Gapless playback?

I’ve been using Rockbox on Fuze v2 for a couple of weeks and haven’t encountered this sound issue.  I update regularly, sometimes building from svn and sometimes downloading a daily build.  At the moment I’m using yesterday’s daily build r26274.  The sound is great, at least the equal of the original firmware imo.  I’ve also been using Rockbox on a couple of old iRiver players, H140 & H340, for a few years.  I recall when the H140 got an iRiver port there were similar anxieties/assertions/questions about the sound quality.  Usually these are in terms that are hard to quantify and more importantly entirely subjective.  It’s impossible to do a reasonable comparison by booting from one firmware to the other.  For one thing the default volume levels are different and they must be identical in order to make any meaningful comparison (however sophisticated we like to think we are, inevitably in blind tests a small increase in volume leads to people identifying the exact same player as a different “better” player).  Perhaps the only way to actually make a decent comparison is to use the headphone socket as a line out and capture to hard disk the same audio from each firmware at measured identical levels, and then to blind test it i.e. abx on a PC.  Unless that kind of objective test is demonstrated, or there is some huge glaring error like distortion or clipping, then  it’s pretty reasonable to assume that these kinds of anxieties do indeed arise from placebo or psychological factors, not from any actual, measurable or demonstrable difference.

It’s also worth checking your settings.  For example is replay gain enabled on one but not the other?  Is the pre-amp level the same on both?  Is one firmware set to album gain and the other to track gain? 

@csavery wrote:

Now that Rockbox is quite usable on Fuze v2 I don’t think the pressure is on so much.

People have an alternative that provides gapless now. I’ve been using it for a couple weeks and it’s working well and provides a whole host of other features. It’s not official yet but it is working.

 

I particularly like using PictureFlow for selecting albums - much more visual with album art than the Sansa “little pic next to  album name that only shows when selected” idea.

csavery, THANK YOU!

I didn’t know this had happened. It seemed like a few months ago, there was NOTHING for us V2 Fuze owners. I just now downloaded the latest build today and gapless is AWESOME.

Thanks for pointing this out. I’ve now added Rockbox to my Twitter so that I can follow it’s development cycle.

 

@ Takla: ALL FX/EQ etc are disabled in both firmwares.  I did not make any adjustments to preamp level on either (so I suppose they could be different), and to be fair just adjusted the output volume on each to make it so it sounds the same volume to my ears. But it’s not just me that’s heard it.  I let a friend compare, and gave him no indication as to which should be better or worse.  He immediately confirmed that the Rockbox one sounds thinner and less smooth.

Coincidentally, after leaving the comment I did exactly as you suggested and recorded sound from both firmwares onto my PC (as a musician I have Cubase 4 software, so it was a doddle to get the sounds recorded for analysis).  The Rockbox is def WORSE!!! No doubt. Sounds a bit like a dither problem - I know from when I have mixed down my own tunes to 16-bit redbook format and forgotten to add dither that it does make the sound very rough.  But that’s still just a guess - it could be something else entirely.

Oh, I almost forgot to mention - the playback on Rockbox is a bit SLOWER too!!! I have not adjusted playback speed at all (it’s all disabled or set to default) but when you line up the tracks together (when recorded onto PC/Cubase from Sansa) the Rockbox one plays back noticeably slower!  Weird.

I will check to see if I have the latest RB release - I didn’t know they were updating it so regularly.

Message Edited by FLACtastic on 05-26-2010 01:55 PM

Message Edited by FLACtastic on 05-26-2010 02:00 PM

@flactastic wrote:

@ Takla: ALL FX/EQ etc are disabled in both firmwares.  I did not make any adjustments to preamp level on either (so I suppose they could be different), and to be fair just adjusted the output volume on each to make it so it sounds the same volume to my ears. But it’s not just me that’s heard it.  I let a friend compare, and gave him no indication as to which should be better or worse.  He immediately confirmed that the Rockbox one sounds thinner and less smooth.

Coincidentally, after leaving the comment I did exactly as you suggested and recorded sound from both firmwares onto my PC (as a musician I have Cubase 4 software, so it was a doddle to get the sounds recorded for analysis).  The Rockbox is def WORSE!!! No doubt. Sounds a bit like a dither problem - I know from when I have mixed down my own tunes to 16-bit redbook format and forgotten to add dither that it does make the sound very rough.  But that’s still just a guess - it could be something else entirely.

