Datasat RS20i BETA test and comparison with ADA Rhapsody/Trinnov TEQ-12!!!! - Page 22

Well I did say



That further clarifies what "perceptually flat" means, but seems to leave open the possibility that the result is objectively inaccurate, i.e., if some live sound were produced in a given room, then reproduced with speakers EQ'd to perceptually flat, it might sound better but different in spectral balance than the original.

There is no "original." You never, ever know how it all sounded to the talent that approved it. They listened to it in a different room with different speakers. So nothing here is about accuracy in that regard.

All you have left then is what is preferred. And we get to know that with listening tests such as being discussed here. From Dr. Toole's Sound Reproduction Book:

"As consumers of these programs, we cannot know what was intended for the sound of any of these programs. We were not there when they were created. We may have been at performances by similar, or even the same, musicians, but they were likely to have been in different venues and possibly amplified. None of us ever placed our ears where the microphones were located to capture the sounds, nor would we want to; we were almost certainly at a distance, in an audience. A simple reproduction of the microphone signals cannot duplicate the experience.

Descriptors like pleasantness and preference must therefore be considered as ranking in importance with accuracy and fidelity. This may seem like a dangerous path to take, risking the corruption of all that is revered in the purity of an original live performance. Fortunately, it turns out that when given the opportunity to judge without bias, human listeners are excellent detectors of artifacts and distortions; they are remarkably trustworthy guardians of what is good. Having only a vague concept of what might be correct, listeners recognize what is wrong. An absence of problems becomes a measure of excellence. By the end of this book, we will see that technical excellence turns out to be a high correlate of both perceived accuracy and emotional gratification, and most of us can recognize it when we hear it."


This is why I think it is so bad when we don't have proper listening test data and instead folks talk about filter types and such. It is important to comply what we collectively without bias indicate as good sound.

It's definitely a good point but in my experience after spending $$$$ on a room consultant, room build and treatment etc... it wasn't nirvana. I'm sure part of it was related to my fixed ceiling height. I spent some bucks getting Adam Pelz up here to integrate my front 2 F113s with my main speakers and it was a big improvement. The Trinnov just takes it to the next (and final for this room!) level. I think the best result will be to have a "decent" acoustic environment and good gear and the Trinnov or Dirac or whatever will be the icing on the cake.

Speaking of the Anthem ARC, I found that if you let it correct all the way to 20khz that the sound was too brash. I limited correction to 1khz or so and it is a remarkable improvement for movies without costing very much money (for someone that already had the Anthem product).

Thanks , I was just pointing out the similarities with the ease and the scope that the Dirac live has compared to the Trinnov. Cheers

There is in the simple scenario I posited.

I do agree with what Toole said.

I suppose that would be one version of "accuracy", not to mention provide for an interesting experiment in equalization: have a vocalist and loudspeaker next to each other in the same room, record the vocalist and try to get the recording to match when it is played back through the loudspeaker. There would be some confounding variables, like the vocalist and loudspeaker being in different locations, lack of consistent repeatability from the vocalist, time gap between hearing the live voice vs the recording, is the EQ compensating for the room or the miking technique, etc.

As far as commercially available music recordings, it is as Amir said: there is no way to judge accuracy since we don't know how it sounded originally. In which case, the best you can hope to do is get results that are pleasing to you. Research shows that most of us humans have similar preferences in this area, which at the very least gives use a useful starting point, after which we can tweak to individual taste (pass the salt please). Since these preferences aren't a secret, it is not surprising to find that some room correction systems have a default target curve that sounds pleasing right out of the box. Which also makes good business sense.

BTW, frequent poster Wayne Pflughaupt has a nice tutorial on how to come up perceptually flat target curve. Again, good starting point (irrespective of the room correction system you plan to use), which you can then salt-n-pepper to taste: House Curve: What it is, why you need it, how to do it.

Thanks for the link, it's a good read.

Dan, I have a few technical questions:

1. Where do you put in the speaker settings? In Rhapsody or TEQ Trinnov?

2. I understand that during the test you did not connect Rhapsody with TEQ Trinnov via ADA bus? Do they need to be connected via ADA bus during normal operation?

3. How do you regulate the volume? I switched the TEQ on Rhapsody and the sound was too loud and I could not turn it down using the master volume. I was told that this was because I did not have TEQ Trinnov but I understand the TEQ Trinnov doesn't have a volume knob.
4. How did you connect the sub from Rhapsody to TEQ Trinnov (to one of the channels)?

