frequency response and phase response respectively. i did include the legends at the bottom but I'll admit they are pretty hard to read!
The wiggles did remind me of the shufflers as well, especially the 0.5ms whose phase response starts wiggling around 1.5kHz.
@ dynomike
Well thanx for taking a byte, or two, & doing the graphs etc
I hope to download your .wav files on sunday from another location, as right now it's a bit slow for some reason ?
I fully expected the much shorter delays to be better. You might like to try the 0.5ms, or even less mid delay, & experiment with Eq & reduced level. For eg, a few db of HF boost with the level reduced from between 10db - 15db blended in with the stereo track is quite interesting ! It doesn't sound dark to me, & the wiggles are not audibly apparent either !
Well thanx for taking a byte, or two, & doing the graphs etc
I hope to download your .wav files on sunday from another location, as right now it's a bit slow for some reason ?
I fully expected the much shorter delays to be better. You might like to try the 0.5ms, or even less mid delay, & experiment with Eq & reduced level. For eg, a few db of HF boost with the level reduced from between 10db - 15db blended in with the stereo track is quite interesting ! It doesn't sound dark to me, & the wiggles are not audibly apparent either !
Look at it this way: all the sound arriving at your ear gets processed by your brain. That brain has been trained all your life to make sense of sounds. It is very capable of making sense of combed signals to catch a conversation. So even though you're not hearing it doesn't mean it's not there.
Your brain will work hard to make sense of the altered sound you created and combine it to something you can interpret. That's what I meant when I said you're kind of faking an added room effect to the center part. The combing and phase deviations make you hear it differently. Part of the brain will actually try and ignore the added effects.
To get some grip on this, listen to your stereo at the sweet spot. (no need for adding any effect). Listen to the music and try to remember the sound you hear. Now record that sound at that sweet spot, and play it back trough headphones. You won't believe the amount of room you will be able to hear listening trough the headphones. You most probably didn't notice the room all that much when you were sitting in the room in that sweet spot.
I once played a recorded song (recorded at the sweet spot) back trough my speakers, doubling up on the room effects. It took me quite a while to unlearn hearing my room after that! But something like this can make you learn to listen for room signatures.
Your brain will work hard to make sense of the altered sound you created and combine it to something you can interpret. That's what I meant when I said you're kind of faking an added room effect to the center part. The combing and phase deviations make you hear it differently. Part of the brain will actually try and ignore the added effects.
To get some grip on this, listen to your stereo at the sweet spot. (no need for adding any effect). Listen to the music and try to remember the sound you hear. Now record that sound at that sweet spot, and play it back trough headphones. You won't believe the amount of room you will be able to hear listening trough the headphones. You most probably didn't notice the room all that much when you were sitting in the room in that sweet spot.
I once played a recorded song (recorded at the sweet spot) back trough my speakers, doubling up on the room effects. It took me quite a while to unlearn hearing my room after that! But something like this can make you learn to listen for room signatures.
Taking another look @ the graphs based on the above, just the mono mid, what are the two graphs depicting exactly ? As i'm trying to figure out why there are wiggles, if you only delayed the mid but did NOT blend it in with the stereo track.
@ wesayso
Yes i see what you mean about the brain & room signatures etc. Interesting experiment you did with the recording. But i'm sure it would have sounded different if you had used "dummy head" mics to record, which simulate our ears. I'm presuming you didn't anyway ?
******
I've now delayed the mono track to 0.01ms & EQ = +2db HF, -3db MID, -6db BASS. Even with the mid level down @ -24db relative to the stereo track, i can hear an improvement. Higher levels naturally are more apparent, but i still don't hear anything nasty etc !
It's a mono simulation - resultant frequency and phase response of the delayed track mixed at -6dB with the original. You're right, if it was just the delayed version there would be no difference in the frequency response. It's the combination that is problematic. You could actually separate the mid and side, and delay the mid channel only (wesayso tried this) but it really destroys the stereo image. Certainly it is effective to decorrelate the center from the sides, but that isn't actually a good thing.
Apart from being 0.5dB louder in the mid channel, I suspect the differences due to delay or EQ are not audible at -24dB. On the other hand this paper concludes that the effects are audible at lower than -20dB. Still, effects of electrical comb filtering (before the sound leaves the speaker) are negative so...
