Topic: 3.8 reports, queries and bugs

[winner]

I'm having issue playing radio stream with XMPlay (latest stuff 3.8.5.20).

So I have this mp3 stream working fine:
https://maggie.torontocast.com:8020/mp3-320

But aac version is having issue:
https://maggie.torontocast.com:8020/aac-320

It plays for 4 seconds and then restarts by itself, like you double clicked again on the stream. And it happens every 4 seconds.
This is with built in aac decoder. I've tried with AAC plugin and it works fine.

In web browser this stream works fine.

Could this be related to XMPlay or its built-in aac decoder?

I am using aac plugin rev. 6c and stream plays fine.

[garson]

I am using aac plugin rev. 6c and stream plays fine.

Try without aac plugin (XMPlay has built-in aac decoder).

[Ian @ un4seen]

Test1.mod exhibit two issues:
 - vibrato depth/speed sounds wrong
 - contains noise or aliasing due to higher root note?

Test2.mod exhibits only the first issue:
 - vibrato depth/speed sounds wrong

Thanks. The problem is that XMPlay is resetting the vibrato with each lone instrument. Here's an update that should fix that:

   www.un4seen.com/stuff/xmplay.exe

Edit: I added another Test3.mod as a separate attachment. It shows more 'noise' or aliasing, with less clear sample swapping transitions in XMPlay than in other players. Taken from pattern 18.

Hmm the noise I'm hearing seems to be solved by the "Amiga resampler" feature that PT2Clone and OpenMPT use. Perhaps that would be a good addition to XMPlay?

Yep, that seems to be aliasing noise. It sounds better when linear or sinc interpolation is enabled? You should get the same noise in OpenMPT if you set its resampling option to "No Interpolation".

I'm having issue playing radio stream with XMPlay (latest stuff 3.8.5.20).

So I have this mp3 stream working fine:
https://maggie.torontocast.com:8020/mp3-320

But aac version is having issue:
https://maggie.torontocast.com:8020/aac-320

It plays for 4 seconds and then restarts by itself, like you double clicked again on the stream. And it happens every 4 seconds.
This is with built in aac decoder. I've tried with AAC plugin and it works fine.

The update above should fix this. Let me know if you still see it happen.

[winner]

The update above should fix this. Let me know if you still see it happen.

The stream played faulty as described with the native aac decoder, but now plays fine on my system with the updated version, XMPlay ver. 3.8.5.21.

[garson]

Thanks Ian.
Issue is fixed.

[brycco]

Thanks. The problem is that XMPlay is resetting the vibrato with each lone instrument. Here's an update that should fix that:

   www.un4seen.com/stuff/xmplay.exe

That fixed it, masterful work!

Yep, that seems to be aliasing noise. It sounds better when linear or sinc interpolation is enabled? You should get the same noise in OpenMPT if you set its resampling option to "No Interpolation".

I wouldn't say interpolation sounds better, as it's cutting off more high frequencies than the Amiga's own filter does. Of course its subjective, interpolated lo-fi 8-bit samples can have an interesting aesthetic when done intentionally.

The Amiga Resampler option in OpenMPT eliminates aliasing while not sacrificing higher frequencies. It can be set to Unfiltered, A500 and A1200, and in all 3 cases the aliasing is reduced significantly by comparison. I quite like the "Unfiltered" mode as it's closest to No Interpolation but without aliasing it seems.

Anyway, cheers!

[Jo Li KMC]

Hey.  Long time no see, huh?

So, I… think I'm having the same problem as Garson?  There's a couple of tunes from an MS-DOS game called Silverball.  They're converted from ProTracker Studio beta modules into S3M.  In any case, these two files – the Duel and Warbot table themes (titled "ROYIII#3" and "rnwiii#9", respectively) – seem to have a rather odd problem when playing under XMPlay v3.8.5, as well as the v3.8.5.21 test version posted above.

