Having done both I would disagree. GLM also doesn't appear to be using a simple algorithm either (like rew).Once you see what GLM does, it is probably easy to re-do it in REW.
Having done both I would disagree. GLM also doesn't appear to be using a simple algorithm either (like rew).Once you see what GLM does, it is probably easy to re-do it in REW.
worth noting that for guys where accuracy is a key concern - because they're recording and engineering the album and transforming it into a finished product - almost exclusively work on bookshelf sized speakers. Then they hand it to mastering engineers who transform it a bit further, generally using floorstanders or very large main monitors.Exactly. And that's assuming the mics were placed far back at the listening position as opposed to placing a close mic at each instrument and mixing it all later. And let's not even get started with vocals recorded in a recording booth and mixed together with synthesized accompaniment later. The point is the concept of "reproducing the original recording" is an abstract concept that isn't readily measurable or reproducible. We should stop referring to it as a desired state.
There is a reason that Harman research tries to solve for "listener preference" and not "accuracy" (another horribly undefined term).
Point source is ideal for imaging and gives you the most accurate reproduction of the source material, including its space. The big sound with an exagerrated sense of height that you get from towers like the Salon2 is more pleasing for classical music and movies. It is just a bigger, more impressive quality of sound that makes the content more immersive and grand. I'm not sure quite how to describe it. It isn't quite as good for properly reproducing the space and objects within it, but it is good for creating an effect of spaciousness.
However, when you add in surrounds and height channels, you also get a huge sound from small speakers like the Genelecs. What sounds better in a full immersive audio setup is another question. I would argue that point-source coaxials are better for this because they can more accurately reproduce height information with their good vertical directivity.
You mean in an immersive setup? Otherwise the one has nothing to do with the other.I would argue that point-source coaxials are better for this because they can more accurately reproduce height information with their good vertical directivity.
I disagree a bit, though that's possibly semantics. As many have noted, reproducing the original performance from 2 channels is impossible, possibly excepting an anechoic performance. (Isn't that how Dunlavy recorded his "live or Memorex" demos?) Accurately reproducing the recording is something else, however. Because stereo theory has been studied at some length, we can predict what certain recording techniques can & should yield. Even more so for synthetic signals.The point is the concept of "reproducing the original recording" is an abstract concept that isn't readily measurable or reproducible. We should stop referring to it as a desired state.
Yes, that's what I mean. I would argue that coaxials would be better for an immereive audio setup because their controlled vertical directivity allows for a proper reproduction of whatever the vertical equivalent of stereophonic sound is. Just as a wonky horizontal directivity response would interfere with stereo imaging, so would the wonky vertical directivity response caused by acoustic lobing interfere with vertical imaging in an immersive audio setup.You mean in an immersive setup? Otherwise the one has nothing to do with the other.
Sorry I read what you wrote 3 times and I couldn't find where you defined "accurate reproduction of a recording" by loudspeakers.I disagree a bit, though that's possibly semantics. As many have noted, reproducing the original performance from 2 channels is impossible, possibly excepting an anechoic performance. (Isn't that how Dunlavy recorded his "live or Memorex" demos?) Accurately reproducing the recording is something else, however. Because stereo theory has been studied at some length, we can predict what certain recording techniques can & should yield. Even more so for synthetic signals.
Record something with known instrument locations, perhaps a string quartet, with no processing. The recording will differ if you use crossed dipole mics versus spaced omnis. Etc. Playbackmaywill not match the actual performance's spatial characteristics, because not all of that information is on a stereo recording. But knowledge of the original location & mic configuration can tell us a good bit about what information can be on the tape, and the spatial illusion which should be created on playback.
Sound absurd? Consider an absurd example. Record a single instrument in an anechoic chamber in mono. Playback over omnidirectional speakers in a tiled room will fail completely at accurate reproduction. Big horns in a damped room will be closer, both to the live performance & what was recorded. Sure, this is an extreme example, but the point is that we don't need to surrender to "if you enjoy it, it's hi-fi!" as soon as the discussion moves from FR/distortion/noise to spatial details. At least for some recordings, we do have enough info to make more complete judgments.
This is just my understanding, not being an experienced recording guy. IIRC, @John Atkinson posted a pretty good article about the subject a few years ago.