I think you will have more significant results experimenting with Mid/Side EQ (specifically to compensate for narrowing low frequencies through side channel EQ, and/or comb effects in the center through mid channel EQ) as wesayso has done, or the rePhase shufflers (if you are in a low-reflection environment).
As the simulations show, blending a delayed copy has objectively negative effects - a smeared time response and comb filtering. Of course simulations aren't everything. In this particular case though, it is known that electronic comb filtering is to be avoided, and its audibility at say -6dB can be confirmed. If you can't hear it that's fine, but why do the damage?
It's interesting to me that a few people are aware of the dull center effect, and others claim to have never heard it. To me, it's obvious. But then I remembered most of the photos I see of speaker systems on this forum and others. Rooms with many hard surfaces, no carpet, low ceilings, bare walls, large glass areas, no room treatments. It all looks clean and slick, but can't sound great.
I suspect that many (most) listening rooms have so many reflections that the comb filtering on the center image is simply erased. This may also be the case in my lava cave listening room, where all surfaces are hard, but none are even close to parallel.
I suspect that many (most) listening rooms have so many reflections that the comb filtering on the center image is simply erased. This may also be the case in my lava cave listening room, where all surfaces are hard, but none are even close to parallel.
Toole is of the opinion that (certain) reflections solve more problems than they cause, and I tend to agree. I still have an interest in "low reflection" reproduction, but I wouldn't say that largely reflective rooms MUST sound bad.
The duller phantom center can really only be properly evaluated in situ. with 3 identical speakers (or a dummy head recording of phantom vs. real center, moving one of the speakers to the center for comparison). Yet, playing with the shufflers made it apparent to me that on nearfield speakers the center has a noticably incorrect tonality. It's just something you may not notice until you hear it corrected.
Looking forward to trying these out on the line arrays once I get those corner reflections dealt with. I have a bunch of panels and diffusors ready to install this week!
The thing I am wondering about if the avoidance of early reflections is also the perceptual difference between a "They are here" and "You are there" type of sound.
As I was shooting for that last one I worked hard to get my first 20 ms as clean as possible (within the boundaries of my listening space). It was also one of the reasons for me to go with line arrays. Another option I had considered using was Synergy type horns.
I did try and get Synergies past the beauty commission here (my Girl did not like the look of any of them at all) but no luck.
I'm still amazed what the shuffler can do compared to mid/side EQ alone. I thought I was pretty much set with my mid/side EQ until I tried that second RePhase shuffler.
As I was shooting for that last one I worked hard to get my first 20 ms as clean as possible (within the boundaries of my listening space). It was also one of the reasons for me to go with line arrays. Another option I had considered using was Synergy type horns.
I did try and get Synergies past the beauty commission here (my Girl did not like the look of any of them at all) but no luck.
I'm still amazed what the shuffler can do compared to mid/side EQ alone. I thought I was pretty much set with my mid/side EQ until I tried that second RePhase shuffler.
Toole would say it depends on the loudspeaker.
With a well behaved loudspeaker (based on his spinorama criteria) his advice is to either use thick treatment at the early reflection points (7cm to 10cm, down to Schroeder frequency) or leave them alone altogether.
His idea is that it is better to get a "coherent" early reflection than a low-passed-ish one.
He also claims that a fair amount of reflections (early or not) are a good thing for hifi stereo reproduction (so called "spaciousness"...).
It certainly helps mitigating problems with phantom image and other discrete comb filtering effects.
The problem with addressing early reflections is that you cannot entirely (and coherently) eliminate all of the them (floor, etc.), and the remaining one(s) will become more and more obvious as you go.
Last edited:
Yes, surely. Mostly I was commenting on the very reflective rooms I see posted so often. I don't think they would sound good, but they should at least mask a lot of problems.
I currently have what is probably a unique listening space, the lava cave. Nothing but reflections off hard surfaces - but the surfaces are so chaotic and non-parallel that any correlated reflection doesn't have a snowball's chance in hell of being heard. It's an unusual environment that brings out the hidden acoustics of recordings. Quite surprising! And the chaotic reflections do a good job of fixing the phantom center.
In normal listening rooms I'm a fan of diffusion behind the speakers, absorption behind the listener, a combination of the two in-between. But I seem to notice the dull center most when the area around the speakers acoustically dead. After reading the research, we know what it's the band above about 1.5kHz that matters for the phantom tonal balance. Making that a dead zone will accentuate the problem. I've certainly done that with room treatment.
Mike,
You've asked me a couple of times about my mid-side EQ. As I was playing around in RePhase and JRiver and plotting my output I might as well check and show what the mid/side balance looks like at the sweet spot. It's only minor EQ but it does work for me.
This is left+right channel (5 cycle Frequency Dependent Window) at the listening position, run once trough the mid EQ and the other trace trough the side EQ. Only a minor difference between both channels as you can see.
The perceptual change this brings is preferred by me. But as with the shuffler it only works in absence of (early) reflections.
You've asked me a couple of times about my mid-side EQ. As I was playing around in RePhase and JRiver and plotting my output I might as well check and show what the mid/side balance looks like at the sweet spot. It's only minor EQ but it does work for me.
This is left+right channel (5 cycle Frequency Dependent Window) at the listening position, run once trough the mid EQ and the other trace trough the side EQ. Only a minor difference between both channels as you can see.
The perceptual change this brings is preferred by me. But as with the shuffler it only works in absence of (early) reflections.
Thanks. I assume the darker trace is the side channel, with the LF boost and HF cut?
I personnally do not have any problem with the tonal balance of phantom images... with stereo musical programs.
But i also agree that speech intelligibility suffers... with movies.
Maybe a pair of stereo loudspeakers are simply not adequate for home cinema reproduction, because movies have not been produced listening to a pair of monitors and are intended to be reproduced with at least a dedicated center channel...?
Btw, i don't understand the comb filtering issue, or, as quoted in the AES paper, their importance is relative, or can be made relative:
I also like pretty live rooms, and listenning not too close to the speakers! But of course others will prefer heavy use of absorbing treatment tending to anechoic conditions...
But i also agree that speech intelligibility suffers... with movies.
Maybe a pair of stereo loudspeakers are simply not adequate for home cinema reproduction, because movies have not been produced listening to a pair of monitors and are intended to be reproduced with at least a dedicated center channel...?
Btw, i don't understand the comb filtering issue, or, as quoted in the AES paper, their importance is relative, or can be made relative:
I also like pretty live rooms, and listenning not too close to the speakers! But of course others will prefer heavy use of absorbing treatment tending to anechoic conditions...
Last edited:
Yep, I think we would all agree with the comb filtering issue - this is a special case though, related to the phantom center, having the same signal reproduced by both speakers is what causes this unique issue. Not talking about comb filtering being an issue in general.
I agree that a real center channel must be preferable for movies since that matches the mix conditions. This is a kind of hack to try to match that sound closer, which may also have positive effects on stereo material. As said before I am now leaning towards a reflective room as the best defence against this issue for stereo material.
What's interesting is that a lot of music is mixed in acoustically dead environments on nearfield speakers... yet we still don't find listening to it that way preferable. One argument is that recording engineers prefer a much lower level of reflections on playback, to better examine the recorded elements. Yet when listening for enjoyment, we don't need to be inspecting the mix to the same degree. Still there are times when I like to really inspect the mix as I find that to be part of the "performance" - the engineer's performance. Casual listeners, when given the choice, will turn reflections up much louder than they naturally occur, even. (Toole)
I agree that a real center channel must be preferable for movies since that matches the mix conditions. This is a kind of hack to try to match that sound closer, which may also have positive effects on stereo material. As said before I am now leaning towards a reflective room as the best defence against this issue for stereo material.
What's interesting is that a lot of music is mixed in acoustically dead environments on nearfield speakers... yet we still don't find listening to it that way preferable. One argument is that recording engineers prefer a much lower level of reflections on playback, to better examine the recorded elements. Yet when listening for enjoyment, we don't need to be inspecting the mix to the same degree. Still there are times when I like to really inspect the mix as I find that to be part of the "performance" - the engineer's performance. Casual listeners, when given the choice, will turn reflections up much louder than they naturally occur, even. (Toole)
I think that mixing this way is important for critical listening, but also the professionnals get more exposed to the tonal issue with phantom image, so that this issue is better adressed and corrected in the recording itself because it's part, a consequence, of the mixing conditions.