My last post reviewed My A/V Rooms equipment and the cabling. Lets see how the room originally measured using some basic and easy to use Studio Six Digital iPhone and iPad applications.
During my early days of testing the rooms acoustics I used a suite of tools and a calibrated microphone system made by Studio Six Digital (S6D). However, while these applications are created to the highest of standards, I found the apps running on smart phones and tablets insufficient for my needs. This suite of acoustic applications has now been significantly expanded together with the supporting hardware. I find they provide excellent professional results but for use in My A/V Room, Room Equalization Wizard (REW) takes some beating. Also the users of acoustic forums where I would post my measurements and files for review, do not use these smart phone applications, most use REW created by John Mulcahy. I find REW to be an excellent acoustics measurement product, it is straightforward to use and is very well supported by its developer. It is provided as freeware from the HTS and AV Nirvana websites and I can highly recommend it. The next post will provide a more detailed REW measurement analysis for the room culminating in the final post where I add two more subs and finally achieve bass nirvana .
I have never spent large amounts of time creating frequency and time domain plots, as I have tended to just rely on my ears. This was about to change! That approach came about after spending years working with experienced studio owners and then designing and working in professional recording studio environments continually comparing what I heard in different rooms. Through experimentation I eventually started to think that I understood what I had to do to a room to make it sound “right”. I did say ‘started’!
Most of my early day measurements were centered around reverberation times (RT60) and obtaing a reasonably flat frequency response in the main listening area. Back then this current focus on having frequency responses flat all the way down to “DC” was not really that important. I worked in a music environment not film production and the Foley stage. I generally didn’t care what happened much below 30Hz provided that most the energy down there got absorbed. I well remember all the hoopla for the first control room in the UK that was designed to be flat down to 20Hz.
When designing a studio control room back then, the RT was determined, often by ear, and once acceptable, EQ was usually applied using 1/3rd octave White analog equalizers, and then again often just by ear. Modern day sophisticated room analysis that requires significant processing power was either too expensive, not available, or just considered not to be necessary, i.e. we can make it sound good. Remembering that when a mix left your studio it must not only sound good in that studio and the next one, but also on radio and at home.
So I still tend to rely upon my ears and memory (even at my age) and my classically trained wife and daughter musicians to determine if something sounds right. On occasions I use a Microtech Gefell M930 (read the review here) to mic up our grand piano (mono) and listen to it hard wired to My A/V Room system.
I like lots of really clean bass so obtaining a good LF response is always of paramount importance to me. Adjusting the mid band and higher frequencies is generally easier to accomplish than obtaining a good bass response, so that is what I focused on first.
The labor of love equalizing and measuring my rooms acoustic performance was accomplished over several small steps and a couple of really big ones:
- STEP 1. AVP Audyssey XT EQ – the original installation only using the AVP built-in XT Audyssey EQ, with no real time measurements
- STEP 2. AVP Audyssey XT EQ using the Studio Six Digital software and hardware for acoustic measurements
- STEP 3. AVP Audyssey XT EQ with the addition of a Velodyne SMS-1 and upgrade to Room Equalization Wizard (REW)
- STEP 4. AVP Audyssey XT EQ with the addition of an Audyssey XT32 Sub EQ and Audyssey Pro (Audyssey Pro now allowed me to go back and “measure” steps 1, 2 and 3)
- STEP 5. AVP Audyssey XT32 EQ – the upgrade of the AVP to XT32 together with the Audyssey XT32 Sub EQ and Audyssey Pro
- STEP 6. AVP Audyssey XT32 EQ with the addition of two subs, making four, and reverting back to the SMS-1
From the very outset the bass sounded poor after Audyssey XT EQ in my AVP had been run, no matter where I placed the microphone and how many positions I used. At this point I could not create any graphs so I have no meaningful record of what Audyssey was actually doing to the room except what the AVP showed for the Audyssey simulated EQ graphs. These were of little use as they just showed the amount of EQ applied per channel and DID NOT even show the sub EQ response, the most important.
The following graphs were taken at a much later date, after I had purchased the Audyssey Pro kit. They are included here to show what Audyssey XT was supposed to be doing to the system. As can be seen they don’t look terrible but all the satellite speakers bass equalization looks poor and the overall bass sounded rather thick and muffled. Audyssey Pro also insisted on either setting the 1038’s to large or crossing over at 40Hz. This made the bass sound even worse than crossing over at 80 Hz for all satellite speakers.
