Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[nabsltd]

If you choose "gapless" in the options, this works for WAV and APE (Monkey's Audio) files, but not for MP3.  I don't use any other formats, so I don't know the status for them, but here is the summary for what MC9.x appears to do:

1. Read some data from music file.
2. Decompress into PCM if necessary.
3. Convert PCM to different bit size/sampling rate if set in options.
4. Add PCM to back of play buffer.

While this is happening, the playback thread continuously grabs data from the front of the play buffer, and sends it to the sound card.

The reason this doesn't result in gapless playback between two MP3 files is because the front of every MP3 contains some blank data, because the original reference encoder/decoder did this, and some decoders require it.  I'm pretty sure that the MP3 decoder in MC 9.x doesn't require it, but, even if it does, the problem occurs after the decode to PCM.

Turning on the "do not play silence" option sort of helps this issue, but it still isn't perfect, and raises quite a few other issues (like two songs that are supposed to have a gap between them won't).

So, could we have some way to "skip blank frames at the start of MP3 files"?

[Matt]

Can you make a simple sample of two mp3 files where some program does this better than MC?  If so, send it to matt @ jriver.com (no spaces) and we'll see what we can do.

Thanks!

[Mastiff]

I use aggressive crossfading for 0.1 second, and I can't hear any gaps between correctly coded MP3 files (Lame audiophile is my setting). All my live albums sounds like the should, and all "double tracks" are gapless.

[rocketsauce]

I saw something kind of interesting on this topic in a thread over at HA.

http://www.hydrogenaudio.org/index.php?showtopic=11661&hl=

zZzZzZz (foobar2000 developer) wrote:

All recent LAME versions write gapless playback info (amount of samples that need to be skipped at the beginning/end of file) into file header (unfortunately those values are incorrect in certain conditions). Currently, foobar2000 (only recent 0.7 betas to be exact) is the only player capable of using that info (even LAME --decode ignores padding). This process does not involve any kind of file analysis or guessing how many samples to ignore; all info needed to play the file gaplessly is stored in the header.
Of course, this header isn't a part of MP3 standard (just like Xing header storing VBR seeking table and length) and existing software/hardware players require modifications to make use of it.
Also, latest foobar2000 beta is capable of editing gapless playback info, so you can basically make any MP3 file (also not LAME-encoded one) play gaplessly as long as you know encoder delay (576 samples for all LAME versions as far as I know), and exact length of track (can be eg. extracted from cue sheet).

Rob

[rocketsauce]

And here's another thread that has some interesting new info on the topic:

http://www.hydrogenaudio.org/index.php?showtopic=7787&hl=

Rob

[nabsltd]

Can you make a simple sample of two mp3 files where some program does this better than MC?

Nothing does it better than MC.  MC is the only player I have found that really can do gapless with WAV files (compressed data is a joke with other players) without hogging the CPU.  For this, I applaud you.

But, this doesn't change the fact that "gapless" leaves a noticable gap when playing MP3 files, which is distinctly different from what happens when playing other files, including other compressed formats.

I use aggressive crossfading for 0.1 second, and I can't hear any gaps between correctly coded MP3 files (Lame audiophile is my setting).

Then, you aren't listening closely.  Gapless with APE or WAV on the first two tracks of "Sgt. Pepper" has a nice smooth transition.  With your aggressive crossfade setting, I hear a blip.

But, with MP3, your aggressive crossfade does reduce the gap compared to "gapless".  BTW, I use the "--nspsytune" variant recommended by some of the LAME developers:
-mj -q2 --vbr-new -V2 --athtype 3 --nspsytune --ns-sfb21 2 -X0 -Z

I saw something kind of interesting on this topic in a thread over at HA.

http://www.hydrogenaudio.org/index.php?showtopic=11661&hl=

Since LAME is the standard MP3 encoder recommended by the MC developers, it would seem to make sense that they could pay attention to the "gapless playback tags" that are put there.  This would completely solve my problem, and also allow all the various crossfade, etc., options to work exactly as they do on WAV files.

