so you can ignore the
> "Preamp gain" from the Oratory settings. See item 12 at
> https://mathaudio.com/headphone-eq.htm for more information.

 

[sweetchaos]

Thanks for the inspiration...

I've added it to my List of Amir's Headphone PEQ filters.
So people are aware of this.