The making of: The Two Towers (a 25 driver Full Range line array)

Your virtual audition of the towers sounds fantastic. Very nice sound - probably close to perfect representation of the recording.

Well I definitely hear a room, my room! 😱
Like gmad said already, a lot less noticeable in real life. Played back on the arrays you get double the effect of the room. Reminds me of the test that Tom Danly suggests to see how accurate your speakers are. Record them and play back till there's a mess left. How many generations of that can your speakers do....

I bet when I change out my ceilings things will get better.
If Kendricks could try to control the room just a little... those are grandiose (as always) but the reflections in the room kill the experience! If only they would add some bass traps and carpet, that would help.

I'm sure you are having loads of fun playing around and making your arrays shine!

I kinda miss mine, and I might be itching to start a new project... oh no!

I quite like their collection of speakers and enjoy the audition they provide.
Sounds like you're almost ready to do it again, remember all the work involved? 😀


Just to get the virtual audition on the new page:
gmad, Thanks! It is fun! just spent some time listening to my room on my headphones. I'll post a clip when I get new measurements because I changed the downward slope that this measurement has with about 2 dB less downward slope and removed the bump I had in the bass. But it did let me enjoy a comparison to 2 youtube videos I liked quite a while ago.
https://www.youtube.com/watch?v=U1tWyOGaylY
and
https://www.youtube.com/watch?v=T09m0IqSV10
2 very different speakers with very different sound
But I got the track to see if I could get near to the big ones in bass.
That goal was for my Car by the way 😀.

Ah, what the heck, I'll just post this one first, I'll notify when I change the song to the latest spec of my arrays 😀.
www.rsr-concepts.com/diyma/linesample.m4a
Right click download and saveas as I'm not sure my server will stream it, it's a full 6 minute aac file @ 350kb/s
Song convolved with left and right IR response as measured at the listening position. I did throw in standard Crossfeed as defined in JRiver.
 
Last edited:
EQs with nasty compromises...

...What processing do you use on the NSB arrays?

My source is a PC. It uses the older Intel Q9550 processor, SSD drive and AMD HD6970 graphics card, which works great for audio and DVD playback. I use the On-board Realtek Audio for movie playback, as that has 7.1 capability (I use 4.2). I can use the built in digital EQ to adjust the sound, but I found that it alters the bit rate. 😱 For stereo and music listening, I use an M-audio Fast track Pro USB DAC. This has a more resolved and laid back sound. No EQ option on this. I have an old SAE parametric EQ, which is very versatile, but it just kills the 3D sound stage. The digital EQ is far more accurate.

Thankfully, I have a Twisted Pair Buffalo III DAC kit, waiting to be built. I would like to add a miniDSP miniSharc kit and USBStreamer so I have exceptional EQ and FIR filters for cross-over experiments. I have not decided if I am building the Buffalo as a four channel DAC to accommodate my movie playback, or go all out and have one Buffalo kit per stereo pair ($$$$).

Currently, I am powering the Arrays with Peter Daniel's LM3875 kits. They play pretty loud with the arrays, but run out of steam with the Avebury. They are more "open" in the mid-range then the Sure class-D TK2050 kits and Pioneer VSP DSP1 receiver I had. I have an SAE 2100 Power amp that I am getting restored by Ken Ealey as soon as the money gets rolling with my new career. I have heard these are quite good if they are properly restored.

I also have a custom designed 1626 tube pre-amp I need to finish wiring. I will let You know my subjective impressions of the sound, as I know You mentioned in interest in tubes earlier in this thread.

Someone also mentioned something about a Nelson Pass Amp kit of e-bay, I believe an Aleph X or something? Link Please... (I did a search and as most of You already know, e-bay needs to work on their search engine.)

My one example of an NSB driver provided by Godzilla for the Subjective Blind Comparo thread is surprisingly good sounding. I can imagine a line array would be quite impressive. They are not as flat in response as the Vifa but nothing a little mild EQ can't take care of to remove that resonance peak.

I need to revisit that thread. I got lost in the chit chat on that one. But I can tell You, the NSB arrays are surprisingly impressive sounding, esp for the money spent. I have been able to EQ out the resonance peak running the arrays through the Realtek DAC. Unfortunately, I discovered that due to the lowered bit-rate the Realtek driver imposes when engaging the EQ, the sound takes on a "graininess" and almost a ringing characteristic. The Frequency Response was amazing, but the fatigue stripped the joy out of listening. So it ends up being more listenable long term with the resonance peak. :spin:

This has got my thinking that the audio chips in my computer are holding me back.

