John Sayers' Recording Studio Design Forum Archive

Hi Stuart,

Two improvements to my studio since last REW data that you saw:

1. Somehow, despite (I thought) being really careful, I managed to have my speakers wired out of phase. OMG... Much better definition now... duh

2. I have six 2 x 4 foot RealTrap minis (bass traps) around the room.

Here is a link to new REW data (today, 8-1) as well as the previous that you saw (7-12), and some photos. Any better at all?

https://we.tl/bxDUESSYo5

note 2 RealTraps across the front corners, barely visible one below each speaker.

I did a few mixes, and they actually worked out really well (even upstairs and in the car lol!). My perception of spacial definition at the listening position is definitely still that it could be better. I need to deal with those early reflections. I still need to deal with the glass window on the right side.

I am doing the REW testing with the pink noise at 75db spl. I'm sure i mix mostly louder than that. Recording data louder would exacerbate the reflections, wouldn't it? What level should I use for REW?

You said my worst reflection was 1.4ft from the speakers. The speakers are mounted with the HF drivers down and in - probably more like 3ft to the ceiling. What about the mirror in the center, between the speakers?

Thanks, getting there very slowly...

Joe

Hi Stuart,

I believe I have physically treated my studio room as much as is feasible - see photos. I have six 4' x 2' RealTrap bass traps spaced from the wall ( 2 are suspended in front of the dark wall beneath the speakers, hard to see). The reflective mirror between the speakers and the glass window behind the keyboards have been treated. Recall that the entire (non-parallel) ceiling, that you called a "compression ceiling", was designed to be one big bass trap. Sound staging and definition are immensely better!

Would a Dirac room correction device such as miniDSP DDRC-22A be a good final addition?

I realize it is probably possible to achieve most of the same results manually if you know what you are doing, but since I am not an engineer, it seems from what I have been reading that the Dirac units do a pretty good job of correcting room EQ and impulse response.

Link to my REW data now and prior to any treatment: https://we.tl/GqXEidQ7Ur

Really appreciate your opinions!! Thank you, sir!

Joe

Link to my REW data now and prior to any treatment:

Hey Joe, that's quite a decent improvement! Things have improved nicely.

Here's some "before" and "after" graphs.

RT-60 Before: RT-60 after: You can see how it smoothed out, especially in the low end.


Here's the frequency response, with both curves on the same graph, smoothed to 1/3 octave. Red is "before", blue is "after": You can see how the blue line is smoother and flatter overall, with none of the large big dips and peaks that you had originally.
That's the "before" waterfall plot, showing how the sound decays over time, and here's the "after" plot:

And finally the spectrogram plots, which show the same data but from a different point of view.

Before:
And after:
In all cases you can see the difference: smoother, flatter, cleaner.

Your hard work is paying off!

There's still room for improvement, of course, but that's a nice improvement that you already got.




I'm betting that you notice the difference! The bass should be tighter, and cleaner, and the overall sound should be smoother, with more detail and clarity... ?



- Stuart

Stuart,

Thank you for the interpretation! I look at the REW data and I don't see that! I guess one has to know how to pull out the data that is important.

Yes, it is much better!

So I'm sure you don't have much to do with Dirac room correction, because the rooms you design don't need it! - but I am using the trial Dirac software, which lets me demo music from my computer only - I would have to buy hardware to put it inline with my power amp and have it work for the studio console.
But it is VERY interesting.

I have generated filters from several measurements (mic positions), and they are quite different from each other. Without question, at least one of them cleans up the imaging immensely and tightens up the bass. It is possible to A-B with and without, as well as one filter against another. With NO correction, the sound is somewhat "bigger" and encompassing - but with the filter on, the imaging gets simpler. Everything moves to the front across the speakers, instead of surrounding me (wall reflections?). Interestingly, the mix from the big speakers with Dirac on sounds more like the mix from my nearfield Mackies, which are less than 3 feet away from my ears- I imagine that should mean that it is more accurate.

That said, the effect is more and less pronounced for different material. There is sometimes an impression that the music without Dirac "sounds better", but without question, everything is more shmeared across the soundstage than with Dirac on. I've been a musician all my life, but not a sound engineer, and I suspect with more experience, I would hear the Dirac as clearly better.

