Become a Site Supporter and Never see Ads again!

Author Topic: Zylia ZM-1, some thoughts  (Read 1008 times)

0 Members and 1 Guest are viewing this topic.

Online WiFiJeff

  • Trade Count: (1)
  • Taperssection Member
  • ***
  • Posts: 808
  • Gender: Male
  • I tape therefore I am.
Zylia ZM-1, some thoughts
« on: December 17, 2018, 09:52:35 PM »
I have been interested in surround sound and ambisonic recording since I got a Josephson C700S some years ago, but so far haven't done too much with it (I mostly use the Josephson for Blumlein stereo, using the third omni capsule to touch up bass response or turn it into a mid/side stereo, not for surround).  As I mentioned on a thread after the NY AES meeting, ambisonics seems to be the flavor of the month, I saw the Polish "third order ambisonic," 19 channel Zylia mic and read the review of it in the October Recording issue and decided to bite.  I have been playing around with it for a couple of weeks, and have recorded two string quartets with it as backup mic (to a Schoeps MK4 ORTF pair and the Josephson C700S). 

The selling point on the Zylia is use for bands in rehearsal or to make virtual reality videos where the sound follows your head.  My intended use as a mic for recording classical acoustic chamber music is a real outlier.  Since the Josephson C700S produces native B-Format ambisonic files, WXY channels, no Z for height info, my idea was to use the Zylia Ambisonic Converter software to produce first order FuMa B-Format (canonical WXYZ channels) to mix and compare directly to the Josephson.


The default setup in the Zylia Studio recording program (I am using an Ockel Sirius A Pro pocket Windows 10 machine that runs all my other audio software) records 24 minute takes, at which point (4 GB at the default maximum 24/48 resolution for 19 channels) it begins a new file, but not seamlessly.  It also normalizes the files separately if you choose to export a standard stereo mix.  If you choose to record wavpack files instead of wav files you can record much bigger files.  The program can convert these to wav, but so far also not seamlessly, I find I have to cut about 10ms out of the join, which is a pain.  So I have learned about wavpack and wvunpack, which are the first command line programs I have dealt with since the early '90s I think.  Anyway, Zylia Ambisonic Converter will make B-Format files from wv files without issues, I have purchased Reaper which handles wavpack files (when I figure out how to use it) for the dealing with the full 19 channels for a custom mix (I should live so long). 


The Zylia uses 19 omni mics with all sorts of channel cancelation and beamforming, I was skeptical how it would produce something I could mix with phase inversion.  And initially what I got was very disappointing; this was not like mixing the Josephson X and Y channels!  I have morphed into a beta tester - it turns out that the Zylia FuMa B format did not use the canonical WXYZ ordering but erroneously had the channels ordered WZXY.  When this was corrected for (they have promised that the next version of Ambisonic Converter will have the ordering corrected), I found that my stereo mix-down had better localization even than the Josephson.  A friend who tested on better stereo equipment was even more taken with the Zylia.  This is astonishing considering another unlikely factor: the Zylia has no preamp or volume controls (they suggest that if the sound clips or distorts, move the mic further away from the instruments, something I am used to doing for my Edison cylinder recording horn but nothing else lately).  For a string quartet about four feet away from the mic, this meant having to boost the Zylia mix in Wavelab by over 20 dB. 

I am interested in feedback on the sound.  I get different impressions of Zylia vs Josephson depending on the headphones I listen on, so I'm curious what others hear on their setups.  The clips I have linked to are from a transcription which bounces musical lines from violins to cello and viola, which I found fascinating in itself.

https://spaces.hightail.com/receive/2oytmsg3vL

ps the Zylia sphere, the size of a slightly too large softball, has an LED equator that I had to cover with a velcroed shoelace belt to keep the psychedelic factor down to classical chamber music levels.

