111 postsPage 3 of 6
1, 2, 3, 4, 5, 6
j0001s wrote
Nip wroteI did two bars as untouched, and two bars fading to infinite automation.
Peak level -6 dBFs, 1 KHz signal 48k....
And even doing manual fade when recording to second track, nothing by automation, it is related to algo doing volume it seems.


Did you disable the Studio One dither before you did the test? It is on by default, and it must be off for a legit comparison.


I don't think dither is applied to 32 bit fp files bounced internally.

It is also not causing intermodulation distortion, only random (preferrably uncorrelated stereo) noise at the least significant bit, as you mentioned in a previous post. And dither noise does not look like in those images. And it's at a much lower level. So I don't think dither is the issue here.
User avatar
by j0001s on Fri Mar 15, 2019 11:01 pm
Couldn't help myself - though I really shouldn't spend time on this now (major business commitment to attend to), I had to get my hands dirty.

Before I write more, I should point out that I have a background in this stuff. So, I ran a bunch of tests to try to understand what was really going on. And, I have to say - they are inconclusive.

I've attached a screen snap. Looks awful, right? It's showing artefacts about -70 db.
Artefacts.png

That's the original sine wave, with certain display settings. Should I conclude that the original sine wave has intermodulation issues with, I don't know what? Not buying that.

I did a bunch of tests, including:

* Turning off dither (minimal effect - I would agree that's not what we're seeing)
* Slicing the waveform up, and changing the clip gain on each section - you see a major vertical line where the gain changes. Keep that in mind.
* Doing a fade out on the clip itself - looks better than the automation
* Changing the amount of automation thinning (default is 50%, tried 0% and 99%) - nothing too obvious
* Changing between 24 bit and 32 bit float - nothing obvious
* Changing from 64 bit engine to 32 bit engine - nothing obvious

So, no clear pattern here.. But, I would caution to not jump to conclusions. There may be a problem, but there may not be. And I'm strongly suspecting the latter.

Now, my DSP theory is in my rear view mirror at this point - but I think we're seeing more of the visual artefacts of the debug tool with the change in gain than any actual audio issues.

If you have a heavy DSP background, and understand spectral leakage, the difference between Hann, Welch, and Bartlett windows, and the impact of the window size, then I think you're safe to interpret this display. As I said, it's in my rear view window, but I do remember enough to be wary.

The one thing that I did confirm - changing gain causes a visual artefact. HOW each DAW chooses to perform a gain change during automation will display differently.

I do need to drop out of this thread now, so I'll leave further debate to others.
User avatar
by Nip on Fri Mar 15, 2019 11:49 pm
j0001s wrote
Nip wroteI did two bars as untouched, and two bars fading to infinite automation.
Peak level -6 dBFs, 1 KHz signal 48k....
And even doing manual fade when recording to second track, nothing by automation, it is related to algo doing volume it seems.


Did you disable the Studio One dither before you did the test? It is on by default, and it must be off for a legit comparison.


Yes, the first thing I did installing StudioOne - since I mostly use Waves L2 on master with those options anyway. Not in this testing case though, nothing in plugins used.

I also suspected dither as in what youtuber did, but it did not affect much. Also seeing what Jemusic wrote - that fixed at -60 dB.

But I found it follows level of volume from automation, and sticking about -50 dB below the actuall signal all the way. And according to my calculations this is introducing 0.3% THD for no good reason just doing automation ramps.

But as I mentioned - I did one with dither on as well.
I did one separate with one option flipped on/off for those I mentioned - there were something about mastering as well, not even using it I thought something might be on anyway or similar.

It could be as StudioOne do things now is very resource efficient. I find StudioOne very efficient in this regard.

But give us an option like "Use high cpu automation when rendering" or similar depending on what the actuall reason is for anomalies.

If to wait a minute extra or more rendering a full project offline does not matter.
If having to wait a little bit more on various transform and bounce options - it does not matter either - to me anyway.

I tested Digital Performer 8.x and 9.x a couple of times when those were released for windows. I saw that cpu was double for the same project compared to competition, but ProTools.

In this youtuber test with automation DP did really well.

One thought is that maybe they do it in a way that simply produce better quality in regard of automation because of the higher cpu.

But my Sonar Artist 2015 did equally well. I think they tested version 2018.11 in video - which might suggest that they introduced some changed in how they do things - since my 4 year old version showed no anomalies at all for automating ramps.

