[lowdown]

That’s beautiful man. Im in awe with my recordings but I have a feeling im still not at the level you guys are experiencing. I wish I had the knowledge to be able to determine what’s wrong and tweak it to perfection

Any chance of a picture of your mics glued to the foam and maybe how they look placed in your ear?

Here's mine:

 

[Brandon7s]

That’s beautiful man. Im in awe with my recordings but I have a feeling im still not at the level you guys are experiencing. I wish I had the knowledge to be able to determine what’s wrong and tweak it to perfection

Any chance of a picture of your mics glued to the foam and maybe how they look placed in your ear?

Sure! Here's a couple pictures of how I'm wearing the mics, along with a standalone shot of the mics so you can see how much foam I'm using. I've experimented with cutting the foam down to different lengths and as long as they are short enough for me to get an insertion that puts the mic capsules right up next to the entrance of the ear canal, I get good results.

Slightly different angle:

And here's the naked shot. These are the Master Series binaural mics from The Sound Professionals. I also have the non-master series version and I completely removed the silicon sleeve that encompasses the mic capsules and this lets me get an even deeper fit, but I don't think they actually produce better results. I guess at a certain point there's no advantage to measuring with the mic further in, and the tonality gets progressively more bright the further in the insertion, from my experience, so this placement in the photo is the most optimal that I've found so far. Also, trimming the foam TOO short makes it difficult to keep the mics facing the right direction due to lack of support from the foam.

The other thing I recommend you use if you're not already doing so is room correction with the Harman-in-room-loudspeaker-target.csv curve. It might SEEM like the bass is overwhelming but it made a BIG difference in getting tonality that was realistic. It was off-putting at first but after changing my mentality from approaching Impulcifer as a SPEAKER emulation rather than a ROOM emulation, something "clicked" and now it makes perfect sense. Such a bassy curve more accurately represents the sound of a speakers that are LOUD in my room, as I would typically get when I'm playing a guitar amp in my room or listening to a movie in a theater, or even seeing a band live. That kind of bass response is not something I get to enjoy on a regular basis and it took me some time to get acclimated to getting THAT kind of sound from headphones.

Overall tonality and frequency response of the BRIR you take makes a HUUUGE difference as to how convincing it will be. When I go back to listen to my older BRIRs the immersion breaking factor is more often than not the thin sound with little to no low-end support. Localization is still good on most of them but without that low end support they don't truly convince my brain that I'm listening to decently loud speakers. Basically, the bass isn't meant to be heard so much as it is meant to be felt, and that made a huge difference to how realistic the speaker virtualization appeared to me. If you can't "feel" the bass from your headphones then you are probably not getting the most out of Impulcifer.

Also, what headphones are you using? I think that can have a dramatic effect on how convincing your results are. My Ananda's are FAR, FAAAR more convincing than my DT1990's, ATH-R70X, HD6XX, or DT770s. Now that I've got a pretty good handle on getting great results out of my Anandas, I'm going to try working on getting the same results out of my other headphones. So far the BRIRs I've made with them have been quite lackluster so I have a lot more experimenting to do.

Reverb management is also a really good thing to try out if you haven't yet. I like it because it lets me set a short reverb time and lets me use my own software plugins like Valhalla's Room and Liquidsonic's Reverberate 3, among others, to create the space that I want to experience. It's wild just how well this works. I can set up a room that is any size and I while I'm changing parameters I can hear the changes in real time and can easily match the reverb settings to create a room of similar size to the one I'm actually in, just waaay better sounding - like, the best-treated room of all time.

I'm still playing around a lot with the channel balance options but I've not settled on any particular setting as being optimal. Sometimes Trend works great and a lot of times it just messed up my bass perception horribly. I still get great results without using any channel balance options, but they just aren't perfect, so more experimenting to be done there. I think manually adjusting left/right balance is probably a better option for me.

EDIT: one more thing that helped me appreciate Impulcifer's accuracy: realize that a truly good BRIR won't make it sound like your speakers are the sound source. They'll make it sound like the PHANTOM CENTER is the sound source. It's a subtle difference but it It's important. Sit down with your speakers and listen to them as normal. Get a feel for the phantom center and really focus on it. Pay special attention to the low end and how it seems to just appear out of nowhere and everywhere all at once. It won't SOUND like the bass is coming from your speakers even though it is.

Next, put on your headphones and your favorite BRIR. Instead of trying to determine if your brain is fooled into thinking that the sound is coming from the speakers, focus on that phantom center again. Use that to gage how successful your Impulcifer results are rather than a general feeling of whether it sounds like your speakers are the source of audio.

