Jump to content
IGNORED

Tweaking Sound with Convolution


Recommended Posts

I have attempted many different kind of tweaks to improve my sound. Cables, tubes, HiRez, power conditioner and stabilzer, cones/spikes. One on occasion, I even submerged the speakers stands in water and sand bagged them with fine sand....... dedicated power supply from solar panels. Far too many to mention here.

 

My current system is rather simple and the best tweak I added is the convolution channels for the stereo playback. The additional price I paid is just about $600 to get the additional twenty over speakers and DACs. I got these.....

 

amplifiers.thumb.jpg.4324ef204f43a92ef4e9371db8032a25.jpgDAC.thumb.jpg.f112650ea83b6dfae969f4499ee84b72.jpgsidespeakers.thumb.jpg.8a7a1ec0af122650039e521239b4e9a3.jpg

 

These addition now allow me to experience far better improvement  as can be seen in the videos below.

 

These videos were made in my room from my listening chair using a cheap binaural microphone (Roland CS-10EM). You have to use headphones to hear the changes. Due to 18 additional speakers recreating the ambiance by adding to the main speakers, the sound played directly through speakers can be muddy and congested and therefore use headphones to hear closer to what you would hear at the listening seat. Due to the limitation of the $75 microphone, the high frequencies are too bright. It also failed to capture the low level ambiance.

 

Some points about these videos:-

 

1) No changes made to the files. All were recorded in 24/96 format and uploaded as such. However, youtube would have compressed the audio.

 

2) They were recorded using Zoom H1 recorder connected to Roland microphone. Level was set at 90 of 100.

 

3) The actual music level while making this recording at  about 214cm from the main speakers were Avg =  76dB and Peak at 87.5dB.

 

4) The actual preference is convolution at 50 or 80%. In this video, the first one was at 50%, second one at 80% and the third at 100% level which I use for large orchestra music with high ambiance content. 

 

4) This setup is my early attempts to recreate the 3D concert hall ambiance in small space. I will add more info about this later.

 

6) You may hear slight timing difference due to my omission when changing the default values when doing this demonstration videos.

 

7) You may hear the air conditioner which is getting a little noisier lately. 

 

8) The videos below are to show the effect of convolution. See here for proper use of them.

 

 

 

 

 

 

 

See here for proper use of them

 

Link to comment

Very interesting experiment.


Not sure what to think of the youtube vids.  It sounds as if the sound stretches from ear to ear though not on a flat line so not stuck inside my head exactly.  The stereo effect is minimal sounding almost like fat indistinct mono with sometimes side reflections.  I get a strong wavering image or flanging effect with the first one, and lesser though still substantial wavering with the third one.   The second one has a barely noticeable amount of this effect.  Second one sounds lower in level than the other two.  I adjusted by ear for that.  Also I hear some noise floor pumping in the second one particularly near the beginning I don't hear with the others. 

 

I wouldn't make too much of how it sounds to me.  I have already noted elsewhere that binaural rarely works well and usually not at all for me. 

 

Oh, and I listened over Sony MDR7510 phones from a Focusrite Forte USB interface.

 

Decided to try some Creative Aurvana Air phones.  Sort of an on ear/nearly in ear set of phones.  With these I had more stereo effect.  The sound however clustered right at each ear with something of a whole in the middle.   The noise floor pumping was more obvious on the second vid.  It seems to come and go in the whole in the middle of the image.

 

Then tried some Beyerdynamic DT880's.  Stereo effect was pretty good with these, but sound moved to the top of my skull filling about the middle half of my head sort of sensation. 

 

Will be interesting to see how it sounds to others for whom binaural works.

And always keep in mind: Cognitive biases, like seeing optical illusions are a sign of a normally functioning brain. We all have them, it’s nothing to be ashamed about, but it is something that affects our objective evaluation of reality. 

Link to comment

I would need to listen to the system in loco in order to provide any meaningful assessment but I agree with Dennis that there are a few disturbing artifacts to the sound; I had already complained about these in the immersive sound thread after listening at the samples in some web site.

I won't comment on the space effects because I don't generally listen to music with headphones though this was how I listened to your videos.

