Author |
Message |
Vozka
Joined: Feb 15, 2015 Posts: 4 Location: CZ
|
Posted: Sun Feb 15, 2015 7:49 am Post subject:
|
|
|
Hello, I bought the AmbiophonicDSP a couple weeks ago and I must say I'm a bit disappointed.
While it probably is the best VST solution there is, I was surprised at how much it colors the sound. I'm honestly still not sure whether it's supposed to do this or if it's some nasty bug. None of the settings is usable at all except maybe Pop, which only cuts some treble, but it's still better than all the others that cut pretty much all the bass and some midbass. Wareing is the worst, but the others aren't much better.
I was able to make it sound close enough to the original (45 microseconds delay, 46% recursion, 10.0 space, 5.3 zentrum), but it was a challenge to make it sound neutral, instead of a challenge to make the stereo sound best.
None of the other VSTs and applications did this, not even the previous 2 knob mosc-ambio plugin. Some did affect the sound color in the middle of the stereo field, but for example the Stephan Hotto transcoder/player sounded pretty much perfect.
This happens both in Foobar VST host and when using Reaper as a transcoder, on two different sound cards, so I'm guessing it must be the plugin - is it really supposed to sound like this? |
|
Back to top
|
|
|
mosc
Site Admin
Joined: Jan 31, 2003 Posts: 18197 Location: Durham, NC
Audio files: 212
G2 patch files: 60
|
Posted: Sun Feb 15, 2015 2:38 pm Post subject:
|
|
|
Voska to electro-music.com
As moderator, I split this from the product announcement topic. I think it's best to be a separate topic.
I'm sorry to hear that you are having this problem. I assure you that the plugin was designed to make no negative changes to the sound. We will try to help you sort this out, or if you wish you can get a refund.
First, which version are you using?
Second, can you get any setting that sounds clean? Try:
Space: 0
Zentrum: 10 (max)
How does it sound on the Bypass setting?
Third, can you make recordings of the input and output so I can hear it?
If any AmbiophonicDSP users have any ideas, please feel free to post them here. We are aware that this is a psycho-acoustic processor and there are some people that do not hear the effect. _________________ --Howard
my music and other stuff |
|
Back to top
|
|
|
Vozka
Joined: Feb 15, 2015 Posts: 4 Location: CZ
|
Posted: Sun Feb 15, 2015 4:27 pm Post subject:
|
|
|
Thank you for the response. I forgot to mention one thing - the stereo effect itself works great. I found that room reflections play a pretty big role in the effectiveness - it works best with speakers that have narrow directivity (mine use pressure drivers with waveguides) so that there are fewer early reflections.
I'm using AmbiophonicDSP version 1.2 with the default settings extracted from version 1.1.
I've made recordings and spectrum analysis screenshots of a white noise generator (I used ARTA) fed into Reaper. I can post actual music as well if you want, but this seemed like a pretty good idea once I took a look at the spectrum.
Here are the recordings as wavs in a zip file: https://dl.dropboxusercontent.com/u/66108396/ambio.zip
This is AmbioDSP with the bypass setting. The zip also includes a recording with the plugin turned completely off, but it looks and sounds the same as this, so I assume no problems here:
This is the electro/game setting:
This is the jazz/acoustic setting:
Setting space to 0 and zentrum to 10 produces pretty much the same spectrum as electro and jazz setings. This is a setting that I have found sounds the best, although not ideal, and the spectrum seems to agree with me:
The pop setting is pretty similar:
And finally, this rather silly thing is the wareing setting:
It seems like the space setting is the main problem - the lower it is, the closer to the wareing frequency response it gets. Setting Recursion at more than 50% also produces strong ringing-like resonance (like a highly resonant filter) at high frequencies, but this vanishes when using shorter (45 microsecond) delay. Getting more recordings shouldn't be a problem, just tell me what you need.
Oh, and in case you're interested in my system specs - my soundcards are Edirol UA-1EX and Cakewalk UA-1G and I run windows 7 64bit. It doesn't seem to matter whether I use ASIO, WASAPI or DirectSound. |
|
Back to top
|
|
|
mosc
Site Admin
Joined: Jan 31, 2003 Posts: 18197 Location: Durham, NC
Audio files: 212
G2 patch files: 60
|
Posted: Sun Feb 15, 2015 5:41 pm Post subject:
|
|
|
Great post. I'll need some time to digest this, but wanted to yet you know I appreciate the effort you are putting into this.
One comment, which may be the source of everything. This is a psycho-acoustic processor, so we are changing the audio signals to provide better stereo. We expect some frequency response variations.
We intentionally designed the controls to go completely out of usable range. For example, we think recursion is too much beyond about 30%, but we allow it to go to nearly 100% so you can tell more of what it is doing. I personally don't use recursion at all, but we included the control to be consistent with the RACE algorithm. In theory, recursion will enhance stereo separation, but YMMV (your millage may vary).
Zentrum and Space are similar. I will use different setting for different kinds of music. I too like space up all the way most of the time, but Robin Miller doesn't. For classical music, I sometimes turn it down.
So, the plugin can in fact seriously change the ultimate frequency response depending on the control settings, as you have correctly observed. The key is to adjust things so they sound best to you, and it will change depending on the type of music you are playing.
There is a system called the BACCH developed by Edgar Choueiri at Princeton University. His system doesn't offer these adjustments as it is for a fixed speaker configuration in an anechoic chamber. This system uses an optimized FIR filter to equalize the output, to compensate for the effects you have observed.