Unfortunately these Audyssey graphs do not tell the whole truth as the AFTER responses are NOT actual measurements. They are only predictions by Audyssey as to what the rooms response SHOULD look like after it has been measured and the resulting EQ applied. To complicate things further, Audyssey in an attempt to broaden the listening area, averages many readings in order to create its final equalization. Ultimately this makes its resulting EQ wrong for virtually all listening positions, while providing the best overall performance for most listening positions. I wanted MY seat, the MLP money seat, to be optimized and Audyssey is not designed to do this.
Much later in this equalization marathon I purchased Room Equalization Wizard (REW). The following graph shows the actual measurements of the rooms acoustic performance before and after the application of Audyssey XT EQ using Audyssey Pro. I have included just one of the many graphs that I created in a vain attempt to make the room sound ‘better.
Even though it can clearly be seen that Audyssey was providing a significant improvement to the rooms response, to be honest the improvements were not as good (flat) as I had expected, and with the sub par bass performance, I was a little more than very disappointed. I assumed that there was a problem with the rooms acoustic design so I realized that I needed some tools to measure the rooms performance in real time. Enter Studio Six Digital.
Step 2 – Studio Six Digital Acoustic Measurements
I had by now obtained the Studio Six Digital (S6D) acoustics measurement package that I was testing for its designer, Andrew Smith. I hoped that it might provide some insight into what was wrong with the bass and the room generally, but I soon realized that the then available range of measurements would not be able to help me.
Please Note: The S6D suite of acoustics software and hardware has been considerably enhanced and expanded since I first used it. It is an excellent suite of very cost effective products for the professional and portable acoustics markets, and can be highly recommended. However, as will be shown, Room Equalization Wizard (REW) is in my opinion, a more suitable suite of software applications that are very simple to use, free, accurate and flexible, and has outstanding support from its creator John Mulcahy. More on this product later. More importantly REW is supported by several large Forums such as AVS, HTS and AV Nirvana that provide access to a vast number of experienced users, and the author, who are ready and willing to help all new comers analyze their files and help with using the application.
I used the Impulse Response (IR) application together with the calibrated iAudioInterface and associated mic. for the room analysis. This application supports various methods of analysis from impulse to swept sine wave. Unfortunately, its response was only 63Hz to 10Khz so it did not provide any further insight into what was going on in the lower octaves, still having to rely on my ears.
The first and simplest test performed was an impulse test created by bursting a large balloon at the front center speaker position.
The decay characteristic from this test indicates that the room has an average RT a little less than 0.2 secs. However at about 8mS there is a clear reflection. This is actually from the rear diffuser. There are no obvious secondary decays.
The decay markers provide clues to the perceived RT60 time and the EDT and T30 all confirm a close to designed RT time of 0.2 secs. The graph below -50dB indicates a potential for secondary decay. However, this is not the case, it is the Apple iTouch noise floor impacting the decay rate.
This shows the decay time per third octave within the first 10mS and correlates with perceived RT60 time. Ideally all third octaves should be equal to 0.2 secs but show a faster decay of approximately 0.17secs. This is upheld by close examination of the initial Schroeder decay. I had not yet determined why there was a significant increase in the RT times around 200Hz and a spike at 2000Hz. However, these only occured during the first 10mS and did not appear to effect the overall RT60 decay time. NOTE: This issue was later discovered to be some woodwork and rack hardware that were ringing just for this initial period.
This provides a closer look at the reflections that occur at the MLP within the first 200 milliseconds. It is useful to look for echoes and repeating decay patterns or non-linear decays. None were seen.
This provides a closer insight into the actual decay times per one third octave, ideally these should all be the same and would be expected to rise at bass frequencies. Within the accuracy of this hardware, they all approximate to the design goal of 0.2 secs decay, starting to gradually rise as expected below 63Hz.
The amount of broad band energy released from a ballon can’t compete with a proper sine wave sweep, especially at low frequencies, so the results obtained above are not as accurate as using a sine wave sweep. All these tests were, therefore, repeated using a 3 second sine wave sweep fed to the AVP front L/R 1038’s and both subs. All levels, distances and EQ having previously been set using Audyssey XT .
The results obtained from this sine wave sweep were very similar to those seen by the simple ballon burst impulse test. The differences mostly resulting from the increased wideband energy available from the sine sweep at lower frequencies and showed a more rapidly rising decay time below 100Hz than was displayed by the ballon burst.