I can also move my head around the sweet spot more without obvious changes in sound with Dirac on.

Some of the other filters, made with the sampling mic in slightly different positions, are not as good.

So what is the take on Dirac room correction from the pro point of view? Is it something you even talk about? Do you think it can be valid?

Thank you!

Joe

I look at the REW data and I don't see that! I guess one has to know how to pull out the data that is important.

Yep! It takes a while to learn how to mentally turn the wiggly lines on the graphs into an understanding of what is actually happening in the room. But hopefully you can see the before and after differences.

Actually, I sometimes do use "room correction" (which in reality is a terrible name for it, since it does not correct the room at all! But that's a different rant... )

I do it using precision parametric equalizers, which you can buy for not-so-much money and are often better than the proprietary systems.

That said, it has to be done carefully, as you already noticed! You can't just do it on any old room, under any old circumstances... You can only use it for certain problems, in certain ways, and only when certain acoustic conditions are already met.

For example, you can never fill in phase cancellation null with "room correction" filters (EQ). The null is there because the wave is cancelling itself out at that position in the room, so if you try to boost that specific frequency, then all you do is make the null even bigger! It cancels itself out even better.

You can only use it where there are no issues with the phase: In fact, you can really only use it successfully for areas of the spectrum where the condition known as "minimum phase" exists. If you have rapid non-linear phase changes going on, then you can't use Eq to fix that: it will just make things worse.

Next, it cannot do much for problems in the time-domain (ringing modes, resonances, etc,), nor can it fix reflections.

And you also cannot use it to fix the whole room: the changes that you make only work for one point in the room, at the expense of all other points. If you improve the response for a specific frequency at the mix position, then the response for that frequency at all other locations in the room will be worse. Thus "room correction" is a terrible misnomer, since it cannot fix the room! It should be called "partial point correction", because it can only fix one point in the room, partially, at best. But "partial point correct" doesn't sound as good as "room correction", for marketing hype.... oh well....

So I use it when it is indicated. I use it when it can make a useful difference, and only when the room has already been well treated, acoustically, such that long resonances are already eliminated and most of the major problems have already been dealt with. If you use it under those conditions, and you know what you are doing, and you do it mildly, just where needed, then it is a great tool. I did use it on my current flagship studio (Studio three Productions. See here: http://www.johnlsayers.com/phpBB2/viewt ... =2&t=20471 ) right at the end of the process, for the final tweaks.

Oh, and I do it by hand, not automatically! I spend them time to analyze and deal with each individual issue on its own, as well as in conjunction with its neighbors, and I try to apply it minimally: as few filters as possible, at the lowest level possible, preferring cuts to boosts. However, it does take time to do that, and it's a service that I offer my customers. Some want it, some don't.

So that's my opinion on room correction: Done right, minimally, carefully. it's a good tool and can produce good results. It can be the final "cherry on the cake", that tweaks a "good" room into a "great" room.

I think your room is a candidate for that. From what I can see in your REW data, I think it could be improved a bit more, with careful tweaking. But I would not do it with a commercial "automated algorithm". As you already saw when you tried to do that with the free demo software, it doesn't work very well, and the results can be disappointing... But if you look at the graphs for Studio Three, you can see what can be accomplished when it is done properly by a real person, not a "one size fits all" algorithm...

- Stuart -

OK - REW Data recorded with and without Dirac room correction!

https://we.tl/fzI53hA8w0

Recorded at listening position, same level (about 74db spl) with Dirac OFF, Dirac Filter #1, and Dirac Filter #2. Number 2 is the one I think sounds good. Scary!

What I see:

Pretty quirky results. Filter #2 has better Impulse Response - less reflection in first 20ms - probably why imaging is better?

What the heck is with the 180 degree phase shift for both filters? I am definitely not hooked up out of phase! And I am not hearing that. If I reverse phase on one channel with the filters applied or off, it clearly goes from in phase to out of phase - easy to hear (center in the center vs. all over the place).

That improved RT60 you mentioned - Dirac seems to make it worse at the low end again... ??

Run like hell? Why does filter #2 sound so much cleaner than OFF??

Joe

Stuart, we crossed paths on my last email - I hadn't seen your last reply yet. Thank you again!