5 My system is 5.1 but I would like to add a second sub. Can I connect the one of the surrounds channels of Rhapsody to TEQ Trinnov or better use Y split from Rhapsody to feed two TEQ Trinnov channels?What about to use the free surround channels on Rhapsody to 2 channels of TEQ Trinnov and to ignore the sub on Rhapsody?

Thanks for the help.

PS If someone has experience with TEQ Trinnov, please join in. Any help will be appreciated.

While you can run the TEQ as a DSP and adjusting the volume on your Rhapsody or other processor with good results, it's not the optimum way to do it.

Ideally you would use the ADA bus. We designed it so that when the bus use, the Rhapsody is always sending full audio levels to the TEQ, while also sending 2 way control of Volume to the TEQ. The TEQ then adjusts volume accordingly and confirms this back to the Rhapsody. That is why when you selected TEQ the Rhapsody was LOUD.

If you forgo the ADA bus, you could get the same results manually: set the Rhapsody or other processor to optimum level and use the TEQ to change Volume level. The TEQ DOES have a Volume Knob and, as well can be controlled remotely via VNC on an iPad or other device.

Both methods will give you optimum listening results.

Bump

Dan & Curt, thank you for the help. I am expecting my unit (TEQ 8) next week and I am trying to get everything ready for the arrival of the new baby...
Appreciated.

Curt, if you control volume on the ADA, would this not be done in the analog domain, and (at least theoretically) be superior than doing it in the digital domain on the TEQ? If so, bypassing the ADA DSP, and just doing decoding, D/A conversion and volume control, and then sending that signal to the TEQ should give the best resutls.

Good point Erik, there is an analog volume on the ADA, but it still turns out better to do it digitally on the Trinnov. If you were to use the analog volume control upstream of the Trinnov, you loose both bits and s/n as you turn it down. If you leave it with optimal levels and then turn down the Trinnov, you only loose bits at the rate of 1 bit per 6 dB.

Example: Lets say one lowers the volume from max 24dB. If you do it up stream of the Trinnov/TEQ, you loose 4 bits and give up 24dB of S/N. If you do it in the Trinnov, S/N is not impacted, and end up with 20 bit resolution.

His "house curve" shows a 16db drop from 20hz to 20khz?? I tried a "house curve" with about a 10db drop (with my TacT) and it sounded awful -- DEAD. Bass heavy. I can not imagine what this would sound like.

Peter and I have been discussing his project off-line. The answer to the question is not obvious, as there are several aspects that are operating in concert when we consider "height channels." Therefore I'm going to provide some insight and some AVS quotes from a Trinnov user who has done all three height possibilities.

At Trinnov we are not after fabricating something that is not in the mix. Rather, we are after spatial accuracy inherent in every mix. We can speculate that the use of height and width in the fixed speaker positions products are there to give a further sense of envelopment, to make the speakers disappear, and render more life like sound. This is not an constraint with Trinnov Remapping, so the approach to height is different: height speakers in Trinnov contribute to building the original acoustic field. Example (thanks Ryan1 and sdurani, R972 Trinnov User Notes thread)....



This makes sense, but it does go beyond this. A characteristic of Trinnov Remapping is to give a sense of width and height space that extends to the borders of the speakers, while the image remains in a focused. It is one of the reasons I've encouraged the use of height in the R972 Trinnov User Guide. I've not used the other systems, so can't comment on the relative merits.


Perhaps I'll be able to go into more details later. For now, I hope the experiences of Trinnov users give you some insight to the possibilities.

Cheers,

So, rather than using DSX or similar, I add extra height and/or side speakers to the mix, and Trinnov remaps them to the original soundfield?

If I used an ADA Reference I could add a couple of heights, a couple of sides, even a couple of "downs" (inverse heights") and Trinnov would remap those to create what should be an enveloping, seamless, "7.x soundfield" created by the 7.x speakers plus the six additional speakers (two additional heights, two additional sides, and two additional "inverse heights")?

As opposed to DSX or similar, which is "synthesizing heights" and/or "synthesizing wides" from the mix?

Is that approximately correct?

I see. Does this mean that if I play a 44/16 2 channel digital source at about -40db on the Trinnov, I loose between 6 - 7 bits, leaving me with just between 9 -10 bits resolution?

Bruce- Yes, you are on the right track... and could experiment with the 7.1 Trinnov unit you now have- borrow back channels to try in the front. Look in the R972 Trinnov User Guide for the "W" configuration using 5 fronts. You'll have fun with that one. Cheers-

I on the other hand find a 16db drop good. I guess speakers could add to the equation, but I think the fact that I am pretty sensitive to highs is also a big factor.

