Does the room acoustics of the mixing/mastering studio have a significant impact on coloration?
Hello markus
What coloration are we talking about?? What's on the recording or what we hear in our rooms due to our own unique enviornment?? The mixing room may have an impact on the finished goods but how much?? Do we really need to worry about it on our end?? I see any coloration from the mixing room as part of the original event. In that respect it is not coloration as far as the end user trying to reproduce what's on the recording.
Rob🙂
Last edited:
P.S. No Earl, I'm not saying that this is a practical approach.
I agree, this is the ideal, but implimentation is the issue. I have heard several "wave field synthesis" types of configurations and, yes, I do hear sounds coming from directions other than the speakers, but the sounds were of such poor quality that I found the effect more annoying than pleasing. The implimentors seemed, somehow, to completely ignore the poor sound quality, instead focussing all of there attention on the "effects". Kind of missed the forrest for the trees IMO.
What coloration are we talking about?? What's on the recording or what we hear in our rooms due to our own unique enviornment?? The mixing room may have an impact on the finished goods but how much?? Do we really need to worry about it on our end?? I see any coloration from the mixing room as part of the original event. In that respect it is not coloration as far as the end user trying to reproduce what's on the recording.
Typically a microphone picks up sound and the recording/mixing engineer listens to that sound over loudpseakers in his control room. He then tries to correct for any colorations. The acoustics of the control rooms plays a major part how this corrections will look like. So yes, the mixing room is part of the original and we would need to account for it in our own listening spaces. The implications of this reproduction process are profound (see circle of confusion).
in this regard I would say that the data is not "direct" but more "implied". By this I mean that we know that Very Early Reflections (VER) cause image shifts and coloration, so one can deduce from this that we don't want VER - not a direct "proof", but pretty logical. And since in a small room is virtually impossible to get to 20 msbefore VER occur, (widely agreed and data supports that there won't be any issues at this length of delay), hence it is, again, only logical to say that in a small room one should "attempt" to extend the VER as far as possible. So is there any direct data that tests this precise hypothesis - no. Is it a logical deduction from what we do know? I think so. Now what is not clear is what the trade-off curve looks like between 1-2 ms VER and 20 ms VER, but I think that it is safe to say that between these linits there IS some tradeoffs involved.Do you have any data that supports this claim? I know the literature and I know where those guidelines historically come from but I also know that there are no studies that conclusively tell us if and how early time gap (or arrival time gap or whatever you'd prefer to name it) is an important parameter in sound reproduction. Studies suggest that level, angle, spectrum is more or equally important.
Not necessarily. If the mixing engineer EQs signals in a way so the timbre sounds correct in his room (direct sound plus reflections), there's no guarantee that the same recording sounds correct in the listener's room because the ratio of direct sound to reflections and the reflections themselves most likely differ.
I think that my comment #4 covers this comment. The "mix" is the dominate factor, not the room or the speakers, so if the "mix" does one thing then the room/speakers cannot overcome this. In other words, good speakers and rooms cannot correct for bad recordings. Nothing new there.
The implications of this reproduction process are profound (see circle of confusion).
I thought that your suggestion to use headphones for mixing was a pretty good idea. Good headphones assumed, of course. I have found some insert phones to be very good - just need to have some externalization applied and I'm all set. Ever tried that? My Chinese client and I are going to do some headphones studies on sound quality.
I agree, this is the ideal, but implimentation is the issue. I have heard several "wave field synthesis" types of configurations and, yes, I do hear sounds coming from directions other than the speakers, but the sounds were of such poor quality that I found the effect more annoying than pleasing. The implimentors seemed, somehow, to completely ignore the poor sound quality, instead focussing all of there attention on the "effects". Kind of missed the forrest for the trees IMO.
Wave field synthesis is plagued by it's own distinctive issues (see Wittek's thesis I've linked to on a couple of occasions). What I'm talking about is good old multichannel reproduction that provides a subset of listening cues. That would be good enough. The real problem is that we still have to define this "subset of important listening cues". A good start would be to define a reference level 🙂
The implications of this reproduction process are profound (see circle of confusion).