Oh, I almost forgot to mention - the playback on Rockbox is a bit SLOWER too!!! I have not adjusted playback speed at all (it’s all disabled or set to default) but when you line up the tracks together (when recorded onto PC/Cubase from Sansa) the Rockbox one plays back noticeably slower!  Weird.

 

I will check to see if I have the latest RB release - I didn’t know they were updating it so regularly.

 

 

Message Edited by FLACtastic on 05-26-2010 01:55 PM

Message Edited by FLACtastic on 05-26-2010 02:00 PM

That’s quite interesting.  If you want to check if dither is a factor then you may have noticed that dither can be enabled in the Rockbox settings meu.  It is disabled by default.  I suggest you have a look at the Rockbox Fuze manual which is downloadable html or pdf or available online at http://download.rockbox.org/daily/manual/rockbox-sansafuzev2/rockbox-build.html

btw if you are using replaygain you will find that the default pre-amp level is different between Sansa and Rockbox firmwares.

Re playback speed:  instead of just comparing Fuze original firmware to Rockbox firmware it would be very interesting to see the comparison made with the original compressed audio file on PC.

The best thing if you find deficiencies which can be substantiated (and it appears you may have) is to submit a bug report along with the data and details of your method to the Rockbox developers.  There is a Rockbox forum thread dedicated to the Fuze at SanDisk Sansa c200v2, m200v4, clipv1, clipv2, clip+, and fuzev2

Development is surprisingly rapid and if you identify a problem which others can reproduce then you can expect some serious effort will go into it and resolution(s) will follow shortly afterwards…it’s a lot quicker than waiting for a Sansa firmware update :wink:

@flactastic wrote:

@ Takla: ALL FX/EQ etc are disabled in both firmwares.  I did not make any adjustments to preamp level on either (so I suppose they could be different), and to be fair just adjusted the output volume on each to make it so it sounds the same volume to my ears. But it’s not just me that’s heard it.  I let a friend compare, and gave him no indication as to which should be better or worse.  He immediately confirmed that the Rockbox one sounds thinner and less smooth.

The issue I have appears to be specific to the Clipv2, so if you think theres some problem on the Fuzev2, could you run RMAA on Rockbox and original firmware and post the results?

@takla wrote:


@flactastic wrote:

 

Oh, I almost forgot to mention - the playback on Rockbox is a bit SLOWER too!!! I have not adjusted playback speed at all (it’s all disabled or set to default) but when you line up the tracks together (when recorded onto PC/Cubase from Sansa) the Rockbox one plays back noticeably slower!  Weird.

 

Re playback speed:  instead of just comparing Fuze original firmware to Rockbox firmware it would be very interesting to see the comparison made with the original compressed audio file on PC.

 

 

If you do some searching, there are several threads on the forum about Clip and Fuze (Sansa firmware) playing back at the wrong speed and pitch. How much and whitch way varies with model.  Clip V1 is ~0.7% fast.  So there’s a good chance that Rockbox is the correct speed.

Message Edited by donp on 05-27-2010 09:10 AM

Thanks for the info donp.  Pretty funny in a way…leads to people complaining Rockbox is faulty when actually it’s more likely to be working OK while the original firmware has the issue.  I think quite a lot of complaints like this (not just Rockbox, all kinds of things) are on the basis of difference, or deviation from what we got used to, without consideration of what should be happening.

Takla: “Pretty funny in a way…leads to people complaining Rockbox is faulty when actually it’s more likely to be working OK while the original firmware has the issue.  I think quite a lot of complaints like this (not just Rockbox, all kinds of things) are on the basis of difference, or deviation from what we got used to, without consideration of what should be happening.”