[Mastiff]

Yes, I'm listening closely. I just checked with the trasition to "With a Little Help From My Friends", and there's nothing wrong. It may be related to sound card or drivers, though.

[nabsltd]

Yes, I'm listening closely. I just checked with the trasition to "With a Little Help From My Friends", and there's nothing wrong. It may be related to sound card or drivers, though.

Unlikely, since my sound card isn't doing that much except feeding PCM out the optical output.

Try re-ripping those tracks straight to WAV (or APE) and compare "gapless" to your crossfading settings.  If you can't hear a difference, well, then I don't know what to say, because the output PCM is radically different.

[Doc_Holiday]

This isn't really a problem with MC9.  The problem is the MP3 format doesn't support true gapless playback.  It looks like the boys/girls making foobar have come up with a work around but I don't think that j.river needs to address this issue.  

If you want gapless playback then use a gapless encoder, not MP3. Plus I think the MP3 will chop off some of the song at the end to make it fit in to an MP3 frame.  So with that info missing you will never have gapless payback.

[nabsltd]

This isn't really a problem with MC9.  The problem is the MP3 format doesn't support true gapless playback.

Except that now there is a standard for how to deal with this.  LAME has been encoding the necessary data for a while, but the new standard should be supported by all reasonable encoders.

If you want gapless playback then use a gapless encoder, not MP3. Plus I think the MP3 will chop off some of the song at the end to make it fit in to an MP3 frame.

No good encoder will do this chop of the end.

I guess my problem is that I read the MC help file ("Gapless. Track changes are perfectly seamless. There is no pause between the end of one track and the beginning of the next. Useful on classical and other recordings where track boundaries don't always occur during silence.") and expected it to actually be accurate.

[Mastiff]

Or perhaps it may be that you expect Ferrari performance for Chevy prices?   In my opinion (sorry that it isn't humble, I'm just not a humble guy...) the version of gapless that MC gives will do for a very large majority of the customers. The best sound quality would of course be a Meridian or similar CD player but then again, that's anoter marked.  

[xen-uno]

If I had to choose between "this" or "that" I would pick "that"...

This = Gapless mp3 data tagging (as ref'd by nabsltd)

That = MC understanding (as in read/write) Replay Gain tagging (such as VorbisGain) other than it's own. It has been around alot longer (if not, it certainly is more utilitarian) and is a solid standard now. First come...first served.

10^-27

edit: Here's an interesting question...
Is tag based RG bypassed when you send PCM data straight out through the SP/DIF port?

nabsltd...I don't mean to dilute your post(s)

[nabsltd]

That = MC understanding (as in read/write) Replay Gain tagging (such as VorbisGain) other than it's own. It has been around alot longer (if not, it certainly is more utilitarian) and is a solid standard now. First come...first served.

nabsltd...I don't mean to dilute your post(s)

That's OK.  I'm all for more support of such older standards, because it would help MC live up to their slogan: "the one that plays it all".

On a related note, I found out that MC can't play MP4 (AAC) files without having Apple QuickTime installed.  Not having to install QuickTime, RealPlayer, etc., is one of the reasons I went to MC.  Oh, well.

If the LAME development team ever fixes the --nogap option so that all files have the correct time in the LAME tag (BTW, MC reads that tag just fine to find out the length of VBR files, so why can't it also read the gapless playback info?), that will solve my problem.  Files encoded with this option play back with absolutely no gap at all.

Or perhaps it may be that you expect Ferrari performance for Chevy prices?

Nope, I just want the player to do the things that the slogans and ads say it will.  Those slogans and ads get people to buy the program...it would be nice if they were 100% truthful.

I've paid my money, and now I'm just asking for some consideration for what I feel is a real flaw in the program.  I think the programmers should be thankful I haven't hauled out my real wish list, which includes such things as being able to mark files with "file X must always be played after this one, even in random play, even if file X isn't in the 'Now Playing' queue", to deal even better with the "two tracks on the CD, and I want to keep that information, but it's really one song" issue.