Offline Gutbucket

  • record > listen > revise technique
  • Trade Count: (13)
  • Needs to get out more...
  • *****
  • Posts: 12882
  • Gender: Male
Re: Zylia ZM-1, some thoughts
« Reply #1 on: December 18, 2018, 11:17:16 AM »
Thanks for posting Jeff, very interesting.  My "seat of the pants"  take on higher order Ambisonics (derived from my research into the subject, although as yet to be confirmed by my own listening which is very limited) is that that 3rd order is where the differentiation of higher-order directionality begins to become readily apparent.

To clarify a bit on what you posted- I assume the 2-channel Zyla sample is a virtual microphone pair output from the Zyla Ambisonic Converter.  Is that output a virtual a pair of 1st-order virtual microphone patterns (output of the same "order" as the Josephsen, in addition to the same virtual pattern shape and angle)?  Or is the Zyla sample an example of a "higher than 1st order" virtual microphone pattern, yet with a similar included angle between the virtual microphones?

I've downloaded your samples, but only have cheap Samsung phone in-ear 'phones to listen through at the moment.  Will be interesting comparing on better equipment when I can.

Thanks again for sharing your experience and the samples!
musical volition > vibrations > voltages > numeric values > voltages > vibrations> virtual teleportation time-machine experience
Better recording made easy - >>Improved PAS table<< | Made excellent- >>click here to download the Oddball Microphone Technique illustrated PDF booklet<<

Online WiFiJeff

  • Trade Count: (1)
  • Taperssection Member
  • ***
  • Posts: 808
  • Gender: Male
  • I tape therefore I am.
Re: Zylia ZM-1, some thoughts
« Reply #2 on: December 18, 2018, 12:14:58 PM »
Yes, the recording I put up is the Zylia virtual mic first order FuMa X and Y channel mix to Blumlein stereo.  I have not tried higher order because I haven't mastered Reaper software yet, where I could apply ambisonic plug ins like ATK (Ambisonic Tool Kit) for Reaper and the Zylia Pro Studio plugins to do fancier stuff.  I assume it is using a fair number of the 19 channels to do this mix, the omnis are MEM (micro electro-mechanical) mics from Infineon.

Looking forward to your reaction on upscale transducers!

Offline rumbleseat

  • Trade Count: (0)
  • Taperssection Regular
  • **
  • Posts: 68
  • Gender: Male
Re: Zylia ZM-1, some thoughts
« Reply #3 on: December 18, 2018, 09:07:26 PM »
What a wonderful musical selection!  Thanks for sharing.
Although the two "mics" have very different frequency responses, I like the separation between the two violins in the Zylia a bit better.
It might be possible to EQ the the two recordings so they match by comparing them on a spectrum analyzer and applying EQ.  On my limited system, I collapsed both to mono and ran one on each channel into a stereo frequency analyzer.  From there you can tweak the EQ to get them to "look" and sound the same.  I could follow through with that if you think it would help.
AKG C480B CK61 cards >  Canare L-4E6S with Neutrik EMC > Tascam DR-680 MKII > memories

Offline Gutbucket

  • record > listen > revise technique
  • Trade Count: (13)
  • Needs to get out more...
  • *****
  • Posts: 12882
  • Gender: Male
Re: Zylia ZM-1, some thoughts
« Reply #4 on: January 10, 2019, 05:24:18 PM »
I don't currently have my good speaker system setup, but finally got around to listening to these samples using Senn HD700 'phones.  I didn't use a high quality DAC or amp, but the better-quality headphones were sufficiently revealing to make the following observations-

Most obvious difference is frequency balance.  The Zylia recording has more low-frequency content and less high-frequency content than the Josephson.  Difference seems to be pretty linear across the spectrum, which I loosely confirmed by using the graphic EQ in VLC to match the Zylia recording to the Josephson (approximately and by ear only), ending up with the equivalent of a linear 'tilt' EQ curve (looking like a straight line rising upward from LF to HF).  This leads me to believe there may be an issue with the EQ calibration of the Zylia, or at least with this particular output rendering of the raw recording data, as the Josephson presumably has a response which is balanced appropriately and it sounds that way.  This is not to say the Josephson recording could not benefit from some EQ, but that seems specific to this particular recording rather than the microphone itself.