 

[morgin]

@lowdown These look perfect I’m going to try using the foam plugs like you have

 

[morgin]

Sure! Here's a couple pictures of how I'm wearing the mics, along with a standalone shot of the mics so you can see how much foam I'm using. I've experimented with cutting the foam down to different lengths and as long as they are short enough for me to get an insertion that puts the mic capsules right up next to the entrance of the ear canal, I get good results.


Slightly different angle:



And here's the naked shot. These are the Master Series binaural mics from The Sound Professionals. I also have the non-master series version and I completely removed the silicon sleeve that encompasses the mic capsules and this lets me get an even deeper fit, but I don't think they actually produce better results. I guess at a certain point there's no advantage to measuring with the mic further in, and the tonality gets progressively more bright the further in the insertion, from my experience, so this placement in the photo is the most optimal that I've found so far. Also, trimming the foam TOO short makes it difficult to keep the mics facing the right direction due to lack of support from the foam.


The other thing I recommend you use if you're not already doing so is room correction with the Harman-in-room-loudspeaker-target.csv curve. It might SEEM like the bass is overwhelming but it made a BIG difference in getting tonality that was realistic. It was off-putting at first but after changing my mentality from approaching Impulcifer as a ROOM emulation rather than SPEAKER emulation, something "clicked" and now it makes perfect sense. Such a bassy curve more accurately represents the sound of a speakers that are LOUD in my room, as I would typically get when I'm playing a guitar amp in my room or listening to a movie in a theater, or even seeing a band live. That kind of bass response is not something I get to enjoy on a regular basis and it took me some time to get acclimated to getting THAT kind of sound from headphones.

Overall tonality and frequency response of the BRIR you take makes a HUUUGE difference as to how convincing it will be. When I go back to listen to my older BRIRs the immersion breaking factor is more often than not the thin sound with little to no low-end support. Localization is still good on most of them but without that low end support they don't truly convince my brain that I'm listening to decently loud speakers. Basically, the bass isn't meant to be heard so much as it is meant to be felt, and that made a huge difference to how realistic the speaker virtualization appeared to me. If you can't "feel" the bass from your headphones then you are probably not getting the most out of Impulcifer.

Also, what headphones are you using? I think that can have a dramatic effect on how convincing your results are. My Ananda's are FAR, FAAAR more convincing than my DT1990's, ATH-R70X, HD6XX, or DT770s. Now that I've got a pretty good handle on getting great results out of my Anandas, I'm going to try working on getting the same results out of my other headphones. So far the BRIRs I've made with them have been quite lackluster so I have a lot more experimenting to do.

Reverb management is also a really good thing to try out if you haven't yet. I like it because it lets me set a short reverb time and lets me use my own software plugins like Valhalla's Room and Liquidsonic's Reverberate 3, among others, to create the space that I want to experience. It's wild just how well this works. I can set up a room that is any size and I while I'm changing parameters I can hear the changes in real time and can easily match the reverb settings to create a room of similar size to the one I'm actually in, just waaay better sounding - like, the best-treated room of all time.

I'm still playing around a lot with the channel balance options but I've not settled on any particular setting as being optimal. Sometimes Trend works great and a lot of times it just messed up my bass perception horribly. I still get great results without using any channel balance options, but they just aren't perfect, so more experimenting to be done there. I think manually adjusting left/right balance is probably a better option for me.

EDIT: one more thing that helped me appreciate Impulcifer's accuracy: realize that a truly good BRIR won't make it sound like your speakers are the sound source. They'll make it sound like the PHANTOM CENTER is the sound source. It's a subtle difference but it It's important. Sit down with your speakers and listen to them as normal. Get a feel for the phantom center and really focus on it. Pay special attention to the low end and how it seems to just appear out of nowhere and everywhere all at once. It won't SOUND like the bass is coming from your speakers even though it is.

Next, put on your headphones and your favorite BRIR. Instead of trying to determine if your brain is fooled into thinking that the sound is coming from the speakers, focus on that phantom center again. Use that to gage how successful your Impulcifer results are rather than a general feeling of whether it sounds like your speakers are the source of audio.

Thanks for the detailed response it’s a big help and a lot of info that I would have not even thought about. I will definitely try your suggestions. Appreciate it so much

 

[lowdown]

@lowdown These look perfect I’m going to try using the foam plugs like you have

The foam on mine could be a bit shorter, it's a matter of finding the balance between long enough to stay in place but short enough to let the mic seat close to the canal opening. I'm used to wearing foam earplugs so inserting these far enough wasn't a problem.