Yet I much prefer the tonality (and even the spatial rendition) of the original 2ch stereo version....yours sound like they've been recorded inside a deep well.

 

 

This is also definitely not what I would call adequate music for testing.

Can you record a cello or violin and piano sonata by recorded by BIS or Dorian?

Or maybe one of Mario's (PlayClassics) tracks?

"Science draws the wave, poetry fills it with water" Teixeira de Pascoaes

 

HQPlayer Desktop / Mac mini → Intona 7054 → RME ADI-2 DAC FS (DSD256)

Link to comment

Thanks for the feedback.

 

The purpose of the video is to show the effect of convolution and how it alters the sound. For music like Mad World the convolution should be turned off completely. I wanted to show the effect but we are quite aware of the well like effect of what's coming out of the cheap microphone recordings. 

 

I am am not really keen with violin or saxophone as it tend to hide room flaws. On the other hand piano is difficult to be captured. Norah Jones is very good but I cannot capture the soundstage. Will see what recording can be used. This another Hifi system playing Mad World. 

 

 

 

 

 

 

 

Link to comment

They were meant for very low volume convoluted sound. Fidelity is not important there and I will be adding more speakers and separating the cables would be impossible.

 

 

Anyway, the above video meant to show the difference caused by convolution. The effects were exagerated. Below is two sample one with Convolution for music without much ambiance and another one for classical music and added with convolution.

 

 

 

And Brahms - Sonata for Violin and Piano.

 

 

Link to comment

Okay, I can hear what the convolution is doing in the second clip; but, unfortunately, for me, the violin is too edgy, particularly when played more expressively - the natural "sweetness" of the instrument is missing, in this presentation.

Link to comment

Thanx for feedback. But this is a clear example, that Youtube cannot replicate what the system sound is accurately. It can show room flaws but it will not reveal everything.  Let alone to use that to pass judgment of stereo setup recorded miles away from the sweet post to tell how good a system is like what you do. Is it so difficult to reduce the gain from the sweet spot and make a video?

Link to comment

This is good ... you're motivating me to get the NAD combo up and running again - it's working without a volume control at the moment; as I point out in my blog this was a significant quality weak link. I've got a replacement, which I've been "too busy" to install, ^_^. So, stand by ...

Link to comment
1 hour ago, fas42 said:

This is good ... you're motivating me to get the NAD combo up and running again - it's working without a volume control at the moment; as I point out in my blog this was a significant quality weak link. I've got a replacement, which I've been "too busy" to install, ^_^. So, stand by ...

 

I like to believe that you have something to offer but I am not getting it. Trust me, if someone were to post saying that placing a bucket of water in your room would make your sound more natural, I will give it a try and then look for the scientist explanation (caveat- it doesn't mean I will do everything). :)

 

So please show us something tangible.

Link to comment
2 hours ago, STC said:

So please show us something tangible.

 

Well, I don't have a clear idea of what you mean by tangible - at this moment, right now, I don't have a system which can produce the quality I talk about; and the NAD setup hasn't been finalised, not by a long shot. And by that I mean it still hasn't given me the "invisible speakers" performance level - close, but not the full monty.

Link to comment

Alright, STC made available to me downloads of the three files in the initial post that had not been thru youtube processing.  This was my response to him via PM about those:

 

Okay, I downloaded the files.  Much nicer.  I suggest no more youtube postings of this.  The compression they are doing leaves too many audible artefacts.  The volume on these are reasonably close matched now.  I wonder with the huge number of files uploaded to youtube if you uploaded the same file twice it might end up with different artefacts in each copy. 

 

The image with my preferred Sony phones is from ear to ear though not strung along a single line.  It seems to have some size, and doesn't feel inside my head in a claustrophobic sense.  More like listening in a miniature room the size of my head.

 

I do still seem to hear more noise floor modulation in the first 20 seconds of the 80% track. I do think the 80% track sounds more believable.  The 40% track does not sound like flanging or wavering now.  Though a small amount of shimmer to the images.  The full 100% track has what sounds like too much sidewall reflection and sounds like an artificial effect.  In general these sound more similar so most of the difference in the youtube tracks must be due to the compression procedure done by youtube.