Hello markus
I am aware of his blog.
So yes, the mixing room is part of the original and we would need to account for it in our own listening spaces.
How are we suposed to do that?? Mixing rooms are all different in both dimensions and the monitors used. Even 2 identical rooms with the same acoustics and monitors can have a different house curves depending on who set it up. We are stuck with whats there you can't correct for it.
Rob🙂
Well, my comparison was between a mixing room and a listening about 1/4 smaller than the mixing room listeing distance was between 2~3 meters in both cases. I could not get the active speakers to sound as loud in the mixing room without clipping the soundcard outputs. The same computer and sound card was used in both cases. In the mixing room, the lower frequencies had to be reduced because in the process of trying to get the same level, the low frequencies were causing clipping.Not with near field monitoring.
in this regard I would say that the data is not "direct" but more "implied". By this I mean that we know that Very Early Reflections (VER) cause image shifts and coloration, so one can deduce from this that we don't want VER - not a direct "proof", but pretty logical. And since in a small room is virtually impossible to get to 20 msbefore VER occur, (widely agreed and data supports that there won't be any issues at this length of delay), hence it is, again, only logical to say that in a small room one should "attempt" to extend the VER as far as possible. So is there any direct data that tests this precise hypothesis - no. Is it a logical deduction from what we do know? I think so. Now what is not clear is what the trade-off curve looks like between 1-2 ms VER and 20 ms VER, but I think that it is safe to say that between these linits there IS some tradeoffs involved.
You know that all of this is too vague for me and we need further scientifically grounded answers to advance sound reproduction.
By the way, the "magic" 20 ms come from Beranek's studies on concert halls. Nothing that has ever been proven to be of any relevance to sound reproduction in acoustically small rooms. In the 70's more advanced measurement equipment became available and we were able to measure the initial time delay (ITD). Don Davis came out with his paper "The role of the initial time delay gap in the acoustic design of control rooms for recording or reinforcement systems". From that point on nobody really questioned his assumptions - we better should have.
I think that my comment #4 covers this comment. The "mix" is the dominate factor, not the room or the speakers, so if the "mix" does one thing then the room/speakers cannot overcome this. In other words, good speakers and rooms cannot correct for bad recordings. Nothing new there.
And still we need to find ways to overcome this.
If one want's to design something that fits reasonably into a living area, it is necessary to do that for the lowere frequencies. Remember that maximum SPL capability is closly related with effective cone area and excursion? When it's not possible to get the desired maximum SPL, and/or when the system response is rolling off, below this point is where one might want to consider extending the decay. Above this, it's still best to keep the CSD as clean as possible.This from the one who is so insistent on the need to have system response with CSDs that show the most rapid and smoothest decay possible? Room modes that extend the decay time is a good thing you say? You want certain regions to be accentuated with longer decay?
Dave
I thought that your suggestion to use headphones for mixing was a pretty good idea. Good headphones assumed, of course. I have found some insert phones to be very good - just need to have some externalization applied and I'm all set. Ever tried that? My Chinese client and I are going to do some headphones studies on sound quality.
I would use headphones only for recording tasks. Here it is important to hear any errors in the recorded signal. Mixing has to be done for the end consumer hence in an acoustical environment that is close to domestic listening spaces.
As to virtualization: there are technical solutions that would allow mixing within a standardized acoustical environment (see Smyth Realiser).
How are we suposed to do that?? Mixing rooms are all different in both dimensions and the monitors used. Even 2 identical rooms with the same acoustics and monitors can have a different house curves depending on who set it up. We are stuck with whats there you can't correct for it.
The magic word is standardization. Without breaking audio's "circle of confusion", all our efforts are just a shot in the dark.
The magic word is standardization.