 

[morgin]

Hi all, just wanted to update anyone having issues with surround sound movies. I've been using MPC-HC for my media playback and had it set up for 7.1 sound output and it performed well, I've also setup and used VLC media player too to compare. I was under the impression that all the players would output 7.1 sound the same as long as they are setup correctly. Last night I decided to try potplayer and configured it to play 7.1 movies, and I have to post that potplayer has made a huge difference in the sounds I'm hearing. So if you use impulcifer for media playback please definitley give potplayer a try

 

[musicreo]

Last night I decided to try potplayer and configured it to play 7.1 movies, and I have to post that potplayer has made a huge difference in the sounds I'm hearing. So if you use impulcifer for media playback please definitley give potplayer a try

You should check the setting of potplayer. I guess it doing some normalisation.

 

[castleofargh]

Hi all, just wanted to update anyone having issues with surround sound movies. I've been using MPC-HC for my media playback and had it set up for 7.1 sound output and it performed well, I've also setup and used VLC media player too to compare. I was under the impression that all the players would output 7.1 sound the same as long as they are setup correctly. Last night I decided to try potplayer and configured it to play 7.1 movies, and I have to post that potplayer has made a huge difference in the sounds I'm hearing. So if you use impulcifer for media playback please definitley give potplayer a try

potplayer is still my favorite player, and beside some very rare artifacts(for a fraction of a second and then all is well for a while, like a buffer issue of sort could do) on maybe 2 DVDs, and one surround demo I found online, this player has been good to me. But IMO if you do get a clear sound difference, it's probably because some setting is causing it, and given all you can fool around with in potplayer, I would bet on it being the odd one.

 

[morgin]

I was also thinking someone who’s had good results should maybe make a YouTube video tutorial. Some of the steps are daunting for newbies like myself and a lot more people would probably try impulcifer if there was a visual guide to follow. Coz this software is too good for just a few to have.

 

[Brandon7s]

Okay, progress update for those who might be interested. I just finished my 60th measurement session and this new BRIR is absolutely BONKERS! I just tried A/Bing it with my Ananda vs. my speakers and I can't tell them apart anymore. This is the first time I've gotten close enough to the real-deal with Impulcifer that I simply cannot tell if audio is coming from my speakers or from my headphones, once volume matched. My last couple of measurements were VERY good, or at least I THOUGHT they were, but this is on a whole different level. You'll probably see why once you see the charts.

I've been having a heck of a tough time getting good channel balance. It's always off somewhere, either the low end or the high end, and getting it nailed perfectly has alluded me until this session. The reason why is that it appears my right and left ears are RADICALLY different, acoustically. How do I know? Take a look at this Headphones plot:

For anyone with relatively balanced left/right hearing, this would sound horrifically out of balance, since there's nothing even close to balanced here! The only point in which my ears hear the same frequencies at roughly equal volume is that small section from 700 to 1,000 hertz, and that's all. But the sound is perfect!

I tried using the average, mids, and trend --channel_balance options and all of them made the BRIR practically unusable. But with no channel balance at all? It's incredible. This explains so much about my experience with headphones in general. I don't get phantom centers with any headphones except for IEMs, and even with those it's weak and ill defined. Most of my headphones sound roughly the same to me when I'm not using an Impulcifer BRIR; however, with the BRIR it's completely different. I can hear all of the headphone's characteristics, just as if I were listening to different speakers.

Now take a look at this corresponding Results plot:

It's nearly dead-on in balance at every frequency when compared to past results. This is the first time I've gotten anything even close to this tight of a balance match.

Here's an example from the prior 2 measurements I've made, both are my previous favorites that I created earlier this week. Here you can see the whacky channel balance plot, which isn't TOO far off of my most recent measurement in the low end but everything passed 1,000hz is completely different and much higher in amplitude compared to my new, perfect measurement:

Results from the same session: pretty tight low end balance and then it gets crazy at just about 100 hz. This is why I could never get the high and mid frequencies to be properly aligned. No --channel_balance setting can fix this wide comb-like pattern:

Here's the second set of measurements from earlier in the week. This sounds a bit brighter but still pretty good sounding. Once again though, it has balance issues between the lows and highs and no --channel_balance options could fix that:

And the results from that session: by the way, the balance issues in the bass regions are HIGHLY apparent, causing the low-end to move left to right depending on the frequency being played. It's super annoying to try to monitor bass guitar with this happening, as you can imagine.

So, what did I do to make this massive improvement from today's measurements? I placed the mics further into my ears than normal, that's all. It looks like there's a critical point at which the highs start to get attenuated to the levels that my eardrum receives, and if I don't push the mics in far enough, I don't get the effect of that attenuation.