Interesting, Dirac claims that it can address phase issues, impulse response and reflections by manipulating phase, differently at different frequencies, to partially cancel some of the problem signals at the listening area (certainly not the whole room!). At least that is what they claim. They claim it is more than fine tuning of EQ. I'm wondering if its like The Emperor's Clothes - no one wants to admit the differences are so subtle that they can't hear them!

My ears are old and (slightly) shot - I do try to take care of them- and I don't have a lifetime of mixing behind me; I could never fine tune the room manually by ear. RTA maybe, but as you said, that ignores what, in my room for sure, are much bigger issues.

I'm really doing this more for enjoyment and learning. I paid a lot for these awesome vintage JBLs, and I want them to sound as good as I can get them to sound - but I'm not looking to make a living from it at this point.

So do you think Dirac's claims are BS?

lol

Joe

OK - REW Data recorded with and without Dirac room correction!

Interesting, isn't it? Not so much what you'd' expect...

Pretty quirky results.

Very true!

What the heck is with the 180 degree phase shift for both filters?

It's not what it looks like! For some strange reason, REW always shows the phase data "wrapped up" into a small vertical scale, so where it slides down past -360° off the bottom, they just wrap it back to +360° at the top. To me, that's a confusing way of showing it. To see it properly, y9u need to "unwrap" it: hit the "controls" button on the far right of the screen, and hit "unwrap", then go to the "limits" button and adjust the scale for the phase so you can see it more clearly. It will then look like this:
Top trace is SPL, bottom trace is unwrapped phase, with the scale adjusted.

So the phase is not actually flipping: it's just they way the graph is drawn by default in REW. Here's the one for "filter #2":
As you can see, the filter isn't exactly "fixing" the phase very much!

The phase will always shift: it has to. Laws of physics. And higher frequencies must shift more. Simple fact of life, due the distance between the speakers and your ears. Low frequencies see less phase shift because the waves are a lot longer, so there's less of a percentage change in the phase for that distance. Higher frequencies have shorter wavelengths, so for the same distance to your head, they must have shifted more. So you should always see an overall drop in phase as you go up the spectrum, and the higher you go, the faster it drops.

But it should be relatively smooth and linear.

Here's the graph for a room I just finished tuning (Studio Three): Same as for your case, top trace is frequency response, bottom trace is phase response. You can see that the phase is dropping in a very liner, very smooth manner, with no sudden up and down jerks, and it is doing so exactly at the correct rate. Also note that the smoothing on that graph is 1/48 octave, which is basically no smoothing at all!

Now, here's what it looked like for the original un-treated empty room, before I started doing my thing with the acoustic treatment, and the Eq tweaking: You can see what a huge difference it makes when both the treatment and the tuning are done correctly!

That improved RT60 you mentioned - Dirac seems to make it worse at the low end again... ??

There's always a trade-off, if the room has not been treated to the max acoustically. There's no free lunch. You can't improve all factors in all ways at the same time for the entire room. Improving one factor will always have consequences for another factor. That's why it is imperative to get the acoustic treatment right first, so that the digital tweaking has as little unwanted effect as possible on other things. But even so, there's still no free lunch: improving one thing at the listening position always costs you something, some place. And that "some place" might also be at the listening position, in some cases.

Why does filter #2 sound so much cleaner than OFF??

Mostly because it is smoothing out the low end a bit. Filter #1 is lousy: it messes up the mid range, as well as a few spots in the highs. But #2 is slightly better. However, I'm pretty sure I could tweak it much better than that, using my method...

Interesting, Dirac claims that it can address phase issues, impulse response and reflections by manipulating phase, differently at different frequencies,

Well, look at the graphs to see what it is actually doing, vs. those claims... You can draw your own conclusions...

to partially cancel some of the problem signals at the listening area (certainly not the whole room!)

They are honest about that part! Good. It's part of that "trade-off / no free lunch" I was talking about.

They claim it is more than fine tuning of EQ. I'm wondering if its like The Emperor's Clothes - no one wants to admit the differences are so subtle that they can't hear them!

You actually can hear phase changes, especially if there are phase differences between the channels.