Pretty funny?  Complaining? Jeez! This isn’t about complaining.  I love the whole concept of Rockbox, and I also love my Fuze! For me this is about sharing issues and desires for improvement, simple as that. Rockbox is not “working ok” - YET!  I am sure they will get there. Thanks for the tip on reporting bugs, Takla - I do appreciate constructive advice, and will look into that. :slight_smile:

Anyway.  Issues for me right now at this time are:

a. It IS indeed slightly too slow (compared to both Sansa and to file on PC/Foobar)

I have now also tested the Sansa Fuze firmware against files on the PC.  Yes, the Sansa is very very VERY slightly fast - really only noticeable when played side-by-side with the source file (they slowly diverge as the Sansa gets ahead gradually - but this is quite a gradual divergence and there is no noticeable pitch difference). The Rockbox is slower, diverges almost immediately, and there is a noticeable pitch difference.

b. The sound quality is a bit thin and scratchy (I speculated earlier that this may be a dither fault - and yes I have the Dither turned on - but I stressed then, and still stress, this is merely a guess, so the cause could be just about anything for all I know).

Just to mention (in case it is relevant) - Any testing I have done has been on FLAC files (I don’t have any lossy files on my Fuze or HD).

saratoga "…could you run RMAA on Rockbox and original firmware and post the results? "

I would be more than happy to run any tests and post the results.  Until now I had never even heard of “RMAA” - but I have now DL’d the necessary software. Let me see what I can do.

UPDATE

I want to do the testing with my best audio leads (QED Reference Audio 1) so I have ordered some adapters (phono to jack) so I can get these into my soundcard’s breakout box (Delta 66).  I have only tested the speed using a cheap mono lead so far, but I will now test sound quality properly also.  Give me a few days to get the necessary adapters, or feel free to do some testing yourselves in the meantime! :wink:

Message Edited by FLACtastic on 05-27-2010 09:33 PM

Message Edited by FLACtastic on 05-27-2010 10:12 PM

Message Edited by FLACtastic on 05-27-2010 10:26 PM

@flactastic wrote:

I want to do the testing with my best audio leads (QED Reference Audio 1) so I have ordered some adapters (phono to jack) so I can get these into my soundcard’s breakout box (Delta 66).  I have only tested the speed using a cheap mono lead so far, but I will now test sound quality properly also.  

 Don’t bother using those silly cables, they won’t matter in this test.  The mono lead should be fine provided it is made out of some kind of metal, I’m not interested in the stereo measurements anyway.

@flactastic wrote:

Pretty funny?  Complaining? Jeez! This isn’t about complaining

 

b. The sound quality is a bit thin and scratchy

A complaint is a complaint.  A bug report is a complaint, formalised.  The word doesn’t have to be taken as being pejorative and I apologise if I annoyed you or upset you.

Sound quality: 

Can you abx it?  Can you identify “thin and scratchy” when you don’t know which file is which?  I believed I could tell moderate bitrate compressed audio from lossless until I did some abx tests.  The unmistakable, clear, indisputable differences mysteriously vanished as soon as I had no idea which file I was listening to.  If you’re using Windows then foobar2000 has an abx plug in which should allow you to perform abx testing quite easily and reliably.

If you adjust the playback speed in Rockbox so that it is corrected can the perceived sound differences still be abx’d?

What!  No Van den Hul cables available?  I guess the QEDs will simply have to suffice. Happy hunting!

Bob  :stuck_out_tongue:

Takla: "A complaint is a complaint.  A bug report is a complaint, formalised.  The word doesn’t have to be taken as being pejorative and I apologise if I annoyed you or upset you.

 

Sound quality: 

 

Can you abx it?  Can you identify “thin and scratchy” when you don’t know which file is which?  I believed I could tell moderate bitrate compressed audio from lossless until I did some abx tests.  The unmistakable, clear, indisputable differences mysteriously vanished as soon as I had no idea which file I was listening to.  If you’re using Windows then foobar2000 has an abx plug in which should allow you to perform abx testing quite easily and reliably.

 

If you adjust the playback speed in Rockbox so that it is corrected can the perceived sound differences still be abx’d?"

Man, you didn’t upset me - I just find you extremely patronising, that’s all. No, I’m not even going to attempt to explain to you what “thin and scratchy” means - whatever I say you will just argue with it. I have already said I will be performing an RMAA test, so save your opinions for when I post that. I am going to measure the distortions, and am going to use a good cable (or a “silly cable” as someone said) so that nothing else can influence the results. The answer to your last question is: yes.

@ donp: I was aware of historical issues with Sansa players and speed. This appears to have now been sufficiently fixed for the native firmware.

And regarding the pre-amp, (as far as I can see) it is only relevant to Replay Gain.  I never use RG as I generally listen to whole albums, not random or mixed up tracks.