 

The convolution you are doing seems particularly susceptible to being messed up by what happens to youtube audio.  So these tracks also sound of higher quality overall.  The youtube is smearing noise all through the track it sounds like. 

 

 

Finally here is a link he provided which works much better for me binaurally with some real 3D qualities to it.  BTW, was this recorded with the Roland CS-10 EM microphones?

 

 

And always keep in mind: Cognitive biases, like seeing optical illusions are a sign of a normally functioning brain. We all have them, it’s nothing to be ashamed about, but it is something that affects our objective evaluation of reality. 

Link to comment
11 minutes ago, esldude said:

BTW, was this recorded with the Roland CS-10 EM microphones?

 

 

Thanks Esldude for the feedback. 

 

Yes, it was recorded with CS-10EM. I thought I had a recording of it playing through the loudspeaker because I use this track to setup my system. Will make another one and see how Youtube is distorting the soundstage.

Link to comment

Oowwww, that would be smarty pants! :)

 

BTW, something "tangible" to what I'm about, is understanding that if I'm asked to assess reproduction that I never listen for what's right in the playback - only to what's wrong ... so, I'm always, always going to point to anything which "gets in the way" of making me believe that what I'm listening to is "live".

Link to comment
On 2017-6-18 at 4:49 PM, STC said:

They were meant for very low volume convoluted sound. Fidelity is not important there and I will be adding more speakers and separating the cables would be impossible.

 

 

The issue is to create cross-talk between the channels due to the induced currents. Cross-talk affects the sound you want to create. Hopefully, at low volumes this is not significant however something to keep in mind. 

 

I tried to listen to the youtube videos however as elsudde pointed out they doesn't seem to make much sense. 

Link to comment
1 hour ago, unbalanced output said:

 

 

The issue is to create cross-talk between the channels due to the induced currents. Cross-talk affects the sound you want to create. Hopefully, at low volumes this is not significant however something to keep in mind. 

 

I tried to listen to the youtube videos however as elsudde pointed out they doesn't seem to make much sense. 

 

 

I am am not sure what crosstalk you are referring to. There are many videos there. The first three and after that another two. 

 

According to Google analytics the average time viewers spent on the these videos was just one minute. With such a short time many would have missed the notes in the video. 

 

 

Link to comment

Looks like it is impossible to capture the illusion created by two speakers with binaural microphones.  I find this very strange because the microphone is not capturing the sound as how we perceive them. There must an explanation. One that I can think of is the close proximity of the two speakers. They are almost touching each other and the bigger than stereo soundstage is heard because of crosstalk cancellation. See the picture below.

 

 

F978447E-A8D0-436A-B853-039E81CCD3E0_zps0gbl1gry.thumb.JPG.5a49b4e02c2be232d185bdb76dde9886.JPG

 

 

 Anyway, I made another attempt by putting a divider in front of my face and I used a pillow. I got better separation but the front stage sounded muffled . Then I used a small block wrapped with cloth. So the result is slightly better. As per post one, the recordings were done from my listening chair. A total of 20 speakers were playing simultaneously for this recordings.

 

Norah Jones' Come away with me.

 

https://youtu.be/i40VgJGjjwU

 

 

And the other two are  2L Lindberg test recordings.

 

https://youtu.be/R8gO5Si3FqI?t=56s

 

https://youtu.be/bN-LjnZuAjE?t=23s

 

Link to comment

Made minor changes by toeing in the speakers inwards and brought down the elevated rear speakers to ear level. I believe that speakers are causing the previous records to sound fous in the middle. But still a $75 binaural microphone cannot capture the exact sound. IMO, the best upgrade was the SIR2 convolution DSP. It reduced the CPU load from 25% to 5%. It sounds better if you use headphones.

 

 

 

Link to comment
1 hour ago, fas42 said:

This is sounding good ... I play your recording, and then the actual source track, available on YouTube - no obvious issues - nice one!

 

Thanks for compliments but the recordings is still not projecting what I wanted to convey.

 

The strangest part of this was we couldn't tell any difference after making the changes but when we compared the recordings the difference was obvious. Looks like our brains are capable of filtering away unwanted sound of the elevated speakers in the earlier setup when listening in the room but could do it when listening with headphones of the binaural recording.

Link to comment

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now



×
×
  • Create New...