Thanks everybody that also did their own tests and shared. This might show different results.

To me StudioOne is incredible in so many ways, and feeling as light on system as Reaper - and even much better in how Presonus implemented video support which is way better than Reaper or anybody elses approach(if you can live without thumbnails on a track).

So I'm going for "know your tools" and just avoid doing ramps of automation over audio - or I do that in another daw like Sonar - over and back into StudioOne. I saw nothing in Sonar above -78 dB or around so - looking mostly at 3 kHz since that is first odd harmonic.

If I'm getting such a large project that VCA's are really needed, maybe even nested VCA's - I will avoid ramps anywhere - it would create ramps everywhere there isn't any already.

But if I got an option to set to use other type of volume changes, since moving a fader manually while recording also produce these things - I would be very pleased. I see no reason for StudioOne to do worse than Sonar in this regard.

As a retired programmer I'm thinking Presonus were focused so much on improving something else that they forgot what it does to ramps of automation.

*** Windows 7 Pro * i7-860 2.8 Ghz 16 GB ram * RME HDSP 9632+AI4S+Audient ASP 800 * GT730 ***
User avatar
by Jemusic on Sat Mar 16, 2019 2:53 pm
I did something interesting yesterday. Take the rendered audio you have created with automation fading from some ref level down to zero and put it on a track and insert the PreSonus Spectrum meter over it. Loop the whole thing and watch what happens.

This is a much more dynamic thing to watch actually. Something that Sonic Visualiser does not show very well. Yes at the start of the fade artifacts do appear. I and not doubting that either. That in itself is interesting for sure and one could wonder why. But when you observe the peak amplitude of the artifacts though I noticed they are exactly 60 db down compared to the tip of the signal level.

The relationship of the artefacts is locked to the signal. They do NOT get louder as the signal fades, but rather they stay 60 db down all the time from the signal. So when they signal is at say -30db the peaks of the artifacts are now sitting at -90 dB. Also the artefacts themselves look like they are bordering on noise rather than fixed frequencies. Yes they are a little node like for sure but I wonder with a complex signal they would in fact sound more like noise. But very very low level noise at that.

So as I have previously said how loud does something that is 60 db down actually sound in relation to the signal. Not loud at all I would say. Inaudible might be the answer to this. The guy on the video says it is audible and even the OP says this too but I wonder if its a case of seeing something and now they are imagining they are hearing it too. And only when automation starts its movement as well.

This is not worth not using automation for either. That is rash objective because automation is a big part of any DAW so why would you suddenly stop using it anymore. Transferring things to another DAW is also not on the cards either. When you are under time pressure this is not the thing to do.

So to Nip and the OP start listening and stop measuring and looking. Do some fades on tracks, buses and a total mix. Have a real close listen people and see if you can actually hear it. I really wonder if in fact you can. In a blind listening test between say Studio One and DP doing the same fade over the same material I really doubt you will be able to pick the DAW. I have conducted many blind tests in my day and people really get into trouble here. Basically they don't know what they are listening to most of the time.

And yes the editor doing the fade is free of the artifacts too. I have always believed the best place to fade a full mix is in the editor not in the DAW anyway. This just conforms that for sure.

This has been interesting and whether they need to fix it I am not sure. As I have said they may not suddenly decide to chase this at all but keep going with other way more important things. But who knows maybe they will look into it. It is interesting that one DAW does it and another does not. But not worth trashing that DAW though.

Automation is also often used when it can be avoided. I like avoiding it. Cut your long audio events up into smaller sections and adjust the gain handles of the individual events that way. It is way faster and much better.

Specs i5-2500K 3.5 Ghz-8 Gb RAM-Win 7 64 bit - ATI Radeon HD6900 Series - RME HDSP9632 - Midex 8 Midi interface - Faderport 2/8 - Atom Pad/Atom SQ - HP Laptop Win 10 - Studio 24c interface -iMac 2.5Ghz Core i5 - High Sierra 10.13.6 - Focusrite Clarett 2 Pre & Scarlett 18i20. Studio One V5.5 (Mac and V6.5 Win 10 laptop), Notion 6.8, Ableton Live 11 Suite, LaunchPad Pro
User avatar
by sirmonkey on Sat Mar 16, 2019 9:46 pm
But.... but... butt? Buuut...