As you can hear from the demo OGG files, certain samples, for some reason, make the audio snap and crackle no matter what the bit rate, "interpolation", or "ramping" audio settings are.  This doesn't happen when the files are played in any other audio player, nor does it happen if I use the OpenMPT input core (v0.5.5.14110).

I love XMPlay like crazy, but hearing my favorite tunes sound like that under the built-in core really bums me out.  I'm not entirely sure what's going on or how to fix it, but hopefully, this is something simple that won't need a lot of fine-tuning on anyone's part.  If not, then I apologize for the inconvenience!

Here's to good news!

[saga]

While it doesn't fix the problem at hand, you should probably throw the original Silverball PSM files into XMPlay instead, as those can be played by the OpenMPT input plugin as well. I know that the PSM files are not the original sources, but they are one small step closer to the original files than those really badly converted S3Ms - in particular the "Duel" tune doesn't sound like a detuned mess that way.

[Ian @ un4seen]

Oops! The internal mixing level was too high for those files. Here's an update that should get them sounding right:

   www.un4seen.com/stuff/xmplay.exe

Btw, I noticed that the MOD/XM/etc volume levels are generally much lower in XMPlay 3.8.4/5 compared to previous versions. Paradoxically, that's because the level scanning became more accurate in 3.8.4. In previous versions, the scanning could give a low reading, which resulted in a higher gain being applied. That gain has now been tweaked for the more accurate level scanning.

[saga]

Btw, I noticed that the MOD/XM/etc volume levels are generally much lower in XMPlay 3.8.4/5 compared to previous versions. Paradoxically, that's because the level scanning became more accurate in 3.8.4. In previous versions, the scanning could give a low reading, which resulted in a higher gain being applied. That gain has now been tweaked for the more accurate level scanning.

Damn, I knew something sounded too quiet but I couldn't quite figure out why, given that in general the volume scanning still seemed to work.

[Jo Li KMC]

[Merged into the next post]

[Jo Li KMC]

Originally posted on: 19 Jun '21 - 01:21 EDT

Oops! The internal mixing level was too high for those files. Here's an update that should get them sounding right:

   www.un4seen.com/stuff/xmplay.exe

Yep.  That absolutely fixed the problem.  That would also explain why some other files had a case of crackles, too – albiet nowhere near as noticeably.

Btw, I noticed that the MOD/XM/etc volume levels are generally much lower in XMPlay 3.8.4/5 compared to previous versions. Paradoxically, that's because the level scanning became more accurate in 3.8.4. In previous versions, the scanning could give a low reading, which resulted in a higher gain being applied. That gain has now been tweaked for the more accurate level scanning.

Hm.  I'm not sure I noticed that, buuut… it's good that you did, Ian!  That sounds like it could be a weird one.

While it doesn't fix the problem at hand, you should probably throw the original Silverball PSM files into XMPlay instead […]

Aaactually… that also fixed the problem, funny enough.  Seriously.  I don't know why.  Especially considering that I did end up converting them to S3M via OpenMPT, anyway, just for consistency!
And yes, "Duel" sounds much better when it has the proper pitch-bends.  (:
(I never played that table much, so I never knew the difference.  I mostly just played "Snooker Champ" for its catchy music~)

Thank you both for your insight and help!  You guys don't get enough props.  Seriously~

Unbelievably, I've tripped over another issue.  And, remarkably, it's with another specific module.

Attached to this post is a 7z file containing a ProTracker Module, "POPCORN7" (that's not its original name – I don't know what it was originally called), and a couple of demonstration OGGs.  For some reason, when the song hits Pattern 12 and starts playing Sample 3 (a string sample) on Channel 3, the sample gets interrupted by Sample 4 (a bass drum) before playing.  It results in an odd-sounding "thud-hum" on every note.  As far as I can tell from the pattern data, there's absolutely nothing that's telling it to do this, or anything even hinting that it should.  The only oddity that I can see is that every note played is set to play at 48 volume.  But… all the other samples have volume effects, too.

As always, it plays fine in other players and even older versions of XMPlay, so… yeah.  Weirdness.

If it's any help, Ian?  I found a bug with a different module in OpenMPT that's just as weird, which I'm submitting to the OMPT bug tracker.  So, I'm not just picking on you~  (;

[gabelincoln]

Report: best player of all time, thank you!!!  How can it be so succinct and predictive when everything else is so deeply flawed?  How liberating it is to be freed from the tyranny of subpar players!  And not a support button in sight!  A true saint, a visionary, a god!

[Ian @ un4seen]

Attached to this post is a 7z file containing a ProTracker Module, "POPCORN7" (that's not its original name – I don't know what it was originally called), and a couple of demonstration OGGs.  For some reason, when the song hits Pattern 12 and starts playing Sample 3 (a string sample) on Channel 3, the sample gets interrupted by Sample 4 (a bass drum) before playing.  It results in an odd-sounding "thud-hum" on every note.

It looks like that file was made with some funky tracker that saved the sample loopstart positions in bytes rather than words (which is the norm), so XMPlay ends up playing some of the next sample too (a bass drum) in the loop. Here's an update that will try to detect such files:

   www.un4seen.com/stuff/xmplay.exe

Report: best player of all time, thank you!!!

Thanks!

[Jo Li KMC]

Yep.  That fixed it, alright.  Cool beans!

I'll let you know if any other oddities pop up, as always.  Hopefully none will, though~

[saga]

That Popcorn track looks like it's a straight conversion from a Ultimate Soundtracker module to a ProTracker module... maybe a bit too straight (in UST the loop start would be in bytes). Probably not a funky tracker but a funky converter was at fault here. You will most likely be able to find the original fine in Soundtracker format somewhere.

[Rah'Dick]

Hey everyone!

I just had a confusing situation that took a surprisingly high amount of tinkering to resolve. I know it's not a bug, but I was truly stumped for a bit.

When you have a search term set in the extended playlist that does not match any track AND you have "skip non-matching tracks" enabled, you can still doubleclick any track to play, but the list doesn't auto-advance and the forward/back buttons stop working, too. Obviously that's the desired behavior, but if you forget that "skip non-matching" checkbox and have to resort to comparing INIs to fix it, maybe there should be some kind of hint or help to resolve this. I'm thinking either of these:
- An info bubble "Cannot change tracks, nothing matches %search%"
- Automatically disable "skip non-matching tracks" when the above condition is met and the forward/back buttons are pressed more than 3 times
- Any other feedback that helps me figure out why I can't skip tracks, even though the playlist clearly isn't empty

What do you think?

[Ian @ un4seen]

A simple option could be to play Windows' error sound when the next/prev buttons are pressed in that scenario. Would that work for you, or do you have the Windows sounds disabled?

[Keltic Danor]

Not sure if it going bonk would be enough information, unless you knew it was the search it would just be bonking in addition to the not behaving as you'd normally expect.

I think either the tool tip indicating the search restriction or maybe just ignore it the setting if there are no results available.
As in if search = on and skip non-matching = on but search results are 0 then ignore it and behave normally. Not disable it as such but suppress it while there are no matches.

[Ian @ un4seen]

OK. I couldn't come up with anything better, so here's an update that should simply ignore the "skip non-matching tracks" setting when there are no matches:

   www.un4seen.com/stuff/xmplay.exe

[Rah'Dick]

Thank you kindly! 

[Metro28]

An .it file gets slow and displays the number of channels incorrectly.
A 57 channel .it file, when I listen to it shows about 230 channels playing.


https://modarchive.org/index.php?request=view_by_moduleid&query=32781

[Ian @ un4seen]

The channel count display includes virtual channels, so that's why it gets so high with that file. You can limit the number of virtual channels in the "MOD" options page.

[Metro28]

Thanks

[Metro28]

Old IT file not work.
An Impulse Tracker IT file does not work if it is from an old version (1.06)