As for pitch - I can’t for the life of me find it on any menus. I know it’s there somewhere, but so far it has alluded me.

Message Edited by FLACtastic on 05-28-2010 12:10 PM

Message Edited by FLACtastic on 05-28-2010 12:15 PM

FLACtastic wrote:

 

Man, you didn’t upset me - I just find you extremely patronising, that’s all. No, I’m not even going to attempt to explain to you what “thin and scratchy” means - whatever I say you will just argue with it. I have already said I will be performing an RMAA test, so save your opinions for when I post that. I am going to measure the distortions, and am going to use a good cable (or a “silly cable” as someone said)

 

And you sir, are being extremely rude (if I may interject my opinion before waiting for your precious RMAA test) to someone who’s only fault was attempting to help you and voicing an alternative point-of-view.

If your perpective and opinion is the only one that matters to you, why post on a public forum in the 1st place?

Thank you Tapeworm.

I made all my posts in good faith and have offered an apology in the same way.  I do think my apology was required as what I wrote was less than ideal.  If it can’t be accepted in good faith then that’s out of my hands.

I repeated the point about perception of sound quality for a couple of reasons:

1: terms like ‘thin and scratchy’ are entirely subjective and impossible to quantify and may mean different things to different people.  If different people have a different idea about what it means then maybe they will think they are dealing with the same issue when actually they each have in mind something else.  If a problem can’t be reliably identified and quantified how to tell if it is fixed?

2: The only sure way to demonstrate that there is a problem of this nature is to abx/blind test it, because the placebo effect is surprisingly powerful and none of us is immune. Some people feel very doubtful about this, some people get upset and feel that they simply cannot be affected and that to suggest otherwise is a personal insult. However placebo effect is well known, factual and measurable and used in all kinds of tests and trials from pharmaceuticals down to humble audio device and codec reviews.  I have to say I felt slightly wounded on discovering that most of the time I can’t hear the difference between a LAME V4 MP3 and flac, but there it is, a simple test shows my ears are not as golden as my ego, and that ears and mind in combination do not make for an objective tool.

Message Edited by Takla on 05-28-2010 04:28 PM

@takla wrote:

 

I repeated the point about perception of sound quality for a couple of reasons:

 

1: terms like ‘thin and scratchy’ are entirely subjective and impossible to quantify and may mean different things to different people.  If different people have a different idea about what it means then maybe they will think they are dealing with the same issue when actually they each have in mind something else.  If a problem can’t be reliably identified and quantified how to tell if it is fixed?

 

2: The only sure way to demonstrate that there is a problem of this nature is to abx/blind test it, because the placebo effect is surprisingly powerful and none of us is immune. Some people feel very doubtful about this, some people get upset and feel that they simply cannot be affected and that to suggest otherwise is a personal insult. However placebo effect is well known, factual and measurable and used in all kinds of tests and trials from pharmaceuticals down to humble audio device and codec reviews.  I have to say I felt slightly wounded on discovering that most of the time I can’t hear the difference between a LAME V4 MP3 and flac, but there it is, a simple test shows my ears are not as golden as my ego, and that ears and mind in combination do not make for an objective tool.

Thank you, Takla.  A few links for those individuals who still think they are immune from the placebo effect:

http://home.provide.net/~djcarlst/abx_plac.htm,

http://bruce.coppola.name/audio/Amp_Sound.pdf,

http://www.nousaine.com/pdfs/Wired%20Wisdom.pdf,

 

More information on ABXing:

http://en.wikipedia.org/wiki/ABX_test,

http://wiki.hydrogenaudio.org/index.php?title=ABX.

 

ABX is a type of double-blind test.

http://gizmodo.com/315250/pear-cable-chickens-out-of-1000000-challenge-we-search-for-answers.

 

Why anti-ABX?

http://en.wikipedia.org/wiki/Confirmation_bias,

http://en.wikipedia.org/wiki/Post-purchase_rationalization.

@flactastic wrote:

No, I’m not even going to attempt to explain to you what “thin and scratchy” means - whatever I say you will just argue with it. I have already said I will be performing an RMAA test, so save your opinions for when I post that. I am going to measure the distortions, and am going to use a good cable (or a “silly cable” as someone said) so that nothing else can influence the results. The answer to your last question is: yes.

