1. This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
    By continuing to use this site, you are consenting to our use of cookies.

    Dismiss Notice

Smyth Research Realiser A16

Discussion in 'High-end Audio Forum' started by jgazal, May 7, 2016.
483 484 485 486 487 488 489 490 491 492
494 495 496 497 498 499 500 501 502 503
  1. Rene Lou
    About optimal A16 microphone placement in your ear canal and the A8 & A16 microphone types:

    The A8 microphone is placed central at the top of a plastic cylinder. The plastic cylinder has to be inserted into your ear canal and the tiny microphone opening (small hole) was facing away from you. As a result, rotating the microphone cylinder in your ear canal had no effect in terms of PRIR capturing performance. Picture from the A8 manual:

    The new A16 microphones are different and have a tiny microphone opening looking towards the attached cable:
    IMG_952_2_sm.jpg IMG_9430__sm.jpg

    In the A16 manual there is a picture about a "correct" A16 microphone placement:
    This picture looks like if you would not have to put the microphones fully in your ear canal (which is not correct and will result in sub-optimal A16 performance - at least for my ears in terms of sound localization capabilities).

    I made several tests with A16 microphone placements as following:

    Initially, I captured my large stereo speakers with microphones fully inserted into the ear canal and with a similar microphone orientation as on the picture in the A16 v0.9 manual (microphone cable leaves downwards and the tiny microphone opening faces the floor):
    The resulting PRIR with the corresponding autoEQ HPEQ sounded much to bright. After I made a manSPK HPEQ, the brightness was gone and the sound was more natural and in a way similar to my stereo speakers, but still not an exact copy of the speakers in terms of localization performance (eg: when playing some music, instruments were not positioned in the room exactly as when my speakers were playing - the speakers did span a deeper 3D-like room in front of me).

    So I tried a different microphone placement, more similar as shown on the picture in the current v0.9 manual:
    The resulting PRIR with the corresponding autoEQ HPEQ sounded still to bright. After I made a manSPK HPEQ, the brightness was gone again and the sound was more natural and in a way similar to my stereo speakers, but this time the localization performance was even worse: listening to music with the A16 sounded still out of my head, but this time much closer to me compared to when the speakers were playing the same music. This was not satisfying since I knew what the A8 was capable of and what I was looking for with the A16 as well.

    So I tried the following: inserting the microphones fully into the ear canal again, but this time rotating them by approx. 90 degrees in a way that the tiny microphone opening is looking towards the rear:
    Success ! The result is absolutely convincing ! The PRIR with the autoEQ HPEQ was still to bright initially as well, but after doing the manSKP HPEQ routine the A16 is now a perfect copy of my speakers. Eg: When a saxophone in a given music is located at approx. 1m behind and 0.5m towards the right of the front left speaker, the result with the A16 is exactly the same - it is an exact copy!

    At least for my ears, this finally worked perfectly. So I did a full 9.1.6ch PRIR with my stereo speakers (as I have explained in a posting earlier in this thread) and did the autoEQ HPEQ and the manSPK HPEQ routines. The sound imaging of the virtual L+R speakers compared with my real speakers is identical. I used several well known music tracks to compare the A16 performance with the real speakers and have now a true copy of my speakers in the HD800 !

    Maybe you want to give it a try as well: to rotate the microphones and do some tests with your own ears.

    Some remarks:

    I tested if I can hear any differences in performance with the Sweep types "12 sec" (single frequency sweeps per channel) and "12sec overlap" (frequency sweeps overlapping, means the right channel sweep starts already when the left channel sweep is still active in higher frequency regions). Result: No difference at all and the 12 sec overlap sweeps are much quicker. The resulting PRIR files are identical to listen to. So the 12 sec overlap sweeps can be used without any negative impact - that is an additional improvement vs. the A8.

    Making the manSPK HPEQ with the HD800: Be aware that even a tiny different placement of the headphone on your head can have a big impact in terms of needed correction - especially for higher frequencies of the test. So be careful not to correct to much in the higher frequency bands just because of non-optimal headphone placement.

    And finally: I think the autoEQ routine could be improved so that the mid/high frequencies do not need such a big correction with the manSPK afterwards. Just doing autoEQ - at least for my ears - sounds always to bright.