I agree which is why I posted THX as far as levels. The bottom line is there is none and most likely never will be with the amount of music being produced outside the major recording studios. We are actually going the other way. The only sound industry that seems to be going in the right direction is movie sound track industry using standardized sound stages for production.
I don't agree all our efforts are a shot in the dark. It has always been this way and we are recreating an artificial event not a real one. If I can get my system as "accurate" as I can I am moving the right direction. That's the best any of us can do.
Rob🙂
It has always been this way and we are recreating an artificial event not a real one.
Either way, what we recreate could be much more accurate and reproducible with proper standards.
These determinations might easily be made with a modern implementation of Ken Kantor's "Magic" speakers for AR, varying the parameters with DSP.Now what is not clear is what the trade-off curve looks like between 1-2 ms VER and 20 ms VER, but I think that it is safe to say that between these linits there IS some tradeoffs involved.
In simpler form, I have several times suggested how we might each accomplish a substantial element of this in our own listening spaces with constant directivity speakers by varying the toe-in. Simple means for doing this from the listening location are easily devised; we can thereby optimize the balance between imaging and spaciousness to our particular pleasure.... 🙂
These determinations might easily be made with a modern implementation of Ken Kantor's "Magic" speakers for AR, varying the parameters with DSP.
In simpler form, I have several times suggested how we might each accomplish a substantial element of this in our own listening spaces with constant directivity speakers by varying the toe-in. Simple means for doing this from the listening location are easily devised; we can thereby optimize the balance between imaging and spaciousness to our particular pleasure.... 🙂
With almost any speaker by varying the toe-in.
My AES membership has lapsed, but I do recall a paper by James M. Kates entitled something like "A perceptual criterion for loudspeaker evaluation" which looked at reflection delays. I don't recall the angle or spectrum. My recollection is that coloration peaked at about 2 milliseconds (most inconvienient - right smack in floor-bounce territory). Not conclusive proof, but useful information.
There is a large and fairly consistent body of anecdotal evidence from owners of fullrange dipoles suggesting what time delay ballparks are acceptable for a full-spectrum reflection arriving from an angle not too far removed from the original source. I have owned and/or sold fullrange dipoles for about fifteen years, and I've paid attention to what other users have posted. I don't know of any fullrange dipole owners who prefer to have their speakers within less then 3 feet of the wall behind them, which would correspond to a time delay of about 6 milliseconds. Generally 5 feet is about the ballpark average for what is judged to give good results, though dipole owners will often use more than 5 feet when feasible (I believed that I could hear improvement out to about 7 feet, but not much beyond that, when I had a 34-foot-long showroom). Now this is not proof of course, but the consistency sets it apart from most anecdotal-based inferences.
Duke
There is a large and fairly consistent body of anecdotal evidence from owners of fullrange dipoles suggesting what time delay ballparks are acceptable for a full-spectrum reflection arriving from an angle not too far removed from the original source. I have owned and/or sold fullrange dipoles for about fifteen years, and I've paid attention to what other users have posted. I don't know of any fullrange dipole owners who prefer to have their speakers within less then 3 feet of the wall behind them, which would correspond to a time delay of about 6 milliseconds. Generally 5 feet is about the ballpark average for what is judged to give good results, though dipole owners will often use more than 5 feet when feasible (I believed that I could hear improvement out to about 7 feet, but not much beyond that, when I had a 34-foot-long showroom). Now this is not proof of course, but the consistency sets it apart from most anecdotal-based inferences.
Duke
Last edited:
With almost any speaker by varying the toe-in.
No, not with almost any loudspeaker. Such a loudspeaker needs to show constant directivity so the direct sound is still sufficiently flat. The majority of HiFi speaker designs is not CD.
No, not with almost any loudspeaker. Such a loudspeaker needs to show constant directivity so the direct sound is still sufficiently flat. The majority of HiFi speaker designs is not CD.
Which he would know if he bothered to do his legwork. It's tiresome dealing with the intellectually lazy.
- Home
- Loudspeakers
- Multi-Way
- Geddes on Waveguides