What does the REW software process through? The DAC or CPU?

Either way, I think the Buffalo III and miniSharc combo will work better for me.

Maybe I can get more out of my NSB "Cheap and Cheerful" arrays yet. The NSBs do not have the intimate detail or downward dynamics of the Alpair 12, but You guys have encouraged me to build a nicer cabinet for them and keep them for surround speakers when I build my "dream" crazy expensive arrays for the front.

Thanks for your interest,

Allen
 
I'm actually running an older Q9300 myself. Dual SSD in mirror for O.S. Bigger WD Black disks for movies and audio storage. I don't use disks, playback is from harddisk. Nvidia Quadro K2000 graphics card for DVD/Blu-ray playback. I have a realtec onboard but do not use it. I have an add on Asus Xonar Essence XT soundcard, I have the extra card to make it 7.1 but not installed yet. So quite similar to your tools there. I had read the Realtec wasn't up to par so that's why I have the Asus card in there. I actually use a Music Fidelity M1 DAC connected to the Asus S/PDIF. Used the asynchronous USB first but that had some troublesome PC power supply noise as did the Asus by itself.
I asked because I think with the current tools available you should be able to get more out of the NSB arrays.
I choose JRiver for playback (I'm on Windows 7 Pro) and use the build in PEQ and it's convolver engine.

REW is only used as a measurement suite. Due to having 2 audio cards, the Asus and the M1 DAC I can loop the sound from REQ trough JRiver and measure/record and confirm the results of my convolved signal. I bet you could do the same and improve the sound that way.
A Measurement microphone is needed obviously, the rest is just a bit of work setting it up.
But miniSharc should work fine too of coarse. You'd need the same microphone to setup miniSharc though. After that it's stand alone.
JRiver uses the CPU for audio tweaks but movie play back (Blu-Ray) far outweighs the strain on my PC vs my audio chain even though a lot is done on the graphic card. JRiver has madVR available for movie playback, an excellent and tunable solution for the picture part for HT. But there are other free tools available that could do the same if bundled right. I like the ease of use of JRiver as it functions as the center of all my Audio/Video needs and accepts additional plugins should I need them. Here's the CPU strain while playing back a stereo track:
cpu.jpg

(captured at a peak, no more than 6% CPU needed (peak) for EQ plus convolution of the audio signal, average is even lower)

The Buffalo III is very interesting though. With an 8 year old kid more and more movies are being played so I had to work on that as well. I had high hopes the Asus could fill all my needs but kept getting PSU noise. I could get rid of it with a ground loop isolator only to find out my boost levels saturated the isolator creating higher distortion levels. I didn't really hear it (only happening at low frequencies) but sometimes knowing it's there is enough to move in another direction 🙂.
 
Last edited:
I have been using Foobar for audio and Media Player Classic for movies. So needless to say I was pretty excited to try JRiver, which could do both:

I downloaded JRiver to check it out. I like the audio features it has. I am having troubles playing certain DVDs with it. It gets stuck in an "analyzing optical disk" window, and the software is unresponsive until I eject the disk. I am looking into how to fix this, but it seems the trouble shooting guide is outdated and does not apply to the version I downloaded. I got my Realtek to work in WASAPI with it, but my M-Audio DAC will not work in ASIO or WASAPI. This is an M-Audio problem, not JRiver. I could not get the M-Audio to work right with Foobar either. So needless to say, I can only enjoy WASAPI output with the Realtek. At least I know it is an M-Audio problem for sure now. This is the all more reason to build the Buffalo.

I do have a Dayton Audio measurement mic and Holm Impulse and REAL measurement software. I have not used REAL much as I have the entry level 1/3 octave version and the higher version upgrades are very costly.

I need to regroup on all of this, as it has been a little discouraging. JRiver seems like a nice piece of software, and of the little bit I heard and saw with the one DVD that did work, the quality is good. I like that You can import FIR filters. The EQ and PEQ seem good too. Blu-ray playback is a plus IF it works. :djinn:

I will keep plugging away at this, and let You know if I get anywhere. Got to stop now and eat dinner so I am ready for my 10 hour night at work. :moon:

I hope I am not hijacking this thread!