They probably do manipulate phase, but I'm not so sure that's a good thing. Remember that I explained why phase has to decay at smooth rate, due to distance and the wavelengths? If you fiddle with the phase too much, you'll mess up that relationship and end up with the wrong phase shift for the wrong frequency... Also, you often find that the places where you need a phase change don't line up with the places where you need an amplitude adjustment, or that the Q is different, but in order to apply a phase change you must also have a filter of some type there! So how do you do that if the phase needs to be broad but the frequency needs to be narrow? You can't have different Q's for each parameter in the same filter...

I could never fine tune the room manually by ear.

That's too different things: I always tune manually, but never by ear. That Studio Three case above: I have never even been to that place! Never set foot inside. So I have never heard it in person. But I tuned it manually! From about 6,000 miles away, actually... Tuning by ear is one thing, and tuning manually is another. I tune using precision measurements and precision filters, carefully applied. I analyze the data to figure out what needs to be applied, but I don't try to do that by ear, because the human ear does not have the ability to do that. You can hear the difference with the adjustments, sure, but you cannot identify the exact frequency and Q anywhere hear accurately enough to apply filters by ear.

Sorry if I mislead you there, but what I to tune a room is not done by ear: it is done by data, but done manually. As opposed to letting a piece of software figure it out, but without really understanding what it is doing. When I look at the data in the graphs, I can interpret in in terms of what it means to the way things sound, the psycho-acoustics response of a real person sitting in that room. That's not how the software does it. It just crunches numbers, based on fixed equations and fixed algorithms. But people don't hear using "fixed equations and fixed algorithms"! We hear using our ears and brains. Since the software has no concept of what a room actually should sound like (just some numbers that supposedly describe that), it cannot do what a human can do.

I paid a lot for these awesome vintage JBLs, and I want them to sound as good as I can get them to sound -

Oh yes, we sure can do that! Not a problem. Manually, but not by ear...

So do you think Dirac's claims are BS?

Not really, no. Maybe just a little "over-exaggerated", but not BS. The product probably does work mostly as claimed, but as with any automated product, it has limitations that prevent it from doing things as well as a human can, because humans actually do have a sense of hearing, which the software does not. We can "feel" the sound, since music is as much art as it is math. Software can deal with the math, but it cannot deal with the art. Our ears don't turn sound waves into numbers that our brain processes: rather out ears turn sound waves into sensations and emotions and feelings. Machines can't deal with those concepts, since they are not conscious. They can tell you what the intensity, phase and delay are for each frequency, but they can't tell you if the bass is "tight", or the overall tone is "warm". Those are psycho-acoustic aspects of how we experience sound, and there's no way that you can put the concept of "warm" into numbers, or create an equation that describes "tight" bass, or "muddy" sound, or "tinny" treble, or "air".

So automated EQ systems can do a few things that help the mathematics of the room, but they can't do much to actually tune them to be "pleasant" and "sound clean and tight" to humans, because you can't put those concepts into equations.

It takes a human to do that... Fortunately for me, or I'd be out of a job!


- Stuart -

Wow, so interesting! You are very kind to take so much time sharing your expertise with me! It is much appreciated!

So can you make my room like Studio 3?? Didn't think so! lol

You have seen my REW data and photos of my room. How much more info is there to measure? More detailed measurements from far away have to involve extremely accurate mic placement, measurement of distances, and perfect description of all room surfaces, etc. Plus construction and installation of the treatments you prescribe has to be well implemented. How are you able to get reliable data from anyone other than yourself, without your being present?? And get results like you do? Hard to imagine! Impressive!!

I have had a few other local folks with good ears come by and listen, and all agree that the mixes are cleaner, with better imaging and soundstage, with my Dirac filter 2 than without it. I guess it will come to, as you say, a human decision - is the subjective listening improved enough to warrant the cost of investment in the gear, regardless of the graphs...

So can you make my room like Studio 3?? Didn't think so! lol

Probably not, but we cab still get it closer than where you are now...

How much more info is there to measure?

You'd be surprised... We've only taken measurements at a couple of locations, but there's lots more to learn about the room by taking more readings in other locations. We took over 500 readings in Studio Three.... Most of those where at the mix position or close by, but many of them were at other locations.