Almost everything I heard and measured seemed to correlate with what the Studio Six software showed me and what I expected from the speakers and the rooms design and treatment (or lack thereof). Furthermore, I am not a “fiddler’ – if it sounds good (but it didn’t) I have no desire to ferret out a problem that I may not be able to fix, this would then drive me nuts.
It eventually turned out that what I was hearing lay in what the measurements did not show me, and what I initially accepted as ‘normal’ for the room and its treatment, in particular its less than stellar bass response!
Step 3 – The addition of the Velodyne SMS-1 Sub EQ
After reading that Audyssey XT did not provide very good bass EQ due to the limited number of frequency points that it could EQ, I opted to add an external equalizer in order to pre EQ the subs and make Audyssey’s job a little easier. I already owned a Velodyne SMS-1 from an earlier HT system. So I decided to use it to pre-EQ the subs before running Audyssey to see if that helped. It would also allow me to check the rooms LF response as it measured and displayed the real-time frequency responses from 20Hz to 200 Hz.
I only have a handful of measurements that I kept from this hardware. Shown below, are the responses of the untreated room and SMS-1 equalized subs.
Based upon the selected MLP you can see the frequency of peaks and troughs correlate well with the various axial room mode shown in the earlier design post. It also appears that there is little LF absorption from the drywall and it nows seems to be behaving more like a rigid wall. I could never detect any resonances or flexing except for the ceiling as mentioned in an earlier post.
Once I had the LF response reasonably flat using the SMS-1, I then re-calibrated the room using the AVP Audyssey XT EQ with Pro. This resulted in the following final equalized sub response.
At least Audyssey seemed to iron out most of the remaining LF peaks and troughs and it corresponds well with the AFTER Audyssey EQ response graphs shown below.
All the resulting satellite speaker responses using Audyssey XT Pro EQ are the same as those shown above in Step 1.
This is what the SMS-1 EQ looked like when I finally got to checking it with REW.
This final SMS-1 EQ (no AVP XT EQ) corresponded well with both the SMS real time graph and the before sub EQ made by the Audyssey software, both shown above.
The bass now seemed considerably better but I was still not happy with it compared to what I remembered from my studio days.
So even after all this, my next leap of faith was to purchase a two channel Audyssey XT32 Sub EQ HT together with Audyssey Pro EQ HT as it is required to run the Sub EQ hardware. I also purchased an Audyssey license for my AVP which allowed me to control it through the Pro application during the EQ process.
Step 4 – The Addition of Audyssey SUB EQ HT
Believing that the SMS-1 was not providing the best EQ possible for the subs I purchased a two channel Audyssey Sub EQ HT. The first decision was how I was going to use this two channel device. To cut a really long story short, and after much testing, I concluded that it provided the flattest and most pleasing bass response when it was used to combine both subs into one. It still measured each sub separately in order to set distance and level but combined them into one when it equalized them. This is the preferred and recommended approach.
After using the Audyssey Sub EQ to pre-equalize my subs the predicted response is shown below.
I then used Audyssey XT in the AVP together with Pro to complete an all channel EQ. This seemed to result in a tighter bass response but it was still not to my liking.
The following graphs are the ‘estimated’ system responses provide by Audyssey before and after equalizing the entire system, including the Audyssey Sub EQ HT used for the subs pre-eq run.
The mid band dip and HF rolloff were deliberately selected by me during the Audyssey Pro EQ process. Both of these are often required to help reduce listener fatigue from an over bright mid range and top end.
The most obvious points to note are that:
- The subs response ‘looks’ REALLY flat, all the way down to 20Hz!! – Unfortunately it didn’t sound like it!!
- All pairs of speakers have acoustically identical responses before and after EQ – within the limits of these predictions.
- The final Audyssey response of all satellite speakers is virtually identical above the selected LF crossover.
- The frequencies of the peaks and troughs of the non-equalized responses correspond to the earlier predicted L, W, H axial mode values.
- The corresponding LF peaks and trough level changes are as would be expected from the placement of the speaker relative to the walls and ceiling.
- Relative to the MLP the values for distance are exact for all pairs of satellites and their level match is good. Level match could be made exact by adjusting the gain controls on the rear of each satellite but +- 0.5dB is more than close enough.
- The equalization for frequencies below 100Hz didn’t improve.
So what did the REW sub measurements look like and how did they compare to Audyssey’s estimates? Again, in order to shorten weeks of measurements I will just include the final REW room sub response that I achieved during this process.
With an LF response this flat (no smoothing) you would have thought that the bass would have sounded really good. Sorry but that was NOT the case, it was still slightly muddy and too full.
