MainSqueeze wrote:
> Well, the short version is: I can extend gormans custom-convert
> to apply equalizer settings to remote streams as well, but in doing so
> Im losing the ability to seek within individual streams. The long
> version follows:
>
> My test case is a Bandcamp URL:
>
> What am
MainSqueeze wrote:
> What am I missing? (It wouldnt be the first time Im missing something
> obvious.) Or is this simply something that cant (yet) be done?
May I suggest a different approach since it looks like you may want to
tackle all audio formats on the same player.
Since the Mac
Well, the short version is: I can extend gormans custom-convert to
apply equalizer settings to remote streams as well, but in doing so Im
losing the ability to seek within individual streams. The long version
follows:
My test case is a Bandcamp URL:
Code:
scala wrote:
> I was a bit slow in catching this very interesting and learned thread.
> I've been using equalization for years, and the fact of not being able
> to seek in mp3 tracks always bugged me: thanks for this solution!
> Although I don't have many mp3s, I did not like the anomalous
MainSqueeze wrote:
> What I'd like to do is see how far we can take custom-convert (for
> equalization, room correction, or personal preference, to compensate for
> room boundaries when positioning speakers, or even just to add reverb to
> every file played back through LMS if that's what it
MainSqueeze wrote:
> Plugins apparently are "streamMode: I" rather than "R",
No. This sweeping statement is not correct .
If you use I - then it tells LMS you want to use STDIN (e.g LMS will
hadle http and ICY stuff)
If you use a rule with R - then LMS will let the trsancode use the
$PATH$
gorman wrote:
> Hi! Did you survive the pub? Waiting for new debug lines :-)
Easy, Buster. I've got a few other things on, too, you know. ;)
Anyway, I've already discovered that not all remote streams are in fact
remote. Plugins apparently are "streamMode: I" rather than "R", even
though they
gorman wrote:
>
> I had to use ffmpeg because, as far as I can see, LAME does not accept
> time offsets when decoding
> (https://svn.code.sf.net/p/lame/svn/trunk/lame/USAGE). Don't know how
> the internal functionality of mp3 transcoding works (when lowering
> bitrates) because that is capable
MainSqueeze wrote:
> Great. Sounds like a plan. What I'd like to do is see how far we can
> take custom-convert (for equalization, room correction, or personal
> preference, to compensate for room boundaries when positioning speakers,
> or even just to add reverb to every file played back
MainSqueeze wrote:
> gorman was so right when he said, in post 20 above:
>
>
>
> Black magic indeed. Each little bit you learn only gives you a better
> idea of how much you don't yet know.
>
> Suppose, for future reference, one wanted to use convert-conf to
> transcode (and equalize) not
MainSqueeze wrote:
> But before I can do that I now have to go down the pub, get drunk and
> catch Covid-19 because Boris Johnson tells me that's the only way to
> save the economy...Enjoy your pint, man! Be safe! Better to drink outside, if
> you can :)
gorman wrote:
>
> As for your question above, with the player.source debug log active, try
> to play an mp3 stream of your choice and post the tokenized command
> here. From there we can work it out, I hope.
Great. Sounds like a plan. What I'd like to do is see how far we can
take
MainSqueeze wrote:
> gorman was so right when he said, in post 20 above:
>
>
>
> Black magic indeed. Each little bit you learn only gives you a better
> idea of how much you don't yet know.
>
> Suppose, for future reference, one wanted to use convert-conf to
> transcode (and equalize) not
MainSqueeze wrote:
> Black magic indeed. Each little bit you learn only gives you a better
> idea of how much you don't yet know.
And that goes for the me as well. I dabble in just one small area of
streaming and transcoding. A recent PlayHLS bug highlighted the gaps in
my knowledge.
gorman was so right when he said, in post 20 above:
gorman wrote:
>
> Convert.conf still is some sort of black magic stuff under many
> aspects... it's good to have clarifications on the forums, when
> possible. For future reference.
Black magic indeed. Each little bit you learn only gives
MainSqueeze wrote:
> One more question, if I may, to add to my little custom-convert
> documentation project: the default convert.conf file lists only three
> "current" device types: slimp3, squeezebox and * (for all types). Do you
> know how one might specify more recent models, e.g. Boom and
One more question, if I may, to add to my little custom-convert
documentation project: the default convert.conf file lists only three
"current" device types: slimp3, squeezebox and * (for all types). Do you
know how one might specify more recent models, e.g. Boom and Touch? What
are their device
MainSqueeze wrote:
> I swear I searched that manual on Sourceforge several times for "-C". It
> turns out the manual uses en dashes rather than just ordinary
> hyphens...