More detailed measurements from far away have to involve extremely accurate mic placement, measurement of distances, and perfect description of all room surfaces, etc.

Not really. Some measurements need to be accurate, sure, but for others it is often good enough to get a "ballpark" measurement. Especially for lower frequencies, where the wavelengths are long. If you are concerned about a wave that is 40 feet it long, you don't need to be too worried about getting the mic accurate to 64ths of an inch...

Plus construction and installation of the treatments you prescribe has to be well implemented.

In some cases, yes, that's true. But most of the treatment for small rooms is large in size, and doesn't need high precision.

How are you able to get reliable data from anyone other than yourself, without your being present??

When we started tuning Studio Three, Rod (the owner) didn't know much about mic placement, acoustic measurement software, how to run tests, etc. So I walked him through it by e-mail, and it only took a few test runs before he got the knack of it. The first few runs took him hours to set up, but by the end of the process he could run a dozen tests inside a few minutes. It's not hard to learn, as long as you can follow my instructions reasonably carefully. on the few occasions where he messed up, nad had the mic in the wrong location, or ran the test incorrectly, or had an equipment problem, it was fairly easy for me to spot that in the data and get him to correct the problem. Sometimes things pop up in the acoustic curves that don't make any sense, so you just sit down and try to figure out what went wrong, then fix it.

It really isn't as complex as you think.

is the subjective listening improved enough to warrant the cost of investment in the gear, regardless of the graphs...

Ahh, but that's the thing: It's not the subjective experience that matters; it's the objective experience. If you want better accuracy in your mixes, you need better accuracy in your room! The thing is, very very often, we can hear that things aren't 100% right, but we can't put our finger on exactly what it is. You might notice that your mixes just don't sound the same when you play them back in your car, house, church, office or wherever, but you can't actually figure what it is... even if someone put a gun to your head, you would still not be able to say what exactly it was that didn't sit right... and that's because the human brain just isn't capable of discerning the exact problem: even though it is aware of it, it just doesn't have the ability to tell you that the problem is that the left speaker is 2.7 dB louder between 590 Hz and 870 Hz, and there's a difference in phase above 2 kHz, plus there's a reflection that arrives at 7.3ms, and an SBIR dip at 73.2 Hz.... All of that can be clearly seen in the acoustic test data, and some of it can be fixed easily... but your brain is just not capable of hearing what the problem is, other than to say that it "doesn't sound right"...

The gear to get quite a noticeable improvement in your room would cost you under US$ 200.... Plus the time it takes to do the tuning. That could be as much or as little as you want, depending on how precise you want to get it.

Your room won't rise to the level of Studio Three, but it can still rise to another level that is rather better than where it is now.

- Stuart -

Ahh, but that's the thing: It's not the subjective experience that matters; it's the objective experience. If you want better accuracy in your mixes, you need better accuracy in your room! The thing is, very very often, we can hear that things aren't 100% right, but we can't put our finger on exactly what it is. You might notice that your mixes just don't sound the same when you play them back in your car, house, church, office or wherever, but you can't actually figure what it is... even if someone put a gun to your head, you would still not be able to say what exactly it was that didn't sit right... and that's because the human brain just isn't capable of discerning the exact problem: even though it is aware of it, it just doesn't have the ability to tell you that the problem is that the left speaker is 2.7 dB louder between 590 Hz and 870 Hz, and there's a difference in phase above 2 kHz, plus there's a reflection that arrives at 7.3ms, and an SBIR dip at 73.2 Hz.... All of that can be clearly seen in the acoustic test data, and some of it can be fixed easily... but your brain is just not capable of hearing what the problem is, other than to say that it "doesn't sound right"...

Ah! I totally get that! I have been spending hours listening to different Dirac filters. I hear the differences, and can describe some of them - but I also see the REW data associated with each, and I would be hard pressed to have any idea what the correlation is between what I am hearing and what I am seeing. And hearing which "sounds better" is also not obvious - but as you point out, is irrelevant. "Accurate", meaning mixes translate outside the room, is an objective thing.

OK, I'll bite. If I were to say, "OK Stuart, I would like you to help me optimize my room in as much detail as possible", what exactly would that involve?

I sent you a PM about a week ago. Not sure if you got that?

- Stuart -