 

 This test takes like 20 minutes.  Instead of arguing about nothing, you could just run the test and settle this already.

I apologise to all who I have offended.  Maybe I over-reacted.  Didn’t like being told not to bother with "“silly cables” and a couple of other comments riled me. My bad, I shouldn’t have reacted.

Yeah, the test is quick enough, I was just waiting for the adaptors so I can connect my silly cables. :wink: The only cable I had to hand was rubbish, so I have only really been able to verify speed differences, not distortions (the cable adds so much distortion and hum itself).

Oh - if anyone could kindly point me to where you change the pitch on the player (in Rockbox), I would be humbly grateful.

UPDATE

Anyway, I have all my cables in order now, and performed the test - or at least I tried to.  Unfortunately this was what I got:

“Cannot finish tests. \nAUDIO ERROR: Cannot detect test signal in the recorded waveform. Probably the input level is too low.  Try to check I/O levels.  No hardware error occured.”

When I did the pre-test, it was telling me the signal was about 18dB too low and suggested to try to increase the levels on the mixer. Unfortunately, I’m running everything at absolute max level (on my soundcard’s GUI and on the Sansa). It’s just that even at max level, the Sansa is very quiet (a bit louder using Rockbox but still very quiet).

I hasten to add, I have disabled any European-regulated volume limitations - and it is still very loud through headphones, but is simply not up to line level.

Good news is, I have a dock on order, which will enable me to get a line level from the player (not via the headphone output). This should enable me to complete the test.

Other good news is, I agree with those who mentioned about the possibility of placebo - I am starting to think this may be the case - none of us is immune to that. It will be good to run the test anyway, but I will not hesitate to admit if it has all been “in my head”!!! :stuck_out_tongue:

However, please let us wait for the test results.  I should have my dock in a few days - or someone else could run the test if they want.

Message Edited by FLACtastic on 05-30-2010 06:47 PM

Message Edited by FLACtastic on 05-30-2010 08:23 PM

@flactastic wrote:

 

Yeah, the test is quick enough, I was just waiting for the adaptors so I can connect my silly cables. :wink: The only cable I had to hand was rubbish, so I have only really been able to verify speed differences, not distortions (the cable adds so much distortion and hum itself).

 

 Cables don’t add hum and distortion.  If you have hum and distortion, its due to something else and you should look into that.

FLACtastic wrote: 

 

 

 

Anyway, I have all my cables in order now, and performed the test - or at least I tried to.  Unfortunately this was what I got:

 

“Cannot finish tests. \nAUDIO ERROR: Cannot detect test signal in the recorded waveform. Probably the input level is too low.  Try to check I/O levels.  No hardware error occured.”

 

 

Double check your PC recording settings.  You have set something wrong.  Even my laptop’s crappy onboard is more then sensitive enough for this test.  

FLACtastic wrote: 

 

Good news is, I have a dock on order, which will enable me to get a line level from the player (not via the headphone output). This should enable me to complete the test.

 

 

Thats not going to work.  Line level is roughly the same level as the headphone out, and even if it was louder I don’t think we have line out enabled in Rockbox yet.  Instead, figure out why your sound card isn’t working properly or else use a different PC.

Saratoga - first you mentioned about my “silly cables” and now you make false statements like “cables don’t add hum and distortion.”  YES they do.  Bad, old, broken cables do.

Stop making assumptions and barking orders (e.g. “You have set something wrong.” “Figure out why your soundcard isn’t working properly.” )

It’s very arrogant, and it’s rude.

Message Edited by FLACtastic on 06-01-2010 10:07 PM

@flactastic wrote:

Saratoga - first you mentioned about my “silly cables” and now you make false statements like “cables don’t add hum and distortion.”  YES they do.  Bad, old, broken cables do.

 

Stop making assumptions and barking orders (e.g. “You have set something wrong.” “Figure out why your soundcard isn’t working properly.” )

 

It’s very arrogant, and it’s rude.

 

 

FWIW,  if my Sansa is turned all the way up it is somewhat louder into my stereo than other line level devices like the tuner and CD player.  It seems very odd that it would be roughly 20 dB too low for a sound card looking for “line level” input.That’s a factor of 10 in voltage.

 

Message Edited by donp on 06-01-2010 10:04 PM