• Welcome to ASR. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!
So you care about extracting every little bit of dynamic range, yet you recommend a -6 dB attenuation (which is quite substantial) just on the off chance that mixed audio might clip? Does audio quality really matter that much when multiple applications are playing audio at the same time, anyway?
-6db is a lot, especially if no improvement in sound quality is to expect, if -0.2dB is already enough to prevent the CAudioLimiter to trigger.

But, if I understand correctly, Windows software volume control would attenuate much more than that. I don't recklon anyone would use it only in the narrow margin of -6dB to 0 dB.

So I don't see anything wrong with a -6db preamp, unless software volume control is wrong in the first place.
 
But, if I understand correctly, Windows software volume control would attenuate much more than that. I don't recklon anyone would use it only in the narrow margin of -6dB to 0 dB.
That is not how Windows volume control works. If it is speaking to a device with hardware volume control (e.g. a Realtek chip on your motherboard, a USB DAC) then it gives the full digital signal to the device and also tells it the volume level to implement in the analog section of the DAC chip.
 
That is not how Windows volume control works. If it is speaking to a device with hardware volume control (e.g. a Realtek chip on your motherboard, a USB DAC) then it gives the full digital signal to the device and also tells it the volume level to implement in the analog section of the DAC chip.
I wrote: software volume control. See @edechamps 's posts.
 
Good stuff and for those that want great sound in windows with their personal flac files there's Foobar2000.
 
Have you tried using some compressor VST within Equalizer APO?
That's a nice thought, but my understanding is that VST plug-ins require a VST host, and I'm unaware of how EAPO might host VSTs. Firing up EAPO, I don't see any facility for it. I'm aware of other things that might host VSTs including Cantabile and VSTHost, but the complexity and effect this might have on audio quality terrify me.
 
Firing up EAPO, I don't see any facility for it.

1647453051202.png
 
I thought this was great, thanks. I note that as much as I love EAPO, you should mention its near-necessary brother Peace (just search Peace for APO). Also I _desperately_ wish I could find something as clean/simple offering compression. Side-stepping the silly argument that compression introduces fundamental inaccuracies (of course it does; altering dynamics makes this a tautology), I _really_ like to compress-the-hell out of much of my listening. I've ended up using the Behringer DEQ-2496 (in pure-digital capacity), which can be found for perhaps $150 used. If there were a good APO for this, that would be fantastic (as you note, freeing hardware, and being cost-free).
Just wondering kind of settings you're using as general purpose compressor?
 
Ah, you are right, but they are crushing bits that Windows can't "see".
all volume control analog or digital reduce bits,that's pretty much how they work; there's no way around it

you crush bits at the source or at the destination (amp) same thing in the end
 
Just wondering kind of settings you're using as general purpose compressor?
You'll hate them! But I don't care because I'm strong, secure, and gosh darn it, people like me. Briefly, the DEQ2496 lets me squash things _so_ much that I use the "threshold" as a volume control. With a 260 watt amp (the 3e EAUMT-0260-2-B with a decent power supply) driving Elac Debuts, I run everything WFO at a 100:1 ratio. (I have a Schiit Sys for preamp volume, but again, it's just pegged). When the threshold knob on the DEQ is -50db the sound is pretty quiet. If I slide to -40db threshold, it's a fine volume. 2.77ms attack, 30db knee (very soft rise), 85ms release. I hear all the air, ambiance, and music the resulting 3 bits of dynamic range can offer. And if a mouse farts in a corner of the building adjacent to the recording studio, if it's during a pause in the music, I'll hear it. Kidding aside, I have to tell you, I love it. No audible noise, hiss, or anything else upsetting. Try listening to Rush's "Limelight" at decent volume and 100:1 compression with decent speakers. Everything prior to this you thought was "living" will be revealed to have been a silly charade. Once Geddy's bass kicks in, you'll know -- You Have Awoken. Over-compression is secret magic.
 
@DDF A couple of doubts.
What is written in the first post is also valid on Windows inputs?
How was the loopback done to make the measurements? (because in Windows reproducing a signal at 0dbfs I don't find anything distorted by limiter)
 
Last edited:
I have to correct my previous post.

This is what I find doing the loopback measurement in REW with a 0 dbFS signal, using the Ultralite Mk5's digital loopback channel (going with the WDM driver in DS mode, and so passing through the Windows mixer first out and then in).

White line is THD and brown one is noise (whatever it means in digital loopback measurement).

loopback.jpg


Same but with Motu's ASIO driver (so Win mixer bypassed, but not Motu's internal DSP, unfortunately).
But I'm not sure that THD behaviour is DSP related and not yet present in signal... I should check but don't know how.

loopback asio.jpg


However I don't understand one thing. When I check levels it tells me that the measured signal is at -4.3 dbFS (sometimes more sometimes less by the way). But then in the graph the fundamental is at 0 dbFS (exactly as set for the test signal). And among other things it also tells me that clipping occurs.

settings.jpg


I'm confused.

Other measures will follow however.
 

Attachments

  • loopback.jpg
    loopback.jpg
    72.2 KB · Views: 88
Last edited:
I have to correct my previous post.

This is what I find doing the loopback measurement in REW with a 0 dbFS signal, using the Ultralite Mk5's digital loopback channel (going with the WDM driver in DS mode, and so passing through the Windows mixer first out and then in).

White line is THD and brown one is noise (whatever it means in digital loopback measurement).

View attachment 200201

Same but with Motu's ASIO driver (so Win mixer bypassed, but not Motu's internal DSP, unfortunately).
But I'm not sure that THD behaviour is DSP related and not yet present in signal... I should check but don't know how.

View attachment 200203

However I don't understand one thing. When I check levels it tells me that the measured signal is at -4.3 dbFS (sometimes more sometimes less by the way). But then in the graph the fundamental is at 0 dbFS (exactly as set for the test signal). And among other things it also tells me that clipping occurs.

View attachment 200200

I'm confused.

Other measures will follow however.
Did you try with -1dBFS?
Didn't check with REW yet, but I just tested a digital loopback with my computer SPDIF In/Out (it shows "Optical" but this output send on both TOSLINK and SPDIF outputs, so cable between both SPDIF RCA plugs, not TOSLINK or an internal loopback) with Multitone software and got a far different result between -1dBFS and 0dBFS (and the 0dBFS signal came back at -0.13dBFS):

-1dBFS.PNG


0dBFS.PNG


And with internal digital loopback (using VB-Audio Hi-Fi Cable) :

-1dBFS - VBA Hi-Fi Cable.PNG


0dBFS - VBA Hi-Fi Cable.PNG
 
Last edited:
Sure. Windows limiter stay away.
However still -90dbFS (note signal is at -4dbFS).
Disappointing...

View attachment 200225

Checked with REW with internal digital loopback
the results are not good with the JAVA driver

REW -1dBFS - VBA Hi-Fi Cable.PNG


REW 0dBFS - VBA Hi-Fi Cable.PNG


But if I use ASIO4ALL, it's clean at -1dBFS and at 0dBFS (even if when calibrating with 0dBFS signal, REW says that it detected clipping)

REW ASIO4ALL -1dBFS - VBA Hi-Fi Cable.PNG


REW ASIO4ALL 0dBFS - VBA Hi-Fi Cable.PNG
 
Last edited:
Yeah, similar results.

Did you disable "other APOs" with EqAPO as suggested in first post?
no, I just quickly tested the digital loopback for difference between 0 and -1dBFS, but not in the same conditions than the first post (no EqAPO and no windows mixer involved, except maybe when I used JAVA driver in REW, I don't know what it uses)
 
Back
Top Bottom