I think that trying to judge channel balance by the headphones.png plots is NOT useful in determining the quality of the final results, and I had spent quite a lot of time trying to figure out how to "fix" that problem... but it was never actually a problem! The problem was channel balance versus the speakers, as shown in the results.png plots. THAT is what you need to try to perfect in order to get a truly amazing quality measurement, I believe.

 

[morgin]

Keep the details incoming. They are a big help. I’m planning on another recording session and want to try all your suggestions because although localisation is very good it still doesn’t feel like I’m listening to speakers in my room so I know I’m nowhere near what I can get.

How far into your ears did you set your mics?

And are you suggesting no balance trend

Also with room correction are you taking readings for both ears in every position with your mic? and if so what angle are you positioning the mic (toward the speaker/facing toward the ceiling?)

I’m thinking for room correction, if I can strap the mic to my head and as close to the ear whilst sitting in the exact same position after headphone and ear measurements without getting off my swivel chair I’ll get a good room recording. But the mic will be facing the ceiling. Then I can do the same for the other ear.

Something like this.

 

[sander99]

I’m thinking for room correction, if I can strap the mic to my head and as close to the ear whilst sitting in the exact same position after headphone and ear measurements without getting off my swivel chair I’ll get a good room recording. But the mic will be facing the ceiling. Then I can do the same for the other ear.

I don't think that is a good idea, because this way the sound you measure is influenced by the presence of your head, in a way the room sound has gone through a part of your HRTF this way. The sound from the right surround speaker at your left microphone for example will be strongly influenced by the head that is full in the path of the sound, the sound has to bend around your head and will be severely filtered. If you compensate for this filtering you will be compensating for a part of your HRTF and thus undoing a part of the HRTF filtering that the speaker virtualisation is doing!

 

[Brandon7s]

Keep the details incoming. They are a big help. I’m planning on another recording session and want to try all your suggestions because although localisation is very good it still doesn’t feel like I’m listening to speakers in my room so I know I’m nowhere near what I can get.

How far into your ears did you set your mics?

And are you suggesting no balance trend

Also with room correction are you taking readings for both ears in every position with your mic? and if so what angle are you positioning the mic (toward the speaker/facing toward the ceiling?)

I’m thinking for room correction, if I can strap the mic to my head and as close to the ear whilst sitting in the exact same position after headphone and ear measurements without getting off my swivel chair I’ll get a good room recording. But the mic will be facing the ceiling. Then I can do the same for the other ear.

Something like this.

Glad to be able to help, hopefully you can get similar results and if there's any way I can help you get there, I'm happy to do it! It sounds like hyperbole but the results I'm getting now are literally life changing. And I mean LITERALLY, not figuratively. As a practically life long guitarists who's been relegated to apartment life for all but a few years, this is HUGE. I can FINALLY enjoy electric guitar at decent volumes and without disturbing my girlfriend or my neighbors. Not only that, it's not a compromise! It's practically indistinguishable from the real deal while I'm in this room.

And outside of the guitar part of this is the, the fact that I can now hear songs as they are meant to be heard is incredible. I love my speakers but I only get the chance to use them at very low levels, far lower than most people use for mixing recordings. So I've always used headphones and I've spent a lot of money on them just trying to get something that sounds even slightly decent. It just turns out that the closest thing I've gotten is so far from "decent" that it's nearly impossible to describe.Then Impulcifer came along and now I want to listen to my entire music collection from start to finish again. I'm hearing things that I couldn't even imagine ever being able to hear, like a pen or guitar pick being dropped on the floor from recordings that I've listened to hundreds of times. It's really hard to understate just how much this has affected my enjoyment of music, both listening to it and playing it. THANK YOU @jaakkopasanen! If there's any way I can donate to you, just let me know and I'll be happy to do it.

Okay, rave over, let me see if I can answer your questions.

Here's a shot of about how deep I think I had it for this last measurement. It's pretty far in, though not further than what I could reach with my pinky finger. That's the front of the mic facing directly at the camera, but as you can see it's obscured by a portion of my ear's geometry. I think how deep you have to get is really just a factor of your personal physiology and ear-shape.

About --channel_balance: I'm not saying that you shouldn't use the channel balance options, I'm saying that when you get a good, proper fit then you won't have to, and that once your measurement is good enough then using the channel balance options just make things worse, because there's nothing to fix. Could you post your headphone.png and results.png from your favorite BRIR that you've made so far? I'd be curious to see if and how they resemble my older, less improved measurements.