IN PRACTICE......can you HEAR anything that bothers you?

I admire skilled engineers and mathematicians and designers of people who make useful products of all sorts. But as a sound tool, can anybody HEAR any undesirable results in any mix? Can we have ONE example of where a bad result in sound actually happened?

Atari 5200, 64K RAM S1PRO Radio Shack Cassette Recorder w/internal Mic, and too many plugins.
User avatar
by Nip on Sun Mar 17, 2019 2:57 am
I turned completely again - and regard this sonic Visualiser as bogus tool, at least what spectrogram is presenting.

I looked at various EQ, frequency analyzers in StudioOne. Bluecat free one, Waves H-EQ and PAZ freqency and analyser.

a) nothing suggest a third harmonic at all - which SV show band on every multiple of fundamental.
b) there is first even harmonic at 2 khz in H-EQ at -50 dB, but not in Bluecat free

I see nothing alarming on frequency content at all - looking at other tools than SV.
Not sure what it is SV is supposed to show.

I am just releaved being free to use anything at any time any way I find useful.

I should have thought of looking at other analyzers/eq first before making a comment even, silly me. :(

*** Windows 7 Pro * i7-860 2.8 Ghz 16 GB ram * RME HDSP 9632+AI4S+Audient ASP 800 * GT730 ***
User avatar
by Jemusic on Sun Mar 17, 2019 5:48 am
I am not sure SV is completely at fault though. But yes it may not be the best tool either. What I found with the spectrum analyser is the dynamic nature of it. There are some artifacts for sure at the start of an automated fade. But what you don't see in SV is how the artefacts are also reduced in level as the signal level goes down in real time. That was interesting. And also the fact they are always 60 dB apart. I still think it is nothing too much to worry about.

But in answer to simonkey's question simply NO. I cannot hear anything bothersome. The guy in the video was very quick to say it is audible but nothing to explain what he was hearing.

Specs i5-2500K 3.5 Ghz-8 Gb RAM-Win 7 64 bit - ATI Radeon HD6900 Series - RME HDSP9632 - Midex 8 Midi interface - Faderport 2/8 - Atom Pad/Atom SQ - HP Laptop Win 10 - Studio 24c interface -iMac 2.5Ghz Core i5 - High Sierra 10.13.6 - Focusrite Clarett 2 Pre & Scarlett 18i20. Studio One V5.5 (Mac and V6.5 Win 10 laptop), Notion 6.8, Ableton Live 11 Suite, LaunchPad Pro
User avatar
by klypeman on Sun Mar 17, 2019 7:36 am
Right, I'll just keep using my ears then..

Forum Moderator.
Please add your specs to your SIGNATURE.
Search the STUDIO ONE 6 ONLINE MANUAL. Access your MY.PRESONUS account.
OVERVIEW of how to get your issue fixed or the steps to create a SUPPORT TICKET.
Needs to include: 1) One Sentence Description 2) Expected Results 3) Actual Results 4) Steps to Reproduce.


AMD Ryzen 7 3800X 8-Core 3.90 GHz Windows 10 64bit Home, MacBook Pro + Mini Big Sur, SO 5.5 Pro, SO 6.1.1 Pro, LUNA, Cubase 10.5 Pro, SLIII32, 24.4.2, 16.0.2 FW, Quantum 2626, Faderport 8, FaderPort 16, Atom, Atom SQ, Console 1+Fader 1, UA Apollo Twin TB/USB, Toontrack, NI Komplete 12, ++
User avatar
by kevinkbell on Sun Mar 17, 2019 2:17 pm
I wonder how this would stack up against modern day DAWs. :roll:


Image
User avatar
by sirmonkey on Sun Mar 17, 2019 10:30 pm
Holy cow... I had a friend who lent me one of those. My memory isn't perfect, but I could swear that it was the exact same model. From 1997 or 1998? Am I close?
What a nice unit that was!

Atari 5200, 64K RAM S1PRO Radio Shack Cassette Recorder w/internal Mic, and too many plugins.
User avatar
by Nip on Mon Mar 18, 2019 12:20 am
Jemusic wroteI am not sure SV is completely at fault though. But yes it may not be the best tool either. What I found with the spectrum analyser is the dynamic nature of it. There are some artifacts for sure at the start of an automated fade. But what you don't see in SV is how the artefacts are also reduced in level as the signal level goes down in real time. That was interesting. And also the fact they are always 60 dB apart. I still think it is nothing too much to worry about.