My biggest wish, though, is that since databases can no longer be merged, why can't I change the "Rating" field so it can accept a different range (like 1 to 10)?  Likewise, every field should be user-editable and deletable.

[Will_L.]

I think MC9 IS the Ferrari of Computer based Jukebox systems.

The Meridian's just a boring old cd player... no where near what goes on with our systems.

[Mastiff]

nabsltd, I find the other points on your wish list a lot more intersting, to be honest. I guess what you want to avoid is for instance track 1 and 2 on Sgt. Pepper not being played in sequence. or "The Hellion" and "Electric Eye" by Judas Priest (the first one is an instrumental that goes over in the second). Smart!  

Oh, and all advertising is general. I remember seing an ad for some high-end DLP projector a few months ago that said something about a perfect picture. My brother agreed completely...but I disagreed totally since no digital projector so far can be perfect - they don't show true blacks. So for him  - and proably for 98 % of the country's citizens - the ad was correct, for me it wasn't. But then I live with a close to 150 pound threeyed monster that has to be adjusted a lot...

Will_L., I disagree. If you want REAL Ferrari performance (which in my book is high end) you'll have to avoid a computer all together and listen to a "boring" CD player (not that I do - I do all my music listening both in my Chevy Suburban and the house via MC) like the Meridian or maybe TAG McLaren. Remember that for instance a Ferrari F-50 isn't made to be as versatile as possible, nor as comfortable to use as possible, it's made with one thought in mind: The best performance possible.  

[nabsltd]

nabsltd, I find the other points on your wish list a lot more intersting, to be honest. I guess what you want to avoid is for instance track 1 and 2 on Sgt. Pepper not being played in sequence. or "The Hellion" and "Electric Eye" by Judas Priest (the first one is an instrumental that goes over in the second). Smart!  

...but insanely hard to do.  It's on my wish list, but I don't ever see it happening with any player.

Likewise, we need an easy way to have one library support more than one user.  "Opinion" fields (rating, maybe even genre) need to be on a per-user basis.  Sure, I can copy my library, and then somebody else can make changes to those fields to suit them, but as soon as I import new media, it will have to be done for both libraries.

There's a million other things on my wish list that are much harder than gapless playback for MP3s.  The advice to go to another format would work, but then I'd have to have three copies of my music...the archival APE, the streamable with other things going on on the network OGG (or whatever), and MP3 to copy to a portable (instead of having to decode/encode every time).

Will_L., I disagree. If you want REAL Ferrari performance (which in my book is high end) you'll have to avoid a computer all together and listen to a "boring" CD player like the Meridian or maybe TAG McLaren.

A "high-end" standalone CD player can't come close to a PC for quality, if you are outputting digitally.  The absolute lack of jitter during the read means that only the link might have jitter.

Then, too, if you do D/A inside your PC with a good sound card, you probably have a better DAC that the one inside that CD player.  PC sound cards have an amazingly short time from development to store shelf, and can easily keep up with updates in technology without forcing you to pay a lot of money for the things that don't need updating: CD transport, remote control, etc.

Most of the audiophiles at AVS forum are shifting over to PC-based solutions not only for the versatility, but the quality.

[rocketsauce]

Likewise, we need an easy way to have one library support more than one user.  "Opinion" fields (rating, maybe even genre) need to be on a per-user basis.  Sure, I can copy my library, and then somebody else can make changes to those fields to suit them, but as soon as I import new media, it will have to be done for both libraries.

You could probably do something similar by just adding a few custom fields and then making separate view scheme groups with view schemes that use each persons corresponding custom fields.

Rob

[xen-uno]

> A "high-end" standalone CD player can't come close to a PC for quality, if you are outputting digitally

C'mon nab...don't exaggerate

There is no technical reason (or other) that jitter correction lies exclusively in the domain of computer CD-x drives. Component CD players can have it, too. In fact, the way that audio mfr's hype specs, I'm surprised it isn't plastered all over the player (maybe it's TOO technical sounding). Jitter occurs in all optical drives, whether standalone or otherwise, because they all use (basically) the same mechanism. ABX a high end component CD player with SP/DIF against the output of a CD-x drive (via the soundcard's SP/DIF's), running through the same DAC, amp, and speakers...at exactly the same volume level. You will NEVER hear a difference.