>
> Still, now that I've recovered from the embarrassment: if "-C" specifies
> the compression rate, in this case for
I swear I searched that manual on Sourceforge several times for "-C". It
turns out the manual uses en dashes rather than just ordinary
hyphens...
Still, now that I've recovered from the embarrassment: if "-C" specifies
the compression rate, in this case for output to FLAC, then it should be
MainSqueeze wrote:
>
> Incidentally, you'll see that I've left in the "-C 0" option for now
> because, having gone through the documentation once more, I couldn't
> find any mention of it in the SoX or the FLAC manuals; does either of
> you know what "-C 0" actually does?
Code:
MainSqueeze wrote:
> Incidentally, you'll see that I've left in the "-C 0" option for now
> because, having gone through the documentation once more, I couldn't
> find any mention of it in the SoX or the FLAC manuals; does either of
> you know what "-C 0" actually does?
In sox help after
I've now updated my custom-convert file, checked the server logs once
more, and it's all good. No more "rejecting" this line or that, just
"found command" and "matched", and the tokenized command on seeking
within a track now reads:
Code:
Tokenized command:
MainSqueeze wrote:
> Ha. I had a feeling you might say that. ;)
>
> The equalization only takes care of four bass peaks that acoustic
> treatment couldn't quite eliminate, and as long as I didn't search
> within tracks it worked, didn't it?No, it didn't work. Check the beginning of
> the log
bpa wrote:
>
> Now you need to wonder - did you ever notice the lack of equalisation
> :-)
Ha. I had a feeling you might say that. ;)
The equalization only takes care of four bass peaks that acoustic
treatment (which, besides several Squeezeboxes, was my best "HiFi
component" ever) couldn't
bpa wrote:
> Now you need to wonder - did you ever notice the lack of equalisation
> :-)Considering the high q values it's not surprising he did not notice the
difference.
In any case, if you want a one program solution, ffmpeg does everything.
Don't know if your system can handle it (I
MainSqueeze wrote:
> Thank you!
>
> So just to make sure I've understood:
>
> The missing "capability T" is precisely the ability to "seek to a start
> time offset" that's defined in convert.conf, I suppose. The one I
> thought I could do without. ;)
>
> When LMS comes across a command that
Thank you!
So just to make sure I've understood:
The missing "capability T" is precisely the ability to "seek to a start
time offset" that's defined in convert.conf, I suppose. The one I
thought I could do without. ;)
When LMS comes across a command that lacks a required capability (e.g.
my
LMS looks through all the rules sort of in parallel so the log can show
"rejects" after an "accept".
All the accept results will then be examined and best match wins (i.e
roughly a MAC specific match beats player type match beats a generic
match but flac vs PCM vs MP3 and direct vs trasncode
MainSqueeze wrote:
> I've attached the log; I hope you can make more sense of it than I can.
> Worryingly, there are several instances of this:
>
> >
Code:
> >
> [20-07-03 17:15:44.3276] Slim::Player::TranscodingHelper::checkBin (252)
Found command: [sox] -q -t
I've attached the log; I hope you can make more sense of it than I can.
Worryingly, there are several instances of this:
Code:
[20-07-03 17:15:44.3276] Slim::Player::TranscodingHelper::checkBin (252)
Found command: [sox] -q -t flac $FILE$ -b 24 -t flac -
MainSqueeze wrote:
> Will do but it'll be a couple hours or so before I'll get round to it,
> I'm afraid. Which log set should I choose from the dropdown at the top
> of the logging page? Or is choosing "Debug" for the player.source by
> itself enough?
player.source to DEBUG is enough - it
bpa wrote:
>
> Can you do a player.source log set to DEBUG when you use WebUI to jump
> to a mid way point in a track.
Will do but it'll be a couple hours or so before I'll get round to it,
I'm afraid. Which log set should I choose from the dropdown at the top
of the logging page? Or is
MainSqueeze wrote:
> Just to confirm, I can both skip tracks (i.e. go to the next/previous
> track) and navigate to specific points in individual tracks (what in
> analogue days, we might have called RWD and FWD).
>
> I'm controlling playback via the web UI (mostly), iPeng (sometimes) and
>
Just to confirm, I can both skip tracks (i.e. go to the next/previous
track) and navigate to specific points in individual tracks (what in
analogue days, we might have called RWD and FWD).
I'm controlling playback via the web UI (mostly), iPeng (sometimes) and
the touch interface of a
gorman wrote:
> So I'm curious about it.
Again it may be a fuzzy choice of words.
However, if there is no time specified as a start time *and* the
transcoding is being used - then seek to a particular time point in
track cannot happen.
