[DavidJr]

So.. any of you think there's any chance of rockbox being enhanced to support PlaysForSure?

 

[hekoman]

Quote:

Originally Posted by DavidJr So.. any of you think there's any chance of rockbox being enhanced to support PlaysForSure?

nope, never gonna happen.

 

[DavidJr]

Quote:

Originally Posted by hekoman nope, never gonna happen.

lol.. I pretty much figured that already...

 

[linuxstb]

Quote:

Originally Posted by dpippel I can deal with UI glitches/omissions, rough edges here and there, etc., but Rockbox on the 5G still sounds like $hit compared to Apple's stock firmware IMO. FLAC support doesn't mean squat to me if my MP3s sound better AND I can use Apple's superior UI. I'll be willing to play as soon as the bedrock issues (read SQ problems) are ironed out, but I think the iPod port of Rockbox has some miles to go yet.

Firstly, are you using a recent build of Rockbox? I fixed the initial "mono and distorted" audio bug in Rockbox on the 5G on the 20 February, and since then no-one has complained about the audio quality.

If your audio problems remain when using a current version, can you please provide more detail - preferably as a bug report on the Rockbox bug tracker:

http://www.rockbox.org/bugs/

Thanks.

 

[swmtnbiker]

Quote:

Originally Posted by linuxstb Firstly, are you using a recent build of Rockbox? I fixed the initial "mono and distorted" audio bug in Rockbox on the 5G on the 20 February, and since then no-one has complained about the audio quality. If your audio problems remain when using a current version, can you please provide more detail - preferably as a bug report on the Rockbox bug tracker: http://www.rockbox.org/bugs/ Thanks.

Thank you for inquiