The real question is do you want to hear the sound with the same errors like the stereo mixing/mastering engineer did, or do you want to hear the sound without those errors?
- Elias
- Elias
What errors would that be? It's the mixing/mastering engineers job to correct errors and get the timbre right while listening in a typical stereo setup. Ambiophonics would need to undo these corrections.
"Maybe you are used to sound of the stereo combs and consider it 'natural'?"
I doubt it as I use headphones a lot.
Although that type of comb filtering is easy to reject, so it doesn't degrade the sound that much if at all.
"I don't get this. With 91us the first comb filtering notch is at 5.5kHz"
Hmm, I viewed the signal (91us) in ARTA and it has a 4.23dB hump at 4.5kHz (using only left channel PN). 3.25dB with 2ch PN.
I guess that is what I heard as tinny sound and comb filtering.
Not sure how to tell if summation is proper though.
I doubt it as I use headphones a lot.
Although that type of comb filtering is easy to reject, so it doesn't degrade the sound that much if at all.
"I don't get this. With 91us the first comb filtering notch is at 5.5kHz"
Hmm, I viewed the signal (91us) in ARTA and it has a 4.23dB hump at 4.5kHz (using only left channel PN). 3.25dB with 2ch PN.
I guess that is what I heard as tinny sound and comb filtering.
Not sure how to tell if summation is proper though.
Pardon me, but in my book, if there is something missing from the recording, then that missing thing is not there for a reason. And if you want to bring it back, you would have to do the recording over again.
I do not understand the reasoning you apply to logic. There are good recordings, there are bad recordings. There are recordings that have no mastering, there are recordings that use way to much compression. Recordings differ so much in quality, from musicians, equipment (some studios use tube gear of various kinds, some, maybe high grade transistor stuff, others use cheap noisy stuff from nameless manufacturers) and random influences, that i consider it a crazy task to try and apply some sort of filter and delay to correct a more or less imaginary problem. But then again maybe I'm the crazy one. I have this interest, regarding a little thing called sound.
I do not understand the reasoning you apply to logic. There are good recordings, there are bad recordings. There are recordings that have no mastering, there are recordings that use way to much compression. Recordings differ so much in quality, from musicians, equipment (some studios use tube gear of various kinds, some, maybe high grade transistor stuff, others use cheap noisy stuff from nameless manufacturers) and random influences, that i consider it a crazy task to try and apply some sort of filter and delay to correct a more or less imaginary problem. But then again maybe I'm the crazy one. I have this interest, regarding a little thing called sound.
Terry J: Same thing happens to me. So far I can only hear wav files on the ambiophonic player. The amb. transcoder works only with external soundcards.I emailed Stephan about this and that´s what I understand from his response.
What about using asio4all?
I only have external soundcards, well, at least the ones that are connected, would hate to restart and move myself 1,5m to connect another cable to my stereo just to test it.
I only have external soundcards, well, at least the ones that are connected, would hate to restart and move myself 1,5m to connect another cable to my stereo just to test it.
It is not really the combing they are used to. It is the central sound source (solo or vocal) that has its lower frequency level doubled compared to side sounds. This artificial stereo warmth is what is really hard to get used to not being there. There is also the reduction in Interaural Level and Time Differences in stereo that one begins to think are normal. So when one hears Ambio without the central bass boost and the wide stage due to having the full value of ITD and ITD it takes a bit of getting used to. Indeed this may be why somebody up above thought things sounded tinny. Some of the Ambiophonic VST plugins allow this central warming boost to be reintroduced.
Imagine a club of black and white photography enthusiasts. They have brought the resolution and contrast of their photos to a high level and discuss lenses, developer, and enlargement paper on their forum. Now somebody shows them a color photograph. Noboody in the club is impressed because the color masks the contrast they have struggled so hard to get just right and the color appears to have reduced the resolution or the fine detail somehow. Also they feel the color is not fully correct. So they do not add color to their hobby.
The point of this analogy is that the stereo loudspeaker triangle (not really the recordings) is an artform and thus its adherents do not have to consider as relevant the realism or dimensional fidelity that binaural reproduction (as in Ambiophonics and other methods) can produce. Like color, I do believe the addition of ITD, ILD, surround ambience (in full bore Ambiophonics) and the elimination of comb filtering does detract or distract one from the purity of tonal accuracy which many if not most audiophiles may treasure above all else.
Ralph Glasgal
Home Page
Ralph
why would the level of a central sound source be doubled? The mixing engineer will EQ sounds to make them sound "right" within the common equilateral stereo triangle. If Ambiophonics cuts lower frequency levels from a central sound then the timbre of that sound will be wrong.
Best, Markus
why would the level of a central sound source be doubled? The mixing engineer will EQ sounds to make them sound "right" within the common equilateral stereo triangle. If Ambiophonics cuts lower frequency levels from a central sound then the timbre of that sound will be wrong.
Best, Markus
If you put a divider between the speakers (no processing), the voice is stonger (seeming 3-6db louder where 10db seems twice as loud at those volume levels and frequencies), and it comes from a solid location (like the size of a head singing in front of you), unlike a 3-4' wide blob of hazy voice when you remove the barrier and seperate the speakers back into the approved "stereo triangle" setup.
The recording engineer is mixing on his setup 3-4' from their monitors at 90-100db sound levels, very unlike 99% of us who listen 10-12' away when the direct information may only be 1/10th of the sound we are hearing.
Since that revelation I've listened mainly in mono (1 speaker for years) and now am dabbling with headphones (no room reflections). Headphones are fun but not natural compared to sound in front of you you've been hearing your whole life, so headphones are not the end all be all of setups.
Norman
The recording engineer is mixing on his setup 3-4' from their monitors at 90-100db sound levels, very unlike 99% of us who listen 10-12' away when the direct information may only be 1/10th of the sound we are hearing.
Since that revelation I've listened mainly in mono (1 speaker for years) and now am dabbling with headphones (no room reflections). Headphones are fun but not natural compared to sound in front of you you've been hearing your whole life, so headphones are not the end all be all of setups.
Norman
Last edited:
ahh, that was about the only thought left that I had (ie use the player download)
thanks for that, will try later.
Remember that this boost only applies to central or mono sound sources. But unless there is a solo mic that is essentially mono, and thus a discrete file that can be processed before mixing, there is no way to equalize or reduce just the center level without affecting the other instruments of the orchestra or off-center sounds. Record producers like David Chesky who don't use spot mics cannot do anything about this warming effect in the center escept to make sure nobody in the ensemble is dead center. For large scale classical music such an adjustment is impossible.
Ambiophonics does not cut anything, it delivers the same response for side images as for central ones. Yes, it is possible that there are some stereo recordings that have been processed to death and that will sound different in Ambio. Then the issue is is different worse? You are assumming that what the recording engineer did to correct for this boost was really to audiophile standards and most tastes.
If you want to hear this effect, play a mono or solo recording of say a low male voice, and listen to it on one speaker only, then two and see if you can hear the effect. Someone once described this to me as a tenor becoming a baritone.
Ralph Glasgal
Home Page
"Indeed this may be why somebody up above thought things sounded tinny. "
No it was definitively due to the huge spikes at 4.5kHz (both positive and negative) especially at 159us and some other values giving really sharp high Q spikes in both directions.
With left ch PN 159us (0-4.5k on)
Both channels get a 10dB hump at 3143Hz due to comb filtering.
But the left channel gets a high Q negative spike of 22dB at 4425Hz
I do not understand how that can sum to an ok signal off axis, even if the on axis would.
And as I said 91us gives a hump at 4.5Khz even though comb filtering starts higher. It sums ok on axis but not 100% "natural". It is especially easy to notice on electric guitars.
No it was definitively due to the huge spikes at 4.5kHz (both positive and negative) especially at 159us and some other values giving really sharp high Q spikes in both directions.
With left ch PN 159us (0-4.5k on)
Both channels get a 10dB hump at 3143Hz due to comb filtering.
But the left channel gets a high Q negative spike of 22dB at 4425Hz
I do not understand how that can sum to an ok signal off axis, even if the on axis would.
And as I said 91us gives a hump at 4.5Khz even though comb filtering starts higher. It sums ok on axis but not 100% "natural". It is especially easy to notice on electric guitars.
This is because headphones by themselves do not correctly play back a "stereo" recording. Stereo requires a specific ammount of crosstalk to work. That is all we are doing here with any playback medium, adjusting/controlling the crosstalk (not a problem with monophonic because there is none).
If you don't like the sound trapped inside your head with headphones you need crossfeed to re-introduce an ammount of crosstalk to bring the sound outside of your head with stereo recordings. With binaural recordings you don't need it, i.e. the recording method matches the playback method.
With ambio, the speakers are close to together to improve upper end frequency response (stereo is really only "accurate" ~1.5Khz before it becomes fake and upper frequencies wander due to bad CTC spacing between L and R tweeters + our hearing mechanism sums upper frequencies to mono.) and crosstalk is controlled to bring back the spacial aspects captured in the recording either real (from the room) or fake reverb-whatever is in the recording.
If you go back into AES archives, these ideas were suggested early on, but was limited by technology. This is not the case today. It might be noteworthy that when "stereo" was being developed there were two sides, Bell Labs with their 3 channel version and the current 2 channel Blumlein idea. 2 channels was more economical in terms of producing records, equipment to manf records, and convince consumers to spend 2 times as much vs 3 times on playback equipment. IIRC, the 3 channel version was preferred but money talks.