    The A16 is an incredible device !
    Last edited: Nov 30, 2019
  2. Rene Lou
    How to optimize head rotation accuracy with the A16:

    I made following experience with the head tacker and PRIR measurements:

    When I started with the A16 I did several quick PRIR tests with just looking approx. 30 degrees to the left and looking approx. 30 degrees to the right to get more familiar with the A16. Sometimes I used the HT assist switched on (Apps --> PRIR measurement (SPK) --> Config look angles --> HT assist). HT assist will support you in looking into the right direction with guiding tones, but does only work when you are looking in the direction the A16 assumes. When you are multiplying speakers (eg: make a 9.1.6ch system with a pair of stereo speakers), HT assist will not work for all channels. When the angles are not perfectly correct during the PRIR measurement, and when you are listening to some music later on and turning your head, the effect of head tracking will not be optimal. Eg: when turning my head to the left, the right virtual speaker was getting louder then it does in real life - the same was true with the left speaker when turning my head to the right - so this was unnatural and not convincing.

    On the PRIR measurement screen of the A16 on the top right corner, the actual head tracking degrees are shown and updated immediately when you are turning your head even slightly by using the provided head-band. Unfortunately the web interface is not ready yet, and when turning your head during PRIR measurements you will not be able to see those tiny numbers - even when the A16 is standing just in front of you.

    I made a comment some time ago about using a video call to get the A16 display closer to you. In the picture below you see my current setup when doing PRIR measurements: Using an iPad camera in front of the A16 and pointing it towards the A16. FaceTime (video call) is being used to call my iPhone and then I simply hold the iPhone in front of me when turning my head. With this, I can rotate my head to any angle I want with a +/- 1 degree accuracy:

    This works quite well. Maybe you want to try this as well until the web interface will be ready. For Android devices there will be a similar solution available I assume.
    Last edited: Nov 24, 2019
    arnaud and Richter Di like this.
  3. Rene Lou
    two times yes
  4. Richter Di
    When we - @sander99 and me - did the measurements of the Focal Aria 906 in my Pilates studio we optimized the room as far as we could
    Carpets on the whole floor, acoustic elements in front of the windows etc., acoustic elements on the sides, plants, basstraps, plus height and distance of the listening place optimized. It is always shocking to see how 1 meter further into the room certain low frequencies are completely gone.

    Anyhow, I measured in one go the auto HPEQ from my Hifiman Ananda and my Sennheiser HD800. Back at home I wasn‘t very satisfied with my Hifiman Ananda measurement. It sounded over HDMI a bit left sided. Just more fuller. Strange things is that over SPDIF Coax this was less the case. I even thought that maybe this left sidedness was still a problem with the room and the placement of the accoustical elements.
    I now tried the HD 800 HPEQ and it is brilliant. So I guess, that it has again something to do with the inEar microphone placement or the way the inEar microphones are built. It would be interesting if one finds out how to use the A8 microphones with the A16. @Rene Lou ?
    Last edited: Nov 25, 2019
  5. Richter Di
    That is strange. Did you use USB or HDMI?
    I used HDMI and Qobuz says it is using WASAPI. On Foobar I have to check again.
  6. Richter Di
    BTW - tried yesterday for the first time to use my Android Samsung mobile with the usb audio player pro (UAPP) and an OTG dongle with the Smyth Realiser A16. While the UAPP recognized the A16 and asked for confirmation to connect with it, there was no sound coming out of the Realiaser, only through the Samsung mobile. As I have used the same configuration with @Jan Meier ’s Soul and his Corda Daccord DACs I am pretty sure that it usually works. Normally the UAPP when connecting with the device for the first time also tells you which frequency and bit rates the DAC is able to process. In this case nothing came up. I used the USB port on the back side. Did anyone else try it? Have I to do something differently?
  7. dsperber
    Am I not understanding something? The HPEQ doesn't have anything to do with "acoustical elements". The headphones are placed over the ears which still have the microphones inserted. The HPEQ reflects solely the headphone characteristics (along with your ears, hearing system, etc.), not the room in which the HPEQ measurement is taken. The HPEQ can be measured anywhere, e.g. in your kitchen or bathroom, back home, anywhere, i.e. doesn't have to be done in your Pilates studio.

    The acoustic character or the room (which is obviously effected by what you do with acoustic elements) is represented by the PRIR, not the HPEQ.

    Granted, I've not yet gotten involved even once with PRIR and HPEQ creation with the A16, so maybe I'm missing something. But am I not correct? If your Ananda HPEQ sounds very different from your HD800 HPEQ (using the same PRIR, I would assume), the difference seems like it must stem from how you placed the microphones in your ears for each HPEQ creation... not anything to do with your room treatments which would affect only the PRIR you're using.