Use logging player.source set to DEBUG to see the full
bpa wrote:
> The single line sox should work but with the suggested but I think the
> meaning of "skip" may be ambiguous.
>
> It may be suitable for you if you do not want to skip to a time point
> *within* a track.
>
> The sox command does not seem to support a start or end time and so
>
gorman wrote:
> I'll try your approach.
The single line sox should work but with the suggested but I think the
meaning of "skip" may be ambiguous.
It may be suitable for you is you do not want to skip to a time point
*within* a track.
The sox command does not seem to support a start or end
MainSqueeze wrote:
> Yes, I can skip (i.e. go to the next/previous track) as well as go to
> particular points in individual tracks using a Squeezebox Touch, the web
> UI (with Material Skin) or iPeng.
>
> Then again, I'm running LMS on Debian and using the Squeezelite version
> in the
gorman wrote:
>
> Can you skip with that line?
Yes, I can skip (i.e. go to the next/previous track) as well as go to
particular points in individual tracks using a Squeezebox Touch, the web
UI (with Material Skin) or iPeng.
Then again, I'm running LMS on Debian and using the Squeezelite
gorman wrote:
> I'm actually not... I don't have material to test... I should re-rip a
> CD with a cuesheet to try.
why - just test the flac command options on a command line with an
ordinary file. No need for cue or LMS. For example, test thta you can
select 15 sec from middle of an ordinary
bpa wrote:
> Are you sure about using "-to" rather than -"t" ? It is one that
> mixes me up and I think i'd use "-t" as it is processing the input end
> time and not the output end time.I'm actually not... I don't have material to
> test... I should re-rip a
CD with a cuesheet to try.
MainSqueeze wrote:
> Thanks for all the info.
> Regarding this:
> > > >
> > With that out the way, I have to admit that I've never seen the point of
> > the commented line (# FT:{START=...) and the first part of the following
> > line, up to the pipe that sends output to SoX. Instead, I've
gorman wrote:
> Ok, found solution for MP3s as well. Since I have music only in flac and
> mp3 my solution is complete.
>
> >
Code:
> >
> mp3 flc * bc:5f:f4:bf:3e:70
> # FT:{START=-ss %t}
> [ffmpeg] -loglevel quiet $START$ -i $FILE$ -f wav - | [sox] -D
gorman wrote:
> If somebody for whatever reason would prefer to use ffmpeg for FLAC
> processing this should work (seek works, END should work for cue files
> but I don't have any to test):
>
> >
Code:
> >
> flc flc * bc:5f:f4:bf:3e:70
> # FT:{START=-ss
If somebody for whatever reason would prefer to use ffmpeg for FLAC
processing this should work (seek works, END should work for cue files
but I don't have any to test):
Code:
flc flc * bc:5f:f4:bf:3e:70
# FT:{START=-ss %t}U:{END=-to %v}
[ffmpeg]
gorman wrote:
> Yeah, I get that... as mentioned, I guess ffmpeg might be an alternative
> for people not wanting to mess with the flac executable. If it works for
> MP3s it'll probably work fine for flac files.
The odd thing was that after a lot of disucssuion, I thought it has been
agreed to
bpa wrote:
> Re why has new Flac not beenincludedin LMS. It's because there are so
> many possibilities and file formats and there is a question whether LMS
> should have a non-standard version of Flac which means more maintenance,
> that committing a change to something as central as "flac" is
bpa wrote:
> I suspect if the Flac file has embedded CUE to identify tracks within
> it. - then there is a "start time" and "end time" for each track.
> Oh, ok, that makes sense. I don't use CUE files, so it's irrelevant for
me. Thanks for explaining this.
Convert.conf still is some sort of
gorman wrote:
> Oh, ok, that makes sense. I don't use CUE files, so it's irrelevant for
> me. Thanks for explaining this.
>
> Convert.conf still is some sort of black magic stuff under many
> aspects... it's good to have clarifications on the forums, when
> possible. For future reference.
Re
gorman wrote:
> Edit: can anybody explain the use of the --until=v% command when using
> flac decoding? With the above line in ffmpeg I've got complete seek
> functionality and I don't get if I'm missing something and what would
> that be.
I suspect if the Flac file has embedded CUE to identify
Ok, found solution for MP3s as well. Since I have music only in flac and
mp3 my solution is complete.
Code:
mp3 flc * bc:5f:f4:bf:3e:70
# FT:{START=-ss %t}
[ffmpeg] -loglevel quiet $START$ -i $FILE$ -f wav - | [sox] -D -q -t
wav - -t flac -C 0 -b 24
ralphy wrote:
> I remember creating a linux i386 and win32 flac binary with the patch,
> but testing stalled after the initial tests confirmed that the change
> worked.