Allen
 
Too bad your M-Audio isn't working right. You might want to check either with M-Audio support or forums, or pro-audio dedicated sites like Gearslutz.com.
I have a Presonus and, although I've had all sort of problems with their drivers, it seems to work well with JRiver. I have 8 outputs, so I can separate my surround channels.

JRiver also has the possibility to add VST plugins, adding even more to the list of audio "enhancements".

Give REW a try. It is free, and a pretty powerful measurement software.
 
OT

My source is a PC. It uses the older Intel Q9550 processor, SSD drive and AMD HD6970 graphics card, which works great for audio and DVD playback. I use the On-board Realtek Audio for movie playback, as that has 7.1 capability (I use 4.2). I can use the built in digital EQ to adjust the sound, but I found that it alters the bit rate. 😱 For stereo and music listening, I use an M-audio Fast track Pro USB DAC. This has a more resolved and laid back sound. No EQ option on this. I have an old SAE parametric EQ, which is very versatile, but it just kills the 3D sound stage. The digital EQ is far more accurate.

Thankfully, I have a Twisted Pair Buffalo III DAC kit, waiting to be built. I would like to add a miniDSP miniSharc kit and USBStreamer so I have exceptional EQ and FIR filters for cross-over experiments. I have not decided if I am building the Buffalo as a four channel DAC to accommodate my movie playback, or go all out and have one Buffalo kit per stereo pair ($$$$).

Currently, I am powering the Arrays with Peter Daniel's LM3875 kits. They play pretty loud with the arrays, but run out of steam with the Avebury. They are more "open" in the mid-range then the Sure class-D TK2050 kits and Pioneer VSP DSP1 receiver I had. I have an SAE 2100 Power amp that I am getting restored by Ken Ealey as soon as the money gets rolling with my new career. I have heard these are quite good if they are properly restored.

I also have a custom designed 1626 tube pre-amp I need to finish wiring. I will let You know my subjective impressions of the sound, as I know You mentioned in interest in tubes earlier in this thread.

Someone also mentioned something about a Nelson Pass Amp kit of e-bay, I believe an Aleph X or something? Link Please... (I did a search and as most of You already know, e-bay needs to work on their search engine.)



I need to revisit that thread. I got lost in the chit chat on that one. But I can tell You, the NSB arrays are surprisingly impressive sounding, esp for the money spent. I have been able to EQ out the resonance peak running the arrays through the Realtek DAC. Unfortunately, I discovered that due to the lowered bit-rate the Realtek driver imposes when engaging the EQ, the sound takes on a "graininess" and almost a ringing characteristic. The Frequency Response was amazing, but the fatigue stripped the joy out of listening. So it ends up being more listenable long term with the resonance peak. :spin:

This has got my thinking that the audio chips in my computer are holding me back.

What does the REW software process through? The DAC or CPU?

Either way, I think the Buffalo III and miniSharc combo will work better for me.

Maybe I can get more out of my NSB "Cheap and Cheerful" arrays yet. The NSBs do not have the intimate detail or downward dynamics of the Alpair 12, but You guys have encouraged me to build a nicer cabinet for them and keep them for surround speakers when I build my "dream" crazy expensive arrays for the front.

Thanks for your interest,

Allen

I would recommend to everybody who have linesource speakers like Wesayso or alike to try the My_ref Fremen Edition amplifier.
It has taken over here as it was (by a pretty wide margin ) better than my previous amplifiers(wich I still have in house) : Goldmund telos 350, Pass different designs, Roenders FC-100, Swordfishy/ASPEN Fetzilla, 6c33 OTL tube amp and others...
Also try a cheap version of a TDA 154X NOS DAC to get a taste of what a DAC with no digital filters can do.
To be able to get the best out of that technique, the cheapo stuff will not do, but you get a sense of the differece from the other soultions (1-bit, 2r2 ladder , sabre aso)

Sorry for out of topic comment, but I couldn´t resist...😀
Koldby
 
Audition of X's Ref speaker with same song Kenrick Audio uses

gmad, Thanks! It is fun! just spent some time listening to my room on my headphones. I'll post a clip when I get new measurements because I changed the downward slope that this measurement has with about 2 dB less downward slope and removed the bump I had in the bass. But it did let me enjoy a comparison to 2 youtube videos I liked quite a while ago.
https://www.youtube.com/watch?v=U1tWyOGaylY
and
https://www.youtube.com/watch?v=T09m0IqSV10
2 very different speakers with very different sound
But I got the track to see if I could get near to the big ones in bass.
That goal was for my Car by the way 😀.