Ah, more like 18 bits left. Trinnov output is 24 bits. When attenuating, your lowering all bits, so it doesn't come off the top, it comes off the bottom. Lowering 6 bits leaves 18 bits.
What do you loose: we start with 16. We process EQ at 64 bits. This gets truncated to 24bits EQ resolution for full output. In this example, we now throw away the 6 bottom bits.

If those of you following along here get lost here because we started with 16 bits, keep in mind we change the original program material when we EQ/remap, and our accuracy in doing so is bit depth. More bit depth is better, because the intermediate math can otherwise eat up your resolution. So to keep resolution, we typically see anywhere from 24 to 64 bits, either fixed or floating point. Trinnov uses 64 bits floating pt. We want to retain resolution on output, as the signal has changed with EQ/remapping, which leads us to Erik's interest. On output, theoretically, we would want to keep all 24 bits, even for 16 bit source material. In a no costs barred world, one might add some analog attenuators after the converters. Some do in their products and they can be an add on as well.

What if you play at full level, but program material that has quiet passages at -40dB? We'd want more then the 9-10 bits of resolution in the program material. That's what 24 bit HD content gives us.

For 16 bit, could you hear the difference if you went to a analog attenuator? Probably not, but Erik might hear it because he has a stunning system. Cheers,

So just to make absolutely sure I understand, when I feed the Trinnov 44/16 from my USB converter, and the Trinnov is in slave mode using the converter as its master clock, the Trinnov still stretched up the word lenght to 24 bit for internal processing?

Irrespectively, I am now sufficiently intrigued to put my very transparent MCH analog preamp (Six Shooter) behind the digital chain, play the Trinnov at 0Db attenuation and see how it compares with digital domain volume control.

I currently use my preamp before the Trinnov for MCH sources only, attenuating in the analog domain before A/D conversion. This is analogous to applying analog domain volume control in the ADA before the TEQ, so if I understand you correctly, feeding the Trinnov full volume analog singal, playing the Trinnov at 0Db and applying analog domain volume control after the final D/A conversion should be better as well.

You'll have 64 bit floating point processing (program material becomes altered by compensation), with 24 bit output.


Affirmative.

It depends on how far your seating distance is from the speakers and what kind of power response they have.Typically, the closer you sit to the speakers the less of a "curve/tilt" that you may need.

I sit pretty much nearfield to my monitors and I use a very shallow "tilt" and it sounds much more natural to me than "flat". The 10dB "tilt" sounds dull and way bass heavy to me.

Perceptually flat (at least to the listener who dialed in the house curve).

OK. Before I start messing with cabling and configuration one final question. If I start out with 44/16 material, and this gets stretched up to 44/24 by the Trinnov for processing, and then cut back to something in the 44/16 - 44/20 range for volume control, am I right concluding I lose absolutely no information? After all, the Trinnov cannot create more information than the original 44/16 out of thin air, so as long as I have at least 44/16 I lose no resolution? This is important, because if that is the case, outputting full volume 44/24, and adding an analog volume control at the end of the chain may not buy me anything. In fact, running through an extra cable and circuit will probably just induce signal degration.

After compensation, you have more then the original 44/16, because the original has been altered. As you have a Trinnov, you could actually look to see how much. As an example, you have level changes due to EQ and speaker differences that eat up dynamic range. If these amounted to 4 bits, well, then you may need the 16+4=20 bits to keep it whole. If you want more detail regarding your specific situation, then we'll need to chat.

Cheers,

I think I understand. If I need to jack up my mains by say 6db to get the same level as my subs, and there is a max of +12db boost in eq, I need 1+2 = 3 bits, so if I drop below 16+3 = 19 bits I lose information. If I chop of say 6 bits for 36db attenuation, I am at 24 - 6= 18, and lose resolution. This is actually a compelling reason to set volume levels of the subs so that very little of no bits are consumed by equalization of subs and mains. Since I do in fact attenuate by over 42 bits late a night I may be losing resolution, and it may be worthwhile trying out the MCH analog preamp for attenutation instead.

It's just 16+2=18 bits to cover a 12 dB boost.


The question of losing resolution also depends on the noise floor in the source and whether the lower level bits are audible in the playback system. It might not be as bad as you think.


Maybe you mean 42 dB?? If 42 bits you are losing a lot more then resolution. The source signal is losing resolution when the volume is reduced (this applies to analog or digital volume controls), but it does not matter since the noise floor (or the quantization floor) of the playback system is not increasing.

If you reduced the MV by 40 dB then added 40 dB in the power amp, the signal would be full volume and have much worse noise/distortion. No one does that, though.

That's what most want, rolled off highs and jacked up bass. I prefer closer to flat response. Not sure I would like that either.