Foobar2000 Dolby Headphone config - Comment & discuss! - Page 9

Quote:

Originally Posted by RPGWiZaRD 

Yea I know about the gain, I've already adjusted all the presets to the point I won't hear any distortion when listening to the songs with the "closest" to clipping point I got, I don't want to lower the whole graph so that the highest peak touches 0-level as it results in an audible sound quality loss in my ears, may be due to the fact I'm not using WASAPI and using the Windows mixer, don't know but I'm sure there's a difference in the result even if I adjust the volume levels to make up for it and since there are more non-bit perfect listeners out there than me that argument becomes a moot point anyway, I always decide all the changes by my ear as I don't put too much faith in theories. With Electri-Q the distortion seems to be appriopriate to the "area" of the whole curve, if there's a range going below 0-level another can be a bit above and still won't result in clipping so therefore different presets of mine can have some of the peaks sticking above the 0-level by different amounts and still sound as "clean". I think the best compromise in this case is to do what I did, just adjust it to the point the most distortion-prominent tracks remain distortion-free by a subjective analysis.

 

I don't like ReplayGain since it modifies the files themselves, I just don't want any modifications to be done to the files at all, they should have the same checksum-match as before even.

Go download this frequency sweep wav file

http://www.mediafire.com/file/dxy7xevh8muwrhb/full%20range%20full%20volume%20sweep.wav

and try to play it back with preamp at 0dB and just the equalizer in the DSP chain.  If you manage to put any point above 0dB without the corresponding part of the wav setting off the clipping alarm on the peak meter, I'm all ears.

Now granted, you may manage to scrape by your music collection with your current settings without setting off the clipping alarm, but that would be just luck, that your music collection doesn't go near 0dB for the parts of the frequency range that you are boosting, even for the loud parts.  Then again, I have had mp3s set off the clipping alarm even when everything on the equalizer was <0 and I'd already dialled in a -2dB preamp, to boot.  Like I said, mp3s encoded without Replaygain can go way past clipping point on decoding, but these can be rescued by ReplayGaining or setting the preamp setting.

Replaygain does not modify the content of the music file, it only adds a tag that tells the player to interpret all the output with a digital multiplier.  The only possible audible effects of changing replaygain are

1) Correct replaygain restores peaks that were clipped off before (unless the clipping is in the recording itself)

2) Negative replaygain raises the noise floor by the corresponding number of decibels (the same thing happens with lowering the curve on Electri-Q

Now, right now I'm just listening straight out of the notebook headphone jack, and could actually use a lower noise floor.  But the only audible effect of the noise floor for me even now is a bit more noise at the beginning and end of music tracks.  Can you hear the noise floor on your current setup now?  Do you hear the noise floor going up if you crank up the main output volume on the windows mixer, or any other analog volume controls you have? (within your *normal* listening volume range, not "I can hear this amp noise if I crank all the knobs to 11.  Oh, and now if I hit a wrong button I'd blow my eardrums out of my head")  If you do, you could theoretically benefit from making the digital signal as hot as possible and setting the main output volume and other analog volume controls as low as possible.  But then if I were doing that I would be so worried about clipping, I would be watching the peak meter on foobar like a hawk instead of of finding time to actually listen to my music

BTW, AFAIK WASAPI doesn't avoid the problem of raising the noise floor.  You still need to modify the digital signal for your EQing and Dolby Headphoning, modifying the digital signal is the whole *point*.  WASAPI just means the modified signal is then sent straight to the DAC rather than going through the windows mixer.

BTW2 here's the EQ curve I'm using to equalize my headphones, as detailed in my review of my current headphones:

Anyway, someone else reading this thread could easily modify your presets with a few clicks depending on who he believes in.  I don't think it's a big deal, it's not like clipping is what's putting me off Dolby Headphone.

Quote:

Originally Posted by RPGWiZaRD 

Yea I explained the loading preset bug a few times already and to avoid crashing you have to load the "Custom: Stock foobar" or "Custom: No EQ" or one of the bottom 3 presets everytime in-between you switch to another as the crashing occurs when there's Electri-Q plugin in the active DSPs list in both the current used preset as well as the loaded preset setting.

 

If you just want a wider soundstage, remove Dolby headphone plugin and play around with the Stereo image width in channel mixer and use the default settings for everything else (disable surround etc).

Well, I'm not talking about changing between DSP chain presets.  I'm talking about going into Electri-Q, hitting the "M" button, and selecting Presets->Import Preset... but thanks anyway.

Does the stereo image width setting have an effect even without Dolby Headphone?  Then that sounds like it would mix the left and right channels for settings <1 and cancel out the parts common to the two channels for settings >1.  Not really what I want if that's what it is.  Is there a readme for this thing?  Thanks.  (don't look too hard for it though--I don't promise to try this again.  I'd probably try Isone first...  Not your fault, you've really overdone yourself preparing all this for Dolby Headphone and I commend your efforts)

Even small amounts of clipping cause very noticable distortion and static-like noise in my setup. While this is mostly mitigated with compression (and limiters are compressors), which I don't use, you reduce the dynamic range, that already was reduced to embarassing levels in most of todays music.

Sure, someone that is not into sound and sound quality doesn't care about this but if you listen closely it certainly is audible and easy to ABX.