Comparing both after rough EQ and level matching to even the playing field, the Zylia seems to provide increased directional spatial resolution at low frequencies, but far less spatial resolution and channel differentiation at mid and high frequencies.  It seems to grow muddier and increasingly monophonic as frequency rises.  The Josephson by contrast sounded sweet, open, precise and far more pleasant overall, particularly in the midrange and above.

The Josephson is a very nice microphone and sets a high bar sonically and the Zylia definitely falls short of that, at least in my subjective listening impression given this particular sample.  Besides listening, I did not otherwise analyze the files other than taking note of the EQ curve used minimized the frequency response differences.

Big thanks for this thread and making these samples available Jeff!  It will be interesting to find if the Zylia output improves once you get more experience with it, including possible calibration corrections and higher-order renderings.  Looking forward to reading more about your experiences with it going forward.

musical volition > vibrations > voltages > numeric values > voltages > vibrations> virtual teleportation time-machine experience
Better recording made easy - >>Improved PAS table<< | Made excellent- >>click here to download the Oddball Microphone Technique illustrated PDF booklet<<

Online WiFiJeff

  • Trade Count: (1)
  • Taperssection Member
  • ***
  • Posts: 808
  • Gender: Male
  • I tape therefore I am.
Re: Zylia ZM-1, some thoughts
« Reply #5 on: January 10, 2019, 06:02:32 PM »
Thanks for this feedback.  I have, since posting these samples, become a novice Reaper user, at least to the extent of being able to load the giant 19 channel Zylia .wv files into it and apply the Zylia Studio Pro plug-in.  This allows me to define a large number of "virtual microphones" pointed at specific instruments with various pattern widths, or even a virtual omni (W channel).  I have reprocessed my earlier recordings with this, and like the results even better.  I am not able to do the sound check for each instrument recommended in the Zylia recording app (Zylia Studio), and it may be that mis-aiming the virtual mics in post continues to weaken the HF response, I'll post those samples next week (I will also be recording a piano-and-wind quintet in a week).  It is also possible to use the Zylia Ambisonic Converter as a Reaper plug-in.  I am really having fun with this mic.

Jeff

Offline Gutbucket

  • record > listen > revise technique
  • Trade Count: (13)
  • Needs to get out more...
  • *****
  • Posts: 12882
  • Gender: Male
Re: Zylia ZM-1, some thoughts
« Reply #6 on: January 11, 2019, 10:41:35 AM »
Cool stuff.  Your comments above parallel my experience with the TetraMic when used on stage relatively near the sources. I found it especially useful and fun to listen while reorienting a single solo'd virtual microphone so as to tweak focus on a particular instrument, or particular portion of the drum kit.  And although I used that technique to determine the best configuration for a virtual stereo pair output, I imagined that mixing multiple virtual microphones, each optimized for a particular instrument made practical by the tighter patterns of a higher order system, would be the most significant improvement of a higher-order ambisonic capture system over an optimally-arranged stereo coincident microphone.

In other words, with stereo output in mind the primary value I see in 1st order ambisonic microphones is the ability to optimally configure and orient a virtual stereo pair after the fact.  That's pretty cool.  Yet I suspect a sufficiently higher order ambisonic microphone would provide enough increased directionality to be able to extend that to isolating sources sufficiently such that it could make sense to build a stereo mix from individual virtual microphones or pairs focused on each individual source - something altogether different and cooler.
musical volition > vibrations > voltages > numeric values > voltages > vibrations> virtual teleportation time-machine experience
Better recording made easy - >>Improved PAS table<< | Made excellent- >>click here to download the Oddball Microphone Technique illustrated PDF booklet<<

 

RSS | Mobile
Page created in 0.174 seconds with 35 queries.
© 2002-2019 Taperssection.com
Powered by SMF