Ah, what the heck, I'll just post this one first, I'll notify when I change the song to the latest spec of my arrays 😀.
www.rsr-concepts.com/diyma/linesample.m4a
Right click download and saveas as I'm not sure my server will stream it, it's a full 6 minute aac file @ 350kb/s
Song convolved with left and right IR response as measured at the listening position. I did throw in standard Crossfeed as defined in JRiver.

Ok, here is the same song (at least first 45 seconds) played on my reference system and recorded at 96kHz 24bit and compressed to 320kbit MP3. Keep in mind that this is a stereo recording of the right channel in mono. The recorder is a Zoom H4 set at mid-gain. Actual sound levels are in the 95dB range, quite loud. Change the extension from .ASC to .MP3 in order to listen. Let me know what you think. I think I need to build the matching stereo pair. 😀
 

Attachments

I have been using Foobar for audio and Media Player Classic for movies. So needless to say I was pretty excited to try JRiver, which could do both:

I downloaded JRiver to check it out. I like the audio features it has. I am having troubles playing certain DVDs with it. It gets stuck in an "analyzing optical disk" window, and the software is unresponsive until I eject the disk. I am looking into how to fix this, but it seems the trouble shooting guide is outdated and does not apply to the version I downloaded. I got my Realtek to work in WASAPI with it, but my M-Audio DAC will not work in ASIO or WASAPI. This is an M-Audio problem, not JRiver. I could not get the M-Audio to work right with Foobar either. So needless to say, I can only enjoy WASAPI output with the Realtek. At least I know it is an M-Audio problem for sure now. This is the all more reason to build the Buffalo.

I do have a Dayton Audio measurement mic and Holm Impulse and REAL measurement software. I have not used REAL much as I have the entry level 1/3 octave version and the higher version upgrades are very costly.

I need to regroup on all of this, as it has been a little discouraging. JRiver seems like a nice piece of software, and of the little bit I heard and saw with the one DVD that did work, the quality is good. I like that You can import FIR filters. The EQ and PEQ seem good too. Blu-ray playback is a plus IF it works. :djinn:

I will keep plugging away at this, and let You know if I get anywhere. Got to stop now and eat dinner so I am ready for my 10 hour night at work. :moon:

I hope I am not hijacking this thread!

Allen

I'm a bit disappointed to read JRiver isn't working as it should for you yet. I'd recommend spending some time on it though, my setup needed some work too with lots of visits to the JRiver forums looking for answers.
Though I don't play from disks, I store them before playing using the MKV container with programs like MakeMKV.
The DAC not working is a bigger problem, have you tried using Asio4all?

On video playback, did you use the Standard Red October or the more advanced Red Oktober HQ? That last one is MadVR, more CPU intensive but a mayor upgrade. It took me a while to set it up mainly due to system limits on my part. But the difference is night and day. It gave me a very beautiful picture quality with lots of settings to tweak. Very rewarding on my (relatively cheap) Panasonic Plasma.

I might have an opportunity to upgrade my PC soon. Looking forward to that, going to a Xeon E5 1620v2 processor and the PCI-e x16 2.0 spec that's recommended for my card.

I would recommend to everybody who have linesource speakers like Wesayso or alike to try the My_ref Fremen Edition amplifier.
It has taken over here as it was (by a pretty wide margin ) better than my previous amplifiers(wich I still have in house) : Goldmund telos 350, Pass different designs, Roenders FC-100, Swordfishy/ASPEN Fetzilla, 6c33 OTL tube amp and others...
Also try a cheap version of a TDA 154X NOS DAC to get a taste of what a DAC with no digital filters can do.
To be able to get the best out of that technique, the cheapo stuff will not do, but you get a sense of the differece from the other soultions (1-bit, 2r2 ladder , sabre aso)

Sorry for out of topic comment, but I couldn´t resist...😀
Koldby

I don't think I'll be able to DIY an amp, I can do lots of things, but soldering was never one of my talents (lol).
I'm still not sure what the best solution for a DAC is. My F1 is an up sampling DAC, you're recommending a Non Oversampling Dac...