10^-27

[Will_L.]

So can we agree that once we go optical sound quality is really up to the amp/receiver (i.e the DAC) not the cd reader?

Then in a comparison with a Meridian CD Player vs. MC 9 with optical out sound quality isn't an issue? If so the Meridian doesn't stand a chance. It can't provide pornography like a computer can... nuff said.

[Will_L.]

Whoops on the Ferrari bit....

You're absolutely right a Ferrari F50 (screw it the F1 hitting the track this sunday) is built solely for performance.

And I say MC9 (and the computer jukebox system in general) out performs ANY cd player. (See note about porno - now that's performance!)

[Veazer]

...I think MC9 IS the Ferrari of Computer based Jukebox systems...

I think it is more like a 40 ft RV. Big, a little slow at times, but loaded with all the goodies and accessories.

[nabsltd]

There is no technical reason (or other) that jitter correction lies exclusively in the domain of computer CD-x drives. Component CD players can have it, too. In fact, the way that audio mfr's hype specs, I'm surprised it isn't plastered all over the player (maybe it's TOO technical sounding). Jitter occurs in all optical drives, whether standalone or otherwise, because they all use (basically) the same mechanism.

Who uses the CD-ROM drive as a source when using MC...not me.  Thus, no jitter, because the rip to hard drive didn't result in any jitter, because there is no way to store serial bit-timing errors on a hard drive.  Now, you can get erroneous reads from the CD-ROM drive, but that's not jitter.

Sure, the SP/DIF out can still have jitter, but that is true for all SP/DIF connections, whether a computer is involved or not.  Removing the possible jitter from the read of the source is a big help, though, and that's what an HTPC does far better than the most perfectly designed CD transport: zero jitter on read vs. some amount of jitter on read.

[hit_ny]

Gapless playback...

i use cue files and one big Mp3, works everywhere. Of course i cant see the tracks unless i use the mp3cue plugin in winamp'

Now if only MC would support cue files like it does APL ones.

so i tried an experiment, downloaded monkey's audio and used a tool included with it called makeapl on a cue sheet...created the apl files.

tried to play them in MC or even winamp, no joy. i suppose APLs expect to be pointing to an APE file rather than a mp3.

so MC team when can we see CUE sheet support ?

[hit_ny]

Bump...

not replying to my own post....some more thoughts on this cue file thing.

Was playing with APE and monkeys audio.  If you download monkeys audio from http://www.monkeysaudio.com/

It has a nice utility called makeAPL that swallows a cue file and spits out individual APL files, these are small binary files ( 2KB) that reference the tracks in the cue sheet to the big APE file. So if the cue has 10 tracks, makeapl creates 10 apl files.

Now MC can  import these APL files and can do audio analysis and all tagging is available. Its totally transparent whether the track is part of a big audio file or separate, which is the way things should be.

i was thinking would it be so difficult for the MC team to do the same with Mp3. Once a cue sheet was imported, could MC create MPL files ( or  whatever they decide to call them) like APL in the same directory as the cue and then import them as it does with APL.

i have a lots of albums ripped with cue sheets and am sure others on this board do too.

C'mon MC team, this would make MC handle all my audio needs. i cant recpommend MC to any ppl i know until it can do this. Would be a first too, cant imagine any tools out there able to do this.

What do ppl think ?

[Veazer]

Can you make a simple sample of two mp3 files where some program does this better than MC?  If so, send it to matt @ jriver.com (no spaces) and we'll see what we can do.

Thanks!

I can send you the files if you absolutely need, but I have a painfully slow connection here. Perhaps I can trim the beginning of song A and the end of song b so you can examine the transistion?