How would you define the term "accuracy" as you used in your response?worth noting that for guys where accuracy is a key concern - because they're recording and engineering the album and transforming it into a finished product - almost exclusively work on bookshelf sized speakers. Then they hand it to mastering engineers who transform it a bit further, generally using floorstanders or very large main monitors.
So yeah, listener preference is the way to go. At some point you just have to decide which you like better and then go for it. If you can't decide, get both and have dedicated spaces for each![]()
Wish it would be as simple reciprocal like that as there are also different radiation patterns. You also wouldn't want to record usual music on array mics and most music is created by a mix of many different near instrument and far field mics which are mixed together to a new art product "stereo recording" as something that gives some kind of realistic illusion when played back in stereo but has usually not much in common with the original cues.If the sound were recorded on line array microphones, the most accurate way to reproduce the original recorded sound would be to play it back on line source speakers.
Fully agree. Also, if we take into account close-miking techniques, sound synthesis and blending of multiple microphones for the same source, we quickly see there is no easy way to define any one reciprocity principle between sound recording and reproduction.Wish it would be as simple reciprocal like that as there are also different radiation patterns. You also wouldn't want to record usual music on array mics and most music is created by a mix of many different near instrument and far field mics which are mixed together to a new art product "stereo recording" as something that gives some kind of realistic illusion when played back in stereo but has usually not much in common with the original cues.
Also you would need loudspeakers with the same exact directivity of the used stereo mics and even there you wouldn't be able to reproduce the original sound field as the 3 dimensional information is lost when captured/reduced to a single dimension per channel.Perhaps if we only spoke of standard stereo microphone technique recordings (without any spot microphones) would this reciprocity hold true - but this kind of recordings are really not the norm for many music styles today (or in general for any kind of stereo content recording).
Adding no additional coloration past what the rooms, microphones & their positioning, and any gear in the chain have already added. So basically, a very flat response.How would you define the term "accuracy" as you used in your response?
I have yet to hear Salon 2's. My Room is L-shaped, I sit in the corner of the L and anything in the front has a huge problem with ~63Hz (ish) standing waves. So being able to place a sub in a more ideal position, before room correction (I assume) will be preferable. I always like to see if moving around my loudspeakers and sub brings me closer to my target response until I either reached the optimum placement or a borderline acceptable placement in the room.I am running 8351B's crossed over with a pair of JTR RS1's. I still prefer the sound of Salon2's, although they are both very nice.
Adding no additional coloration past what the rooms, microphones & their positioning, and any gear in the chain have already added. So basically, a very flat response.
Interesting. How do you define "coloration?" How do you measure it?
And when you say flat response, is that a flat anechoic response? On-axis? Averaged over a range of axes? And what about reflections from walls? Or does playback need to occur in an anechoic chamber?
Corrected to be as flat as possible in the listening area, or a purpose built space that needs minimal correction. So that'd be averaged around where the engineer's head is going to be most of the time. RT60 needs to be very short because you will be adding reverb in the mixing process so the room is likely heavily treated or built to direct reflections away from the listening position.
So no, not an anechoic chamber, but it would be significantly deader and flatter than most consumer listening spaces. And it all gets easier if you start with a speaker that's flat, on axis, measured anechoically... hence the chart of "best speakers" on here being dominated by studio monitors.
I didn't mean to imply that I have all the answers, just that a recording is distinct from the original performance. That said, I feel that it's reasonable to consider a speaker's performance as we would an amplifier's: the output signal should be exactly proportional to the input signal. If any changes are inaudible, that's probably close enough. Yes, I realize there are a million variables with directivity, where & how you measure, etc.; but it seems difficult to argue it's the wrong goal. And, yes, we have a long way to go.Sorry I read what you wrote 3 times and I couldn't find where you defined "accurate reproduction of a recording" by loudspeakers.
I think the only way to solve your problem will be more subs.I have yet to hear Salon 2's. My Room is L-shaped, I sit in the corner of the L and anything in the front has a huge problem with ~63Hz (ish) standing waves. So being able to place a sub in a more ideal position, before room correction (I assume) will be preferable. I always like to see if moving around my loudspeakers and sub brings me closer to my target response until I either reached the optimum placement or a borderline acceptable placement in the room.