SV show as much going on at every multiple of fundmental frequency - you see nothing of this, like third partial, in other analyzers. I have not found one other analyzer/eq that show any activity at third at all.

What I saw at most, was -60 dB at first partial being 2 kHz which is the nature of various warmifiers that generate some at octaves to sound phatter etc. Do ramps is StudioOne and you have a built in warmifier, kind of. Very subtle, but still.

Bluecat was good in it going down to -120 dB so you can follow through the fade in action.

SV is buggy and does not represent a truthful picture. It shows bands of anomalies at every multiple of fundamental, and same levels too. This is bogus as I discovered. They a cumulating something in absurdom.

Interesting in that different daws show different degree of those artifacts in SV, but don't know what that is - certainly not on the levels portrayed which I believe would be audible. THD of 0.3% would be audible for someone that spent enough time mixing(THD is measured at first odd partial being 3 x fundmental).

I spent an hour with some classical music with a lot of brass and choirs and stuff - rich on frequencies - and created two same tracks and small ramps of 1 dB all the way, up and down, on one, and nothing on the other and used exclusive solo between them trying to spot anything different - and fail finding any differences.

I feel perfectly safe to use automation, VCA's and nested VCA's as needed in StudioOne.

*** Windows 7 Pro * i7-860 2.8 Ghz 16 GB ram * RME HDSP 9632+AI4S+Audient ASP 800 * GT730 ***
User avatar
by niles on Mon Mar 18, 2019 2:55 am
I did a simple test too by rendering a 440hz sinus with a short fade in both Studio One and Live. I isolated the sound that was added by the automation ramp. In contrast to Live, there is some noticeable noise (when knowing what to listen to) at the part of the automation ramp in Studio One's result.

What this means in real life situations with dynamic content, a broader frequency spectrum and probably a lot more going on, I'm not sure. Still it can be valuable information for certain delegate tasks.

For those interested, here's a video of the test I did:

phpBB [video]


Additional info: Both Cubase and Studio One introduced noticeable noise. While Live, Bitwig and Reason introduced less obvious noise. I also noticed that similar noise introduced by plugins is equal across DAW's. For instance Plugin X performs just as good/bad in DAW B or C.
I also noticed that changing the processing buffer, also changes the noise pattern.

OS: Windows 11 Pro | HW: Gigabyte Z690-UD-DDR4 • INTEL i7 12700K • 64GB • 3x EVO 860 • NVIDIA GT1030 (@WQHD) • RME AIO
User avatar
by Skaperverket on Mon Mar 18, 2019 1:36 pm
Thank you so much for sharing your results and thoughts, niles! Much appreciated!

Unfortunately, you are right, it's clearly audible once we know what to listen for.

I finally found some time myself to test this, and here are some of the things I've come across:

1. Event fades produce slightly less noise than fades done by automating the volume fader.

2. The noise occurs also when playing back smaller automation moves, say a dB up, a dB down. I.e. not only long fades.

3. The noise can become very noticeable once I start applying processing after the automation, f.i. if I use distortion after the automation. In some cases, when using heavy distorion and a hi pass filter after the volume automation, a chattering sound of this noise was the most prominent sound in the whole signal.

Other points of interest:

4. The curve of the event fade and the curve of the volume fader automation do not match, so when both are set to a linear curve, they'll sound very different. I never compared this before, just going by ear, but still interesting and nice to be aware of.

5. When an instrument track/channel is record enabled, you get zipper noise when moving the volume fader while playing bass notes.

6. I tried to find a replacement for the volume fader for automating volume. Unfortunately, the Mixtool plug-in also makes noise. In fact, most of the plug-ins I tried did make some noise, but slightly less. The gain fader of Goodhertz seems to be the cleanest one I've found so far, with no audible problems under the same (extreme) conditions as described in 3.

7. The amp envelope in SampleOne also makes a similar noise. Applying distortion and filtering after this is not unheard of, so this should in my opinion also be looked into.
User avatar
by Jemusic on Mon Mar 18, 2019 2:42 pm
Well I have changed my tune as well. I agree with both Niles and Skaperverkert that this can be heard. And here is the way to really hear it. I have been trying to figure out a way to hear these artifacts,

2nd Sense Audio are giving away a free EQ. Here:

https://2ndsenseaudio.com/plugins/

Scroll right down to the bottom of the page for the free EQ. What is cool about this EQ is that in cut mode it can provide a max of 60 dB attenuation. Something that most EQ's cannot do. So notch out the tone. Set a single band for 440 Hz (or the exact frequency of your tone) and very steep Q and max attenuation. This EQ needs to be inserted on your main output mix, not on the track itself. (if its on the track you want hear anything at all, artifacts also get attenuated!)

Although the tone is still there it is super quiet now. What you will hear is the artefacts generated by the fade. You will need to crank your monitors though right up! Niles is also correct in that changing the buffer size also alters the sound of the noise at the fade. I still say that in most cases compared to the full signal level these artifacts are harder to hear. But once you do this notch test and you really start to hear them then you know what you are listening for.

I agree it should be fixed. Ableton is dead quiet with the fades as well. (note in the original video Ableton looks bad but he is very wrong on that. I am hearing nothing in terms of artifacts from Ableton in fact) My fades with the editor are also silent too.

Update

Fading out at the end of an audio event e.g. doing a 1 or 2 second fade by dragging end fade handles in, does not produce any artifacts either. So this mainly happens with automation being drawn over tracks etc.. I know a fade out at the end of an event won't solve more complex automation requirements within an event but if its just a fade out you are wanting to do that is best way to do it. I tend to do a lot of that as well hence the reason I have not heard this before.

Specs i5-2500K 3.5 Ghz-8 Gb RAM-Win 7 64 bit - ATI Radeon HD6900 Series - RME HDSP9632 - Midex 8 Midi interface - Faderport 2/8 - Atom Pad/Atom SQ - HP Laptop Win 10 - Studio 24c interface -iMac 2.5Ghz Core i5 - High Sierra 10.13.6 - Focusrite Clarett 2 Pre & Scarlett 18i20. Studio One V5.5 (Mac and V6.5 Win 10 laptop), Notion 6.8, Ableton Live 11 Suite, LaunchPad Pro
User avatar
by Nip on Tue Mar 19, 2019 12:41 am
I saw -50 dB at second order partial, octave, by nothing at third.
Bandpass out second order, sure you will hear that cranked up.
Or notch out test frequency.

So if you hear at octave it will not sound harsch, like if anything at third fundamental.

So test signal 1 kHz, put bp at 3 kHz, or notch at 1 Khz
If 440 Hz put bp at 1320 Hz, or notch at 440 Hz..

Depending on steepness of filter, 6 dB/octave, 12 or 24 etc - matter how much the filter affect noise/artifacts as well.
If roughly 1.5 octave to third partial - improve hearing it by 9, 18 or 36 dB filtering out test signal.

Personally, not seing anything above octave from test signal in analyzer/eq regarding content, especially Bluecat analyzer going to -120 dB I have no real problem with it. But lost confidence in Sonic Analyser in how severe it shows it to be, not idea what they are cumulating to get those results.

But if a blank sheet and nothing but fundamental, fine too of course. If severly increase cpu, I would like a setting how to do it - if playback and rendering so we can choose what matters more.

Interesting find that clip envelopes show nothing.

EDIT. Used Waves H-EQ and did a notch, digital 2 slopes to notch out 1 Khz.
So at about -65 dB is first even partial, and first odd partial at -75 dB which means 0.017% THD.
Amplifying this you can hear a little zipper noise, just at the end 1-2s or so.

This is big difference from 0.3% THD that Sonic Visualiser suggest in how they present anomalies..

BluecatAudio analyzer show nothing at third for some reason, until you get to zipper noise very much below test signal.

For me this stays OK, nothing I really worry about.
Last edited by Nip on Tue Mar 19, 2019 3:09 am, edited 1 time in total.

*** Windows 7 Pro * i7-860 2.8 Ghz 16 GB ram * RME HDSP 9632+AI4S+Audient ASP 800 * GT730 ***
User avatar
by Jemusic on Tue Mar 19, 2019 3:02 am
Despite measurements and how things look, once you almost null out the tone and hear what is left it is quite audible. And once you get to know the sound you can hear it all the time. And as Skaperverket also mentions any processing down the line might actually emphasise it as well.