So instead of reviewing all the data that REW was creating for me that would have led me to see what the problem was, Denon was offering an upgrade to the AVP that also included XT32 with Sub HT EQ. So guess what? Yes, I upgraded the AVP, again hoping to cure my bass problems. Read on!
Step 5 – AVP Upgrade to Audyssey XT32
Even though the mid and higher frequencies, stereo imaging and depth of image were all quite good, I was still not thrilled with the bass response. So again after reading that the AVP could now be upgraded to XT32 with many of the features that the Audyssey Sub EQ HT had, yes you guessed it, I upgraded the AVP. That also meant adding a pair of height speakers as this upgrade included additional height processing. Equalization runs had now taken over my life and were taking days to complete.
In theory the upgraded AVP could now do what the external Audyssey Sub EQ HT could. However, NOT SO. It turned out that it CANNOT EQ the sub channels as one and tries to EQ each of the three available sub outputs individually. I had already concluded this was not the best technique and having the AVP driving both subs as just one removed some of the added benefits of the external Audyssey Sub EQ HT.
The two sub graphs below are included just to show the results of the AVP Audyssey XT32 Sub EQ HT running as two channels together with the external Sub EQ HT running as two independent channels, each channel driving just one sub. I can promise you it never sounded as good as using the external Audyssey Sub EQ HT as descibed above.
So, in the interest of brevity the following Audyssey predicated responses were how the system was left after the XT32 upgrade using the External Audyssey Sub EQ HT to pre EQ my subs as one pair.
On paper it looks good, yes?
- The second pair of surrounds (B) are now missing
- A pair of height speakers had been added
- The sub response still looks about the same as before the upgrade. It even seems to kick up a little at 20HZ!!
- All speaker pairs have virtually identical responses
- The overall response of every speaker is virtually perfect. (Well, with 1/3rd octave smoothing they should be!!)
- Frequencies below 100Hz have now been equalized much better due to the significant increase in the number of equalization points that are available from the XT32 upgrade.
When comparing the XT equalization to the XT32 equalization it is clear that there have been improvements in the process, especially below 100Hz. The overall frequency responses are smoother and it resulted in an overall cleaner soundstage and slightly tighter and less boomy bass.
So what did the final REW measurements look like for the front channels?
The drop in level at 1KHz was introduced by me using Pro, as it seemed to compensate for the speakers sounding a little too forward. The HF roll-off above 10KHz is also deliberately introduced using Pro in order to tame the top end a little. Ignoring the frequency response changes made using Pro the overall frequency response was within + – 3.5dB. This is often considered quite acceptable, but in my opinion there was still plenty of room for improvement, especially in the bass region, despite it looking nice and flat to 15Hz.
As with the S6D electronics, the RED residual electronics noise floor limits the measurement of the NC rating. The PURPLE increase in the noise floor below 40Hz, with the HVAC active, is to be expected but it is inaudible, see the NC graph below. The room doesn’t exceed NC26 due to the two peaks at approximately 60Hz and 120Hz. I suspect that they are both supply frequency related to the fan speed. See below for the S6D NC graph.
Well, for step 6 results, including the addition of the two extra subs, you will have to wait for the coming posts.
This is where I discover the benefits of working with Room Acoustic Wizard (REW). The challenge was to be interpreting all the acoustical data the software generates.
After completing step 5 I was reasonably happy with the acoustic performance of the room. Imaging and depth perspective were very good. The bass integration and the LF response, for either music or film, was acceptable, with OK extension and ‘slam’ and it was quite tight and reasonably clean. However, the room still fell short of really deep bass AND it still sounded a little thick. Almost as if the bass was decaying too slowly. It was, see coming posts.
The room acoustic isolation from the main house is outstanding, as measured earlier, and could not have been improved further without significant additional construction expense.
The graph below shows a background noise value of NC26 with or without the HVAC on. This is because the background noise floor of the room far exceeds what the Studio Six Application can measure due to the mic, preamp and iPhones electronics noise, so the final room NC rating is not known. However, the projector alone will prevent it going below NC25 and would need to be either re-located outside the room, put in a sound absorbing enclosure or replaced with something quieter. None of which are likely to happen in the near future.
Clearly noise below the 63Hz octave is below the threshold of hearing. I suspect that the rooms actual background noise is well below NC20 with the projector and HVAC off when I am just listening to music. I can honestly say the room is really silent even with the HVAC on!
The next post. Room review and Room Equalization Wizard (REW) results in more detail.