Interesting, Marc, do you have an idea of the phenomena at work?
Do you have links to a study of this well-'known' feature of our auditory system (which I did not know)?
It interests me especially since 700Hz was exactly the crossover frequency of the two channels of the best speakers that I have heard (and worked with) in my life. To comply with the 'Wife Acceptance', my personal speakers use the same driver, the same horn formula (spherical waves), but with a crossover one octave higher. Less convincing.
AD1865 has a bit of a cult following and a lot of the old Audio Note 0 x OS used this chip. I've heard a few of them but for all the good that they did, I found them just too colored. With transformer I-V, tube stage and transformer OP, stands to reason. I'd like to hear a decent implementation of this DAC chip.
It's an interesting chip in that it's not complete R2R but segmented into 15 equal weighted, most significant bits + a 14 bit (binary) R2R ladder. Lower glitch energy.
Best R2R DAC I've heard was based on a PCM1704. Came in for new USB bridge and clocks. It sounded pretty darn impressive and different to any DS I've heard. After playing that DAC for a few nights I really got why some people like them so much.
Choose your DAC's and pants carefully
For reproduction ? What do you mean ?For reproduction I am reserving judgement until I have tried it. For older recordings and esp off vinyl it may have benefits. Holman considered a width control necessary on a pre-amp so why shouldn't I
It is more simple to mix more or less the two A & B channels if you want to add some crosstalk. But who want this ?
2A = M+S = (A+B) +(A-B)
2B = M-S = (A+B) - (A-B)
M-S was invented just because it allows to get a stereo image with two close capsules in the same mike body. Easier at the end of a boom. The problem is a cardioid and a fig-8 one never have the same response/phase curves. So the result is, let's say, approximative.
Not to forget that the two capsules are usually at a different distance from the target in the axe. Not ideal.
Last edited:
Those are not quite the descriptors I would use for this effect, guess that's the engineer in me...
. However, you are the only person I have seen state this other than myself maybe 15 years ago.. We are in violent agreement.
John
Please, John, stop this violence !
(My poor English: When I said 'absolute', I mean some little delay all over the frequency range.
180 ° of phase shift is too trivial for the engineer who sleeps in you and torments you on the nights of full moon.)
Last edited:
There is a direct mathematical relationship between L/R and M/S as you note. So you can swap between them in processing easily, either in analog or digital domain.
Once you are in M/S world you a new set of things you can play with as you can adjust side and leave Mid alone. Examples being.
-For vinyl users a proper 'rumble' filter that filters on side and removes signals introduced by the cartridge. A mastering elliptic EQ for playback in effect. (and yes maybe vinyl users like that LF munge)
- Stereo width control. Useful on some recordings
- Side only filtering. Again on older recordings this can tighten up the perceived soundstage.
All of the above is known by the guys who mix and master records but totally ignored on playback. Worth experimenting with I say. And all defeatable if required.
It is about having fun after all?
Once you are in M/S world you a new set of things you can play with as you can adjust side and leave Mid alone. Examples being.
-For vinyl users a proper 'rumble' filter that filters on side and removes signals introduced by the cartridge. A mastering elliptic EQ for playback in effect. (and yes maybe vinyl users like that LF munge)
- Stereo width control. Useful on some recordings
- Side only filtering. Again on older recordings this can tighten up the perceived soundstage.
All of the above is known by the guys who mix and master records but totally ignored on playback. Worth experimenting with I say. And all defeatable if required.
It is about having fun after all?
I first noticed the effect because my early on sound card used one D/A and used two S/H for left and right. The interchannel delay messed up the soundstage and pan did not fix it properly. Images tended to not want to follow the pan correctly, some frequencies more than others.
I've also found that vocal localization center stage tended to be not so dependent on interchannel level to a point. Almost as if the source localization was "detented" so to speak, no change up to a certain force then movement.
Edit: I don't know if it was a deadband, threshold, or non linear in nature. But it didn't follow the pan law.
John
Last edited:
Don't know offhand, but at least long enough for some people to notice noise floor modulation.
In general I would tend agree that block-averaging techniques seem a very interesting methodology that could probably be used to help identify and or measure some audible effects for which fixed tone measurement techniques may not be particularly well suited. It might be nice to learn more about specific procedures for how to best go about doing it (application of useful software, hardware, etc), which might save some of us time vs having to learn by ourselves from scratch. I understand it may all seem rather obvious to someone who has been applying it for awhile.
Last edited:
Mark4... The reverb tails you refer to - if this is what it sounds like it is, it should be readily measurable? Did I miss some measurements?
On the localisation (soundstage, stereo image...) point - people hear differently. The illusion works well for some, not so much for others, it's very much a subjective phenomenon. Same with surround sound stuff. Perhaps it's no surprise that small effects would make a difference to different people therefore - but that should still be measurable.
Have you tried running mono information in dacs?
On the localisation (soundstage, stereo image...) point - people hear differently. The illusion works well for some, not so much for others, it's very much a subjective phenomenon. Same with surround sound stuff. Perhaps it's no surprise that small effects would make a difference to different people therefore - but that should still be measurable.
Have you tried running mono information in dacs?
Is that Dork Audio? I hope you noticed the antenna symbol rather than ground used on the label.
I have not made any effort to measure it. Not quite sure how to go about that either. I suppose one could record the output of two dacs that vary in how well reverb tails are preserved during playback, then try to take the difference between the two recordings. However, the two dacs may vary in a number of ways compared to each other. Autocorrelation might help find the best alignment between the two for taking the difference, but not sure what would end up being in the difference file. Also, not sure that people with certain kinds of dacs would be able to play back the difference file and hear reverb tails in that case either. At this point I don't know what obscures the tails exactly, although I think I can probably show that there is a link between more jitter and less reverb tails.
In relation to jitter, I just found a few published papers online, and a few pages in a book on S-D dacs, all of which address the effects of jitter in various subtypes of S-D dacs, and how the jitter effects may be mitigated. Will have to study those some before I can say much more.
People also have different dacs, speakers, rooms, etc. Here at my place I can provide a pretty good stereo sound field at the speaker centerline. Probably reasonably close to as good an opportunity to hear and experience the illusion as one is likely to find outside of a good recording or mastering suite. Not perfect, but I think most people could probably close their eyes and hear, say, a vocal performance as emanating from directly between the speakers.
Regarding mono mode, I have used it for certain things, but not for trying to listen for an illusory sound source width effect. That would be something I would not expect to occur in mono, but various things have happened that I wasn't expecting along the way in my dac building journey.
I get that with a ****** piano between the speakers. What I look for though is not a teeny tiny flat orchestra between the speakers, but one that appears to exist above, behind and out side them. When all things fall together the illusion can come from outside the walls of the room.
If you clicked on the stereophile link posted here a couple of days ago JA noted that he uses both LEDR and pink noise in assessing imaging. As these test files are widely available and relatively easy to report on (people can draw/describe what they hear without going into contentious language) I wonder if they might be of use here?
Don't recall that. Do you have the Link?
EDIT: Regarding a very wide soundstage beyond the speakers, I would need to apply a shuffler for that. Either that or swap the phase on one channel above bass cancellation frequencies (not such a good thing if one cares about mono compatibility, but sometimes people make records like that anyway).
Last edited:
Since you have a multimeter, try the level matched listening comparison and see if this persists.
- Status
- Not open for further replies.