It is something that hopefully will be addressed. Yes fading the ends of events is fine so that does work OK. But more complex automation should be able to be done without any additional noise interference.

Specs i5-2500K 3.5 Ghz-8 Gb RAM-Win 7 64 bit - ATI Radeon HD6900 Series - RME HDSP9632 - Midex 8 Midi interface - Faderport 2/8 - Atom Pad/Atom SQ - HP Laptop Win 10 - Studio 24c interface -iMac 2.5Ghz Core i5 - High Sierra 10.13.6 - Focusrite Clarett 2 Pre & Scarlett 18i20. Studio One V5.5 (Mac and V6.5 Win 10 laptop), Notion 6.8, Ableton Live 11 Suite, LaunchPad Pro
User avatar
by Nip on Tue Mar 19, 2019 3:22 am
I see no way you hear this together with good signal and something -75 dB below.
I think the Remove Silence tool has a setting for -80 dB for default threshold.

If good signal lower to -50 dB the anomalies are -75 dB below that.
It's a function of what is there.

These things can float around in project, at the bottom and depending on processing be emphasized - sure.
Other math done can float this - in theory.
But feel I'm with non-alarmists in how reality is.

But always good to know your tools, I guess.

*** Windows 7 Pro * i7-860 2.8 Ghz 16 GB ram * RME HDSP 9632+AI4S+Audient ASP 800 * GT730 ***
User avatar
by Skaperverket on Tue Mar 19, 2019 4:08 am
Just a quick update/tip:

Automating the Mixtool plug-in (especially on low frequency content) can also cause a bigger problem, at least if the automation curve passes through 0dB on the gain knob in Mixtool. An audible click occurs. As if it's changing an algorithm or something. The click is not present if you stay below (or above) the 0 dB mark all the time. Good to be aware of, and hopefully something than can be fixed in an update.
User avatar
by multifederal on Tue Mar 19, 2019 5:45 pm
What can be more important than our audio? In the digital world we expect no anomalies, only the ones we introduce on purpose. In a 64bit double-precision environment we expect 'pristine'. My fades should not be introducing distortion.


SO, the S1 developers SHOULD BE CHASING THIS.

The dismissive ones are the typical hacks, I don't listen to them. If the Ardour DAW can do it clean, then why can't S1? < that's a fair question!

Ardour (a cheap affordable DAW) won the test. S1 is supposed to be in the major league.


Some hacks think for cars oil is oil, but they are wrong. And some of them change the oil every two or three years, when the experts say to change it every six months. Can we hear the destruction that is happening when we don't change the oil on-time, no! Does it matter? Yes! But the hacks will still say "don't worry about it"


I say again... I WANT MY FADES PRISTINE!
User avatar
by Jemusic on Tue Mar 19, 2019 6:32 pm
multifederal wroteWhat can be more important than our audio? In the digital world we expect no anomalies, only the ones we introduce on purpose. In a 64bit double-precision environment we expect 'pristine'. My fades should not be introducing distortion.


SO, the S1 developers SHOULD BE CHASING THIS.

The dismissive ones are the typical hacks, I don't listen to them. If the Ardour DAW can do it clean, then why can't S1? < that's a fair question!

Ardour (a cheap affordable DAW) won the test. S1 is supposed to be in the major league.


Some hacks think for cars oil is oil, but they are wrong. And some of them change the oil every two or three years, when the experts say to change it every six months. Can we hear the destruction that is happening when we don't change the oil on-time, no! Does it matter? Yes! But the hacks will still say "don't worry about it"


I say again... I WANT MY FADES PRISTINE!


Read the whole thread. I have changed my point of view on this and they are looking into it.

Specs i5-2500K 3.5 Ghz-8 Gb RAM-Win 7 64 bit - ATI Radeon HD6900 Series - RME HDSP9632 - Midex 8 Midi interface - Faderport 2/8 - Atom Pad/Atom SQ - HP Laptop Win 10 - Studio 24c interface -iMac 2.5Ghz Core i5 - High Sierra 10.13.6 - Focusrite Clarett 2 Pre & Scarlett 18i20. Studio One V5.5 (Mac and V6.5 Win 10 laptop), Notion 6.8, Ableton Live 11 Suite, LaunchPad Pro

111 postsPage 3 of 6
1, 2, 3, 4, 5, 6

Who is online

Users browsing this forum: patricemazmanian and 39 guests