By the way, while you're here, what's the window you use at 20 kHz? How many samples? The loner I make that window, the more focus I get. Yet it becomes a bit unnatural at some point. Not sure where the limit is.
 
Last edited:
Ok, here is the same song (at least first 45 seconds) played on my reference system and recorded at 96kHz 24bit and compressed to 320kbit MP3. Keep in mind that this is a stereo recording of the right channel in mono. The recorder is a Zoom H4 set at mid-gain. Actual sound levels are in the 95dB range, quite loud. Change the extension from .ASC to .MP3 in order to listen. Let me know what you think. I think I need to build the matching stereo pair. 😀

Glad you're joining in on the fun X! You should make it a stereo set.
It sounds a bit crisp on my lines, will try headphones next. Might just be living with a "warm" speaker setup for 20+ years on my part 😀.
 
I would recommend to everybody who have linesource speakers like Wesayso or alike to try the My_ref Fremen Edition amplifier.

I have not heard of that one, I will look into it...

Also try a cheap version of a TDA 154X NOS DAC to get a taste of what a DAC with no digital filters can do.
To be able to get the best out of that technique, the cheapo stuff will not do, but you get a sense of the differece from the other soultions (1-bit, 2r2 ladder , sabre aso)

Koldby

I am assuming by "cheapo stuff" You mean the Realtek and M-Audio, in which case I fully agree. The Twisted Pair Buffalo III DAC kit I have to assemble is supposed to be pretty good even though it oversamples. What are your thoughts on the miniDSP stuff? Mainly the miniSharc and USB streamer. I have not bought those yet, so it would be no loss to me if You feel they would not work for array applications. I eventually want to be able to stream a 4 channel surround system out of my computer. Eight channels might be better, allowing subs and maybe multi-way capacity for future speaker experiments - I have not decided if I want to do this with one Buffalo DAC configured as eight channel or 4 Buffalo (or maybe others) DACs configured as 4 stereo units.

I could assemble the Buffalo as a stereo DAC and EQ my arrays in JRiver like wesayso is doing. (Or is this what You are suggesting not to do.) I am wide open to any ideas You have.

Thanks for your help,

Allen
 
gmad, Thanks! It is fun! just spent some time listening to my room on my headphones. I'll post a clip when I get new measurements because I changed the downward slope that this measurement has with about 2 dB less downward slope and removed the bump I had in the bass. But it did let me enjoy a comparison to 2 youtube videos I liked quite a while ago.
https://www.youtube.com/watch?v=U1tWyOGaylY
and
https://www.youtube.com/watch?v=T09m0IqSV10
2 very different speakers with very different sound
But I got the track to see if I could get near to the big ones in bass.
That goal was for my Car by the way 😀.

Ah, what the heck, I'll just post this one first, I'll notify when I change the song to the latest spec of my arrays 😀.
www.rsr-concepts.com/diyma/linesample.m4a
Right click download and saveas as I'm not sure my server will stream it, it's a full 6 minute aac file @ 350kb/s
Song convolved with left and right IR response as measured at the listening position. I did throw in standard Crossfeed as defined in JRiver.


After listening the convolved track i'm positive as the two reviewers below and give 6 stars out of 6 for sound and 6 stars for beautiful design and look.

Hope others agree here is my more detailed sound review comparing the convolved track to the unprocessed original track.
A little harder edge on sound picture and the room is only difference and trying concentrate listening to the frequency response it sounds just as the real track does this is fantastic reproduction, missing only a little on very high frequency. Given it has a little harder edge than the original unprocessed track i understand you prefer FIR for the high pass filter verse IIR in that FIR normal soften edge on sound stream compared IIR that can make edge even harder.

Because we have seen many pictures of your living room i actually feel i just been there 🙂 it was a pleasure great thanks sharing.


Wow, that sounds great! Way better than either of the youtube videos. The bass from the towers is amazing. This is fun!

.....Your virtual audition of the towers sounds fantastic. Very nice sound - probably close to perfect representation of the recording.
 