>
> Links to the binaries and details are in the 'slimserver-vendor pull
> request #72'
gorman wrote:
> LMS version is: 7.9.3 - 1579195498 @ Thu Jan 16 21:08:22 WEST 2020
It seems the fix to flac was never put into "production" your language
settings look OK to me.
Looking at your sox command - is the trailing "-" missing ?
Can you post the custom-convert.conf file you are
philippe_44 wrote:
> In your PC locales settings, what defines the decimal separator? There
> is a bug/issue with flac and decimal separator for some non-english.
> Flac expects time in the locale settings, and that might be mm::ss,ff
> and LMS sends mm::ss.ff (I have the issue being a
I remember creating a linux i386 and win32 flac binary with the patch,
but testing stalled after the initial tests confirmed that the change
worked.
Links to the binaries and details are in the 'slimserver-vendor pull
request #72' (https://github.com/Logitech/slimserver-vendor/pull/72).
philippe_44 wrote:
> No it's still here - I almost lost my mind the other day because I
> thought it was my remote-header PR that created an issue. I've verified
> the problem by invoking flac manually
OK then there must be a use case that wasn't tested. I remember testing
and setting locale
bpa wrote:
> Is this still an issue ? I thought it was fixed it see
> https://github.com/Logitech/slimserver/issues/277
No it's still here - I almost lost my mind the other day because I
thought it was my remote-header PR that created an issue. I've verified
the problem by invoking flac
gorman wrote:
> I thought I had it active but maybe I didn't,
This is expanded command line is only output on DEBUG.
Code:
[20-06-30 02:52:07.9655]
Slim::Player::TranscodingHelper::tokenizeConvertCommand2 (616) Using command
for conversion:
philippe_44 wrote:
> In your PC locales settings, what defines the decimal separator? There
> is a bug/issue with flac and decimal separator for some non-english.
> Flac expects time in the locale settings, and that might be mm::ss,ff
> and LMS sends mm::ss.ff (I have the issue being a
gorman wrote:
> I thought I had it active but maybe I didn't, I don't know. Redid the
> whole thing here: https://pastebin.com/xQRYVBn3
>
> You can see it's playing "I'd Do Anything for Love But I Won't Do That",
> I seek and it gives, at 02:52:08.0695, some error about end of file...
> but
bpa wrote:
> Are you using player.source set to DEBUG ?
> There seems to be other logs active but not player.source set top DEBUG
I thought I had it active but maybe I didn't, I don't know. Redid the
whole thing here: https://pastebin.com/xQRYVBn3
You can see it's playing "I'd Do Anything
What I can't understand is why this is happening. The first portion,
which is the one "deciding" seek is identical to other lines in
convert.conf
flc mp3 * *
*# F*B:{BITRATE=--abr
%B}*T:{START=--skip=%t}U:{END=--until=%v}*D:{RESAMPLE=--resample %D}
*[flac] -dcs $START$ $END$ --
gorman wrote:
> What happens when I seek is that the playback stops and it jumps to the
> next song in the queue.
>
> This what I have in debug log from when the song gets transcoded with
> sox to when I seek and it errors out, jumping to the next in queue.
Are you using player.source set to
Ok, as I said it was a while ago and I guess the "seek" problem was
never solved. I moved on to Inguz DRC/EQ shortly after and it seeks
fine. Here is what Inguz has in its custom-convert.conf:
Code:
flc flc * 00:00:00:00:00:00
#
Unfortunately the problem remains. When decoding FLAC and inputting into
sox, seek jumps to the following track in the queue. When decoding MP3
and inputting into sox, seek restarts the current track.
It's funny because it was this five years old post of yours that put me
on to this solution.
I haven't used an "equalizer" custom-convert.conf for a few years, but I
think it worked back then. Some slight differences to your own, so
maybe worth a try:
Code:
flc flc * 00:04:20:00:00:00
# FT:{START=--skip=%t}U:{END=--until=%v}D:{RESAMPLE=-r %d}
What happens when I seek is that the playback stops and it jumps to the
next song in the queue.
This what I have in debug log from when the song gets transcoded with
sox to when I seek and it errors out, jumping to the next in queue.
Code:
[20-06-29 15:51:18.4361]
I've recently bought a DAC (Topping D10) to connect to my headphone
amplifier. I'm using Beyerdynamic DT 880 and I'm quite pleased with
AutoEQ values for them.
Now, I used to use EqualizerAPO but that doesn't work with ASIO/WASAPI
and... being stupid that way I like the display on the DAC to
68 matches
Mail list logo