For room correction: I'm just using a general room mic placement. I did 5 different measurements with the mic vaguely near where my head would normally be to room.wav and then used this command to process the BRIR with room correction:
python impulcifer.py --test_signal="data/sweep-6.15s-48000Hz-32bit-2.93Hz-24000Hz.pkl" --room_target="data/harman-in-room-loudspeaker-target.csv" --generic_limit=2000 --dir_path="data/my_hrir"

I'm not using specific room measurements. I did try that once but the results weren't any better than I was getting with general measurements. I'd focus on trying to get a very accurate measurement in left/right balance first, and worry about the room EQ afterwards. It can actually make the illusion of listening to your same speakers even weaker, because we're somewhat familiar with how our room sounds and you can recognize that easier if you don't EQ out the imperfections.

If there's anything else you can think of, feel free to ask!

 

[morgin]

I don't think that is a good idea, because this way the sound you measure is influenced by the presence of your head, in a way the room sound has gone through a part of your HRTF this way. The sound from the right surround speaker at your left microphone for example will be strongly influenced by the head that is full in the path of the sound, the sound has to bend around your head and will be severely filtered. If you compensate for this filtering you will be compensating for a part of your HRTF and thus undoing a part of the HRTF filtering that the speaker virtualisation is doing!

This actually makes a lot of sense so when room recording do you think just being present in the room will effect the measurement? Because originally I’m sat down but when doing the room recording I’m in a different place in the room

 

[morgin]

Glad to be able to help, hopefully you can get similar results and if there's any way I can help you get there, I'm happy to do it! It sounds like hyperbole but the results I'm getting now are literally life changing. And I mean LITERALLY, not figuratively. As a practically life long guitarists who's been relegated to apartment life for all but a few years, this is HUGE. I can FINALLY enjoy electric guitar at decent volumes and without disturbing my girlfriend or my neighbors. Not only that, it's not a compromise! It's practically indistinguishable from the real deal while I'm in this room.

And outside of the guitar part of this is the, the fact that I can now hear songs as they are meant to be heard is incredible. I love my speakers but I only get the chance to use them at very low levels, far lower than most people mix recordings at, for sure. So I've always used headphones and I've spent a lot of money on them just trying to get something that sounds even slightly decent, it turns out that the closest thing I've gotten is so far from "decent" that it's impossible to describe. Impulcifer came along and now I want to my entire music collection from start to finish again. I'm hearing things that I couldnt' even imagine ever being able to hear, like a pen or guitar pick being dropped on the floor from recordings that I've listened to hundreds of times. It's really hard to understate just how much this has affected my enjoyment of music, both listening to it and playing it. THANK YOU @jaakkopasanen! If there's any way I can donate to you, just let me know and I'll be happy to do it.

Okay, rave over, let me see if I can answer your questions.

Here's a shot of about how deep I think I had it for this last measurement. It's pretty far in, though not further than what I could reach with my pinky finger. That's the front of the mic facing directly at the camera, but as you can see it's obscured by a portion of my ear's geometry. I think how deep you have to get is really just a factor of your personal physiology and ear-shape.



About --channel_balance: I'm not saying that you shouldn't use the channel balance options, I'm saying that when you get a good, proper fit then you won't have to, and that once your measurement is good enough then using the channel balance options just make things worse, because there's nothing to fix. Could you post your headphone.png and results.png from your favorite BRIR that you've made so far? I'd be curious to see if and how they resemble my older, less improved measurements.

For room correction: I'm just using a general room mic placement. I did 5 different measurements with the mic vaguely near where my head would normally be to room.wav and then used this command to process the BRIR with room correction:
python impulcifer.py --test_signal="data/sweep-6.15s-48000Hz-32bit-2.93Hz-24000Hz.pkl" --room_target="data/harman-in-room-loudspeaker-target.csv" --generic_limit=2000 --dir_path="data/my_hrir"

I'm not using specific room measurements. I did try that once but the results weren't any better than I was getting with general measurements. I'd focus on trying to get a very accurate measurement in left/right balance first, and worry about the room EQ afterwards. It can actually make the illusion of listening to your same speakers even weaker, because we're somewhat familiar with how our room sounds and you can recognize that easier if you don't EQ out the imperfections.

If there's anything else you can think of, feel free to ask!

The actual help and willingness I'm getting from you guys is just beautiful.

here are my png

I'm curious as to the volume you did your recordings at. Its difficult to say I know but
how loud were your speaker when doing the recordings. Just at comfortable listening volume or something loud as to disturb the people in the house

Did you keep the same amount of volume for both headphones and speaker so they sounded the same to your ears, or did you concentrate more on headroom being the same