I would suggest if you plan on compromising playback by sitting off-center, do not use high recursion settings and stick with the simple setup.
Last edited:
Inside-the-head-localization is not a crosstalk issue. We localize sounds inside our head when the individual influences of our ears, head and torso are removed (see Head-related transfer function - Wikipedia, the free encyclopedia). Klensch did some experiments (see Blauert "Spatial Hearing").
Best, Markus
Best, Markus
I think I tried playing around with the simple barrier method with two speakers years ago and didn't have great results. After discovering this thread I tried it again. The barrier was a footon mattress about 4-5" thick. The results of this were pretty subtle. Listening to some classical CD's at times I thought the timbre of instruments placed at the center of the soundstage matched better with those near the edges. Soundstage width didn't change much compared to stereo. With a narrower 20-30deg. spread between the speakers I got a similarly narrow sound stage. With a wider spread things seemed to work alright. Differences were subtle but didn't seem to be any worse subjectively than normal stereo.
Last night I downloaded the ambio_one vst plugin and tried it in my basement setup. That was a completely different proposition than the barrier! I was able to get the ~180 deg. soundstage people have mentioned. For this setup, my speakers were arranged far from optimally. I had to sit back about 16 ft. to get the 30 degree window but the results were very encouraging. I still need to play with the settings to dial things in and try a closer listening position. I wanted to post this since I saw there were at least a few others in this thread that had trouble with the barrier method like I did. If you can run one of the RACE plugins, it is definitely worth a try even if the barrier doesn't work for you.
Dan
Last night I downloaded the ambio_one vst plugin and tried it in my basement setup. That was a completely different proposition than the barrier! I was able to get the ~180 deg. soundstage people have mentioned. For this setup, my speakers were arranged far from optimally. I had to sit back about 16 ft. to get the 30 degree window but the results were very encouraging. I still need to play with the settings to dial things in and try a closer listening position. I wanted to post this since I saw there were at least a few others in this thread that had trouble with the barrier method like I did. If you can run one of the RACE plugins, it is definitely worth a try even if the barrier doesn't work for you.
Dan
Thanks for the tip about ambio.one.
It has no "tinny" sound, only comb filtering.
I have only tested it with headphones but seems to work great without changing the sound in a bad way.
I guess the JAVA based app is broken in one way or another.
If you want a VST wrapper for ambio.one to use it as a transcoder you should try vvvv : a multipurpose toolkit | vvvv - a multipurpose toolkit
Works great.
Time to try it with real speakers.
It has no "tinny" sound, only comb filtering.
I have only tested it with headphones but seems to work great without changing the sound in a bad way.
I guess the JAVA based app is broken in one way or another.
If you want a VST wrapper for ambio.one to use it as a transcoder you should try vvvv : a multipurpose toolkit | vvvv - a multipurpose toolkit
Works great.
Time to try it with real speakers.
The quote is correct. It is easy to demonstrate. Open your mouth and hum to yourself. Then bring your hands up to your ears and cover them a bit. Let us know what you hear. Now just cover one ear and let us know what you hear.
Ralph Glasgal
It is not easy to get a barrier that works down to the lower frequencies. First the speakers should be placed very close together just about the thickness of the barrier apart. The barrier wall should look like a thick large mattress place on edge and should be about ten inches thick. The mattress does not have to extend to between the speakers. It should be between your legs with your nose just an inch or so from its edge. it also needs to be tall enough to be one or two feet above your head when sitting at the end. I had such a barrier wall made by EchoBusters years ago, and the results with most large ensemble LPs and CDs were exceptional. But then concert and opera videos entered the picture and things had to change.
Ralph Glasgal
I'm not talking about bone conduction, I'm talking about the fact your landscape is trapped between your ears i.e. no depth, it does not sound as if it coming from infront of you but rather you are sitting in the middle of the band instead of the audience. Try it with speakers instead of headphones. Not enough crosstalk depth collapses, too little (cetered speakers) and width collapses. It's all about controlled crosstalk, not neccesarily removal. The recording crosstalk needs to match the playback setup, otherwise none of it works correctly.
http://www.linkwitzlab.com/headphone-xfeed.htm
Last edited:
Neither did we.
Not sure what you're trying to explain. You have to differentiate between localization of natural sound sources and phantom sound sources. The perception of depth is another issue. Crucial parameters of depth perception are (see http://hauptmikrofon.de/HW/Wittek_thesis_201207.pdf):
a) Level (sound pressure)
b) Direct-to-reverberant energy ratio
c) Reflection pattern (timing, level and directions of early reflections)
d) Frequency spectrum (for very near and for far sources)
e) Binaural differences: acoustic parallax as well as intensity/phase differences
f) Motion parallax (changes of perspective with source/listener movements)
g) Interaction with source familiarity and non-acoustical cues such as visual cues
Best, Markus