IMHO, the frequency response of the output electrical signal isn't as important as the sound you hear. The important thing is the frequency response at your head, with your ear/mind. In an ordinary stereo system with no processor, the frequency response is terrible. It won't show up on any test of the electrical signals, but it is a result of the comb filtering produced by the interaction of the sound in your ears coming from the two widely spaced speakers. While the post processing optimized filter used by Edgar Choueiri may work in an anechoic chamber with fixed placement of the speakers and a specific head size, it doesn't work in a real room with various speakers and speaker placement and head sizes.
That said, you may wish to try using an equalizer after the AmbiophonicDSP. In my experience, this is more trouble than it's worth, but I'd be interested if you would find it useful.
One more point. Conventional stereo (60 degree separation) works best for stereo sources that have very wide separation fields, like for electronic music that has different material routed to each of the two speakers; it works worst for mono. Unfortunately, most recorded stereo music has the most important sonic material at the center of the sound stage - the center equally routed to left and right - the mono component of the stereo signal. Our AmbiophonicDSP does the best at presenting the center stage - cross talk comb filtering is reduced by the close placement of the speakers. We use the processor to cancel as much of the cross talk as possible to spread the signal to the left and right. Fortunately, most people hear this as improved separation; wider and more precise than conventional stereo. So, with the AmbiophonicDSP you get better center and left and right extremes. _________________ --Howard
my music and other stuff |
|
Back to top
|
|
|
mosc
Site Admin
Joined: Jan 31, 2003 Posts: 18197 Location: Durham, NC
Audio files: 212
G2 patch files: 60
|
Posted: Sun Feb 15, 2015 6:09 pm Post subject:
|
|
|
I've reviewed your files and screen scrapes. I think you are seeing what is happening, and if you don't hear any crackles, clicks, hums, whines, or other obvious distortion, then the processor is working as designed.
Your favorite settings are pretty much mine too. I use less recursion and a little more delay. Delay is a function of the distance between your speakers, the distance from the speakers to the listener, and the width of the listener's head. There are minor effects of altitude and humidity too as these influence the speed of sound. It's hard to hear the effect of delay, so we made the control have a very wide range.
As for the Wareing sounding sounding pretty bad, we agree. We put that setting in to honor David Wareing who we respect very much. As far as we can determine, he was the first to start experimenting with DSP for cross talk cancellation. At the least, his algorithm was the first one we discovered, and it triggered our research and development. We think we've come a long way from his first attempts, but he was a pioneer and deserves our respect. When you start with the Wareing settings, you can hear what each control does as it enhances the algorithm. _________________ --Howard
my music and other stuff |
|
Back to top
|
|
|
Vozka
Joined: Feb 15, 2015 Posts: 4 Location: CZ
|
Posted: Mon Feb 16, 2015 4:31 am Post subject:
|
|
|
Ah, in that case everything is as it should be I guess.
I understand that simple measuring, like spectral analysis, may be misleading when talking about psychoacoustics, but I used it because it quite accurately represents how music actually sounds to me.
I do use an equalizer after the AmbiophonicDSP plugin and had no problems with it (I raise the treble a bit and cut a specific bass frequency where my monitor setup produces a lot of resonance). Theoretically there shouldn't be any as long as you apply the same EQ on both channels, right?
As for the default settings - considering I've had nearly the same results with three different pairs of speakers in three different rooms, wouldn't you think it would be best to use a setting like yours or mine as the default?
In a similar vein, while it is interesting to hear what the wareing setting sounds like, or trying what some extreme settings sound like, wouldn't it be better to hide those under some button labeled "advanced" or something similar?
My reasoning for these suggestions is this: ambiophonic playback is amazing and brings considerable improvements even on very cheap speakers, but for the average user it's just too complicated to setup. It will never work completely out of the box since one has to fiddle with the speaker positions a bit and preferably adjust a couple settings, but the closer to painless the software is, the more people will be able to use it. From this point of view, making the settings too broad doesn't make sense, neither does including the wareing setting - the end user doesn't care unless he's already interested in the technical side and it will just confuse him or turn him away. While I am interested in the technical side, none of my friends whom I recomended ambio to are. |
|
Back to top
|
|
|
mosc
Site Admin
Joined: Jan 31, 2003 Posts: 18197 Location: Durham, NC
Audio files: 212
G2 patch files: 60
|
Posted: Mon Feb 16, 2015 1:36 pm Post subject:
|
|
|
Your suggestions are very well thought out. I'm working on another version of the software, but I was going in the other direction, one more knob.
I guess a lot of knobs is what you get from a couple of engineers.
Time to do some re-thinking.
Anyway, thanks for your feedback. I enjoyed this conversation.
Happy listening... _________________ --Howard
my music and other stuff |
|
Back to top
|
|
|
Vozka
Joined: Feb 15, 2015 Posts: 4 Location: CZ
|
Posted: Mon Feb 16, 2015 2:17 pm Post subject:
|
|
|
Thank you, I'm looking forward to try any new features!
As a compsci student I understand the desire to add more options very well . I was thinking myself that it might be useful to set the "space" setting (or amount of cancellation?) for each channel separately, since the most annoying problem in two different bad acoustic settings for me was that one side of the sound stage extended further to the side than the other.
I'm going to try to write my own transcoder prototype just to try little things like this when I have more time since it's such an interesting topic. |
|
Back to top
|
|
|
|