For the music I have ripped with EAC & Lame APS, the Winamp 2.91 default DirectSound output plugin seems to perform a near perfect gapless transition.  I think it works with the default settings, but just in case here are my settings for the v2.2.6 of out_ds.dll:

DEVICE: SigmaTel Audio, 'Create Primary Buffer' enabled
BUFFERING: All are set to 1000ms
FADING: Only 'Pause/Stop' and 'Seek" enabled, 500ms each (NOTE: Enabling start or stop fades disables gapless output!!!)
OTHER: 'Remove silence...' enabled, 'Volume control" enabled, 'Smooth volume..." enabled,  Volume control set to Linear, Logarithmic fades enabled

Try it with your own music first and see if it works for you. If not, let me know and I'll send you my samples.

Thanks Matt.

[hit_ny]

Are people happy with the gapless play of MC 9 ?

i find it works sometimes and otherwise not. SO i decided a long time ago to rip with cue sheets and one big file.

Gapless play on any player regradless of wheter the format or the player supports it.

Of course now, you cant see the tracks. SO its the limitation of the player that requires you to rip separately and hope the player can do it.

You can get gapless play with winamp, just forward the next trak buffer to buffer bout 300ms...and it seems to work. ( but only winamp ver 2.9 )

But if you use portables you are SOL.

i would like to see CUE file support in MC, there is a working version out there for APL. I dont think there are any technical issues to my knowledge that prevent this form being implemented in MC. If there are i would be very happy to know about it.

SO once again CUE FILE SUPPORT FOR MC 9..

[rocketsauce]

You could also use another lossy format (like MPC or Ogg Vorbis) that does not have the gapless playback limitations that the MP3 format has.

Rob

[hit_ny]

You could also use another lossy format (like MPC or Ogg Vorbis) that does not have the gapless playback limitations that the MP3 format has.

What if you decided to rip lots of stuff way before you ever heard of MC. So called legacy stuff. Cue files take care of gapless problems. This is one reason that mp3s will be with us for a long,long time to come.


What if you like to hear your stuff on the go ? Are there any portables that support the so called gapless formats. there might be in the future, but none right now i want to listen to stuff NOW.


Would i be expected to re-rip to something else. just to be able to use MC ? No, i could just continue to use winamp+mp3cue. This is a hard argument to overcome when trying to interest people to use MC for the first time.

Its even harder when the no one in JRiver even cares to comment on it nor commit to it, even though its been asked for almost a year. If players have cue support the issue of gapless is moot.


I'd expect any "winning" product to be able to cater to what i have rather than make me do more work. Making the transistion to a newer/better way of listening to music easier. Take away the reasons not to use MC.

[rocketsauce]

What if you like to hear your stuff on the go ? Are there any portables that support the so called gapless formats.

Are there any portables that support cue sheets? You were talking about using cue sheets for playback. You didn't mention using a portable, so I assumed you were talking about PC based playback where the audio format is pretty much a non-issue.

Would i be expected to re-rip to something else.

I don't know. All I was doing was suggesting a way to get gapless playback + track access using MC.

Rob

[hit_ny]

Are there any portables that support cue sheets? You were talking about using cue sheets for playback. You didn't mention using a portable, so I assumed you were talking about PC based playback where the audio format is pretty much a non-issue.

no portables that i am aware of handle cue files. but the playback is flawlessly gapless even on the cheapest of players, which is the point of ripping to a big file.


its a non-issue in terms of playback on MC yes.but you can't see what you are listening to. since MC has no way to understand track info that is written into the cue file.

But getting MC to display the cue file is only half the story, i am looking for APL like support for cue files.

MC's strongest point is the way it organisies stuff. with smartlists its easy to pick tracks you like and burn to cd-rw to listen later on portable cd-players. IT's easy to see what you have, many times the best tracks are in mixes so a way to get at them would be great.

Sometimes you want to hear certain tracks that are in a mix, but since you ripped as a whole file there is no way to access that particular track. i admit in this case gapless play is not really wanted. If you decide you want to listen to the track as it was in the mix, you have the whole  album ripped as one.

But the ability to make custom playlists of tracks embedded in one big file is very attractive, pretty much the way it does with APL currently.