The Professional Scenario is an utter mess IMO. These days, most, and especially the so called 'best' currently do their best to create an anechoic path from speaker to ear. Newell. His clone regularly achieves below 100mS down to the lowest LF.
Others do the opposite, Bogic Petrovic. Many others do something in between, with lots of reflections from the Side Walls, but angled to send HF well away from the ear.
Some designers, including one of the so called 'best' aggressively argues that absolutely flat response is optimal. This could not be further from a typical listeners room, which has 1dB per octave LF up tilt.
But it remains the case that small rooms cannot accommodate deep enough LF trapping. There are very expensive solutions such as Modex and PSI Active.
DD
DanDan this encompasses so many different issues, as you well know, and a lot of what up discuss actually highlight the real problem. When I took courses on acoustics, we studied classic control room/mastering room design. We studied people like Newell. The view at two different Universities in which I took courses (Purdue and Univ. of Illinois Urbana Champaign) was that many of the early folks and the designs they espoused were based largely on bad science and bad ideas. Not everything of course, but much of what was held as critical was based around thinking that didn’t have a lot of sound science behind it. Worse yet was some was based on science for which a full understanding of the psychoacoustics was not understood and so bad designs evolved around them (the dangers of early reflections for example). Newell is actually more of a contemporary designer and his ideas make a lot of sense in modern studios where multichannel virtual environment mixing prevails. It makes no sense for a 2-channel but they are very common. I don’t see a good argument for non-environment mixing rooms outside of multichannel or 3D artificial environment mixing. However it’s my understanding he is a proponent of non-environment rooms which were first popularized and are still commonly used for 2-channel mixing (we studied two case studies or his as examples of good and bad rooms for their intended purpose). Unless mixing for headphones I don’t see how that leads to a natural sounding recording.
LeDe was a dumb idea. I get the idea but I think it’s well established that wasn’t the way to go. It kind of makes sense in home listening rooms as long as they are larger and the speakers are out into a room.
Controlled image and reflection free seems to be more common today and make a lot more sense when done right. What I find is a critical misunderstanding of what they are about. Reflection free zone does not equal no reflections at the mixing position. It’s about mitigating acoustic mirrors that are so early in the formation of the sound that they simply color and smear the sound and image. That means that where and how a sidewall is absorbed depends on the size of the room and placement of the speaker. Controlled image simply taking this notion to the next level. You have to also take the speakers own directivity into account. A speaker with narrower dispersion naturally has less of these “bad” early reflections. Where a lot of guys go wrong is they continue to absorb “early@ reflections that are actually late enough that our brain uses them as cues for the room environment sound, and if not replaced by effect speakers, can lead to an overly dry and artificial sound. Like first reflection absorbers.
In the end, for mastering, mixing, or home listening, some environment must be in the room if it is 2 channel. Otherwise where will the environment come from? It is my opinion that non-environment rooms that were mid-designed as basically anechoic chambers have lead to some really “over-mixed” recordings that sound really unnatural.
I don’t have much experience with studio rooms that are nothing but environment, but I would imagine they can be just as problematic. If the decay times of the mixing room far exceeds what you find in real rooms then I would imagine the recordings would be too dry.
As for the flat vs room curve idea. That is so misguided that I’ve wanted to do a video and paper on the topic for a while now. If people took the time to read the AES journal articles (and not just books) by Toole they would understand where that comes from better. If they the. Read the source papers (Toole didn’t invent that idea) they would probably be able to start applying it accurately. Neither flat nor 1dB per octave is right. Those are both extreme generalizations.
Here is the science:
Speakers are all, for the most part, at least somewhat directional. That is, at some point in the frequencies the response begins to fall off in level as you move off axis. When it gets extreme people like to call that beaming. An omnidirectional microphone captures the direct and reflected sound without prejudice. It can’t tell the difference. Nor can the capturing software. All it can tell is the arrival time. What that means is that since most speakers become increasingly omnidirectional at low frequencies and since the period increases in length at lower frequencies, eventually the portion of sound captured In the steady state contains more and more reflections. The frequency response plot thus contains more energy at low frequencies than at high (which, due to less reflections, contains less energy). So that means in a room with reflections (I.e. any normal room) you expect the captured steady-state response to have a tilt. How much depends on the directivity it the speaker. The narrower the directivity over the more reduced bandwidth, the more tilted it would be. An Omnidirectional speaker would have a flat in room response.
What you hear in all those scenarios is flat because human hearing is very selective. Our ears are directional and our brain is really smart. It can tell reflections from direct sound and knows to filter out those reflections. So what we hear, tonally, doesn’t match what we measure in room.
In an anechoic chamber or any environment for which there are no reflections, a speakers response should be flat. There is no excess energy from the reflections in the steady state because there are no reflections.
Now that means in a non-environment room that is basically truly anechoic flat night in fact be a better match to what people hear in their room. But let’s be realistic, how many non-environment rooms are tricky anechoic at all frequencies. Anechoic chambers aren’t anechoic at all frequencies. Most anechoic chambers are physically too small to be much good below 100hz, 50hz at best. Basically the giant ones for testing rocket ships or cars are about the only ones good down to 20hz or below. That’s because it’s impossible for acoustic absorbers to reasonably absorb all of the sound. Some of it always reflects back, even at mid and high frequencies, but especially low frequencies.
So I would argue that there is a desirable room curve for a non-environment room and it’s neither 1dB per octave rising nor is it flat. It is, instead, a function of the speakers directivity and the rooms actual absorption characteristics at the mix position. And quite frankly nobody knows what that should be, but it’s also unimportant. What you want to do is simply build a speaker that has a flat listening axis response In the anechoic chambers environment, and whose DI is as flat as possible as low as possible in at least the frontal hemisphere. The level of the DI is a matter of debate, but it is my opinion that 0-1dB is just too wide dispersion and provides a very diffused image that is not realistic. That 2-5dB is in the range of most normal speakers and provides a good natural image. That 6-10 is what we often call controlled directivity and provides more pinpoint imaging (something I think is desirable, especially in virtual environment reproduction), and anything above that is special purpose for the most part. The desired room curve then is whatever the speaker naturally makes, you don’t fit a speaker response to a room curve, it’s a property of the speaker and room interacting as captured by a flawed ear proxy. Eq is reserved instead for removing the bumps that are caused by SBIR and modes, or actual flaws in the speakers own response.