    Am I missing something?
  8. Richter Di
    No, you are completely correct. It was only me who thought listening to the PRIR and the HPEQ of the Ananda, that we did something wrong with the room acoustics. but to my surprise it was the HPEQ of the Ananda, since the HD 800 was great without the any left sidedness.
  9. audiohobbit
    I already asked Stephen about the new microphones and why the vents are looking downwards. He says it is only because of the form factor of the circuit boards. It is rectangular, so can't be placed in a small round shape.
    He says the mics still work omnidirectional, same as the mics of the A8. They did comparisons between A8 and A16 mics and heard no difference.

    So this contradicts your findings a bit, Rene.
    I asked him about that though, but don't expect an answer soon, seems they are very busy at the moment.

    @HT Assist: You can use this also when not facing directly the speakers: Either place the set top at the new center angle or build a rotating disc with an angular scale to rotate the head top on the head band to the desired center angle.

    Stephen recommends 12 s non overlap sweeps and 15 degree increments in horizontal look angles.
  10. blubliss Contributor
    I use USB now but have used HDMI. I'm pretty certain the Smyth PC driver is ASIO.
  11. sander99
    So, as you may have read already I visited @Richter Di and we did some PRIR and HPEQ measurements for the both of us. We only managed to do 2.0 because we needed time to carry some absorbers, bass traps and carpets into the room, and set the speakers up etc. And then we first had a few less satisfactory measurements due to microphones moving out of position.
    But the end result: magnificent! Very nice to hear the Realiser in action again after the 2.5 years that passed since my demo in Munich. It was a confirmation for me that those old memories weren't just a dream, the A16 is real!
    And now actually it was the first time I listened to normal stereo over a stereo pair of virtual speakers instead of Dolby ATmos 7.1.4 over a 7.1.4 set of virtual speakers.
    And the speakers (Focal Aria 906), and the way we set them and the room up was probably much better sounding than the Smyth demo setup, except only limited to 2.0 channel of course.
    So thank you Richter Di for this oppertunity!

    My PRIR and HPEQs are on the Google Drive managed by @Rene Lou: https://drive.google.com/drive/folders/1MWWb8KELPBbAC5WSHf82UXeV73v1O84W?usp=sharing
    jgazal, sm0rf and Richter Di like this.
  12. Richter Di
    Does the headtracker guidance work with 15 degrees and how?
  13. Rene Lou
    Then Stephen is wrong, sorry. I know what I'm hearing. Maybe the A16 microphone is omnidirectional when measured 1m away from the microphone, but it is definitely not omnidirectional in the near-field. Did they measure the A16 microphone in the near-field, near-field be in the 3-5mm range ? If you have a look at the picture of my ear, reflecting ear surfaces are between 1 and 10 mm away from the microphone - and these ear reflections are important to catch when you try to get a perfect PRIR in terms of sound localization.

    To verify that the A16 microphone is not omnidirectional in the (mm) near-field I just did a very simple test you can do for yourself as well. I used an online frequency generator at https://www.szynalski.com/tone-generator/ and played a 3kHz sine tone over an iPhone in ear headphone (EarPod). I used the A16 App 'Listen to microphones on HPB' as a test and tried to get the same reading in dB when moving the in ear headphone around the microphone in very close distance. For a reading of approx -30dB I have to hold the headphone directly to the rear of the microphone (0mm distance). In front of the microphone opening, I have to extend the distance to approx. 5mm to get approx. -30dB. When I then move the headphone around the microphone at approx. 5mm distance I get a reading from behind the microphone of approx. -40dB. In the near-field, the A16 microphone is definitely not omnidirectional (this will be the reason why some of my PRIRs are working very well with a specific microphone orientation and other PRIRs with other microphone orientations do not).

    Are you informing Stephen about that, or should I ? In my posting about optimal A16 microphone positioning in the ear I just described what I'm hearing when comparing PRIRs and my high-end speakers, and I'm very happy that I could finally make a copy of them with the A16. For doing so, I had to rotate the microphones.

    When you have an A16, try it for yourself - it's quite easy, just make some PRIRs with one microphone position and some others with another position.
    Last edited: Nov 25, 2019
    arnaud and Richter Di like this.
  14. Rene Lou
    I will ask James about how to connect the A8 microphones to the A16 but will need some time to do new PRIR tests after I have received the information.
    Richter Di likes this.
  15. audiohobbit
    I already asked Stephen about your rearward facing mics, but they are very busy at the moment testing the next firmware update which should come out maybe next week or so.
    Last edited: Nov 25, 2019
    Richter Di likes this.
483 484 485 486 487 488 489 490 491 492
494 495 496 497 498 499 500 501 502 503

Share This Page