Hey RPG, I've been testing several upmix plugins: Neural Upmix, Starplugs plugins, V.I Suite, and FreeSurround. Neural Upmix and Freesurround provide a wider soundstage but I feel this kind of sound is better suited to watching movies (I tried the default setting for both plugins). Starplugs plugins have more echoey/reverb than channel mixer and a more wider soundstage than Neural Upmix and FreeSurround plugins. Years ago I had tried the V.I Suite but the width slider always returned to it's default setting and the Foobar seek function not working properly... but now works fine... and for my ears is the best sounding upmix plugin. Almost no echoy/reverb and with a more balanced soundstage... if you have not tried it yet maybe you should do... kudos for FreeSurround's depelover because his plugin sounds almost identical to Neural Upmix but this is a commercial plugin!

I had been testing the Isone Pro plugin with V.I and several VST plugins: Oxford EQ, Trident A-Range, SPL TwinTube, Sonalksis Ultimate-D and Sonalksis FreeG, mostly after reading too much in this forum . Isone Pro gives me a better instruments separation in each song but  V.I with Dolby sounds great too and this setup is completely free!

Finally I wanted to ask you if you were so kind to upload the Electric-Q presets only (not the DSP chain presets). As I understood the EQ is to balance the use of the dolby plugin right? And as I have a Sony XB-500 like yours I think the Electric-Q presets might suit my setup very well.

Thanks.

I'm using your settings with my LCD2's and I really enjoy it. The resolution seems to suffer a little, as if there was a thin veil over the music. But it's so thin that I think it's worth. One of my problems with LCD2 was it's cramped soundstage. Now things are different. I wish I could afford an amp for 1500 $ that  had a wide soundstage, but this is more economic compromise..

I'm looking forward to further development of these kinds of projects, both software and hardware. I mean new clear sourrund EQ's and products like the Smyth Realiser(but cheaper!)

Quote:

Originally Posted by RPGWiZaRD 

 

Yea this veil depends how the EQ is adjusted to make up for the Dolby Headphone "frequency response"-veil from the veiled treble-range (and boosted low-range), I tried fixing this with Electri-Q as good as possible I think I'm pretty close. I used to use another default EQ preset that might have sounded tiny bit less "veiled" but it also provided some distortion due to not having lowered the gain at all and had the treble range going to about ~+1.5dB above zero. As I lower the gain I also hear that I'd need to boost treblerange further to make up for it but still have to try avoid distorting. With my own hardware DSP EQ in my soundcard this would be such an easy fix as it won't distort at all even going a huge amount above 0-level without lowering preamp but unfortunately software EQs like Electri-Q doesn't quite work like that and it's a constant battle for "distortion vs veil". I'd need to be able to boost the frequencies like on my hardware EQ without getting distortion to completely take care of it.

Man, I really think this has to be all in your head, just wish there'd be some way for you to A/B test it yourself.  Maybe find the two volume knob settings that would match output volume for two different preamp settings and have someone A/B the two preamp / volume settings for you while you turn your back?

Or, it could be a problem with how the EQ is interacting with the Dolby Headphone plugin.  See, the plugin takes its input to be the source material and applies its HRTF accordingly.  So if there's something in the algorithms that takes an input at x Hz and outputs a signal at y Hz, and your EQ tones down x Hz by 3dB, you'd also toning down the output at y Hz by 3dB (at least the part added by Dolby Headphone) unitentionally.  I don't know how the HRTF model would actually do something like that, it may be some other kind of interaction I'm missing, but I know I did hear quite a different sound when I placed my radical headphone EQ at the front vs at the back of your DSP chain.

So you could try moving your EQ to the back of the chain and see if the preamp still makes a difference to the sound signature?

I don't really see how there could be a hardware EQ (as in a physical RLC network) that knobs on the computer screen could be changing.  It's probably just another software EQ, but seamlessly integrated with the analog volume control on your sound card so that when you push a slider up 5dB, the sound card dials up the volume output by 5dB then applies -5dB to every other frequency band.

Whoops, nevermind.  Going to take a lot of tweaking to get exactly the sound I want, but so far I am very much liking the default "out of the box" sound with the Graphic EQ over what I had been listening to.  Having it out of my head is a pretty huge plus.  Now I need to get it to sound almost as accurate...

Right now, listening to Opeth FLACs just don't sound that good.  They sound fine until the growls happen, and then it's all jumbled into one huge muddy mess--for the rest of the acoustic and clean vocal parts though it sounds fine.  No idea why it does that to the growls.  Switching to something like "Dante's Inferno" from Iced Earth however is a "Whaoh!" moment. It almost sounds like a completely different song, and not in a bad way.

Hey RPG,

I have Realtek HD Audio AC622. I am using your config for Realtek Onboards, that you have posted on the fourth page of this thread and I have to say it sounds really great and much more natural than SRS Audio Sandbox (had to uninstall this). The only thing I corrected was the master gain. Had to put this to auto because of distortion and too high peak levels.

However, with SRS the sound used to be like thrice louder, so, I have to put all the bars up to get ~medium loudness. Any suggestions how to improve loudness? Also, there wasn't any distortion no matter how loud my headphones played. I have no idea what to do as there is an awful distortion after I increase master gain just by 0.5 dB?

Thanks

Nevermind about the loudness, I am so dumb, the "Wave" bar was like at 30%  After I replace your Core.cfg with the default one, where can I find the presets? In the preferences > DSP manager > DSP chain presets they are not displayed. Also, the peak level is above 0 dB with Electri-Q preset, should I just apply Advanced Limiter for that?

Sorry for being a newbie and thanks!

subscribed