Last edited:
BYRTT, I haven't actually listened to the reference track, but in my experience, I've found that the room can impart a certain "hardness" on the sound that is more easily noticed when listening to a microphone pickup. I don't think that filter type is necessarily a factor in this case. By the way, I'll be providing a virtual audition soon of my room in which I think the room sound is more noticeable than wesayso's (a little bit more "hardness" if you will), yet still not very noticeable in real life.
 
Thank you, BYRRT for that very kind review, but you leave me no room for improvement 😱.
After all I'm still working on the eyebrows...
zappa.jpg


I notice the hardness you speak about when listening with headphones. After hearing it played back in my room (twice the room effects) it's hard to unlearn hearing my room while playing normal music. After all, for the most part we hear what we want to hear or what we are most focused on.

Looking forward to hearing your room gmad, also curious as to what music you will choose. I'll upload a different song next time, one that I actually really like.
Might be an all instrumental piece, maybe a track from Rodrigo y Gabriella as I've been playing them a lot lately.
 
Last edited:
I have not heard of that one, I will look into it...



I am assuming by "cheapo stuff" You mean the Realtek and M-Audio, in which case I fully agree. The Twisted Pair Buffalo III DAC kit I have to assemble is supposed to be pretty good even though it oversamples. What are your thoughts on the miniDSP stuff? Mainly the miniSharc and USB streamer. I have not bought those yet, so it would be no loss to me if You feel they would not work for array applications. I eventually want to be able to stream a 4 channel surround system out of my computer. Eight channels might be better, allowing subs and maybe multi-way capacity for future speaker experiments - I have not decided if I want to do this with one Buffalo DAC configured as eight channel or 4 Buffalo (or maybe others) DACs configured as 4 stereo units.

I could assemble the Buffalo as a stereo DAC and EQ my arrays in JRiver like wesayso is doing. (Or is this what You are suggesting not to do.) I am wide open to any ideas You have.

Thanks for your help,

Allen
No I mean cheapo non oversampling like Muse TDA1543 or similar from e-bay.
It is a quite different technologi than m-audio and realtech. They use conventional dacs with digital filters.
Dacs with no digital filters are more analog like and to my ears (among a lot of others) looses the kind of synthetic sound so related to digital. You face a lot of other challenges to approch the absolute(live music) but you should at least hear it for yourself so you can make your desicion.
Koldby
 
I'm still not sure what the best solution for a DAC is. My F1 is an up sampling DAC, you're recommending a Non Oversampling Dac...

By the way, while you're here, what's the window you use at 20 kHz? How many samples? The loner I make that window, the more focus I get. Yet it becomes a bit unnatural at some point. Not sure where the limit is.

I don´t really know what you mean by a window in a non oversampling no digital filter DAC as there are non (re the term no filter😀).
Sometimes I use resampling in Foobar, wich is a digital filter too (has to be), but the algorithm for resampling are very different from brick wall digital anti aliasing filters and hasen´t the same sound signature.

But as I said--- Try it it is cheap to do so and you get a feeling for what it can do to digital sound. If you don´t like it stay with Buffalo and the like (very good DACS) but you mss out an opertunity if you don´t.

IMHO😱
 
Hi koldby, I meant the window used in Audiolense to correct the response. I believe in Audiolense it would be referred to as the amount of Cycles. It can be seen in the Correction Procedure Designer. 5 cycles at 22.050 Hz would be 0.227 ms for instance.
I was wondering how you set that window.

I still want to try and upsample in JRiver before it goes to the DAC and have my FIR correction at say 96 KHz.
 
Glad you're joining in on the fun X! You should make it a stereo set.
It sounds a bit crisp on my lines, will try headphones next. Might just be living with a "warm" speaker setup for 20+ years on my part 😀.


Here is the freq response/phase and harmonic distortion of my reference speaker along with the IR. Perhaps it is the massive single overshoot on the impulse that makes it "harsh". The frequency response is relatively flat +/- 2.5db and not tilted down as HF's go up like how have yours. Perhaps that is what the "brightness" is from? The IR is typically reversed because I think the pre-amp in my mic produces a reversed signal? I have noticed it on all drivers even simple full range at near field (yes I checked polarity of driver with battery).
 

Attachments

  • Ref-phase.PNG
    Ref-phase.PNG
    92.3 KB · Views: 560
  • Ref-HD.PNG
    Ref-HD.PNG
    162.1 KB · Views: 555
  • Ref-IR.PNG
    Ref-IR.PNG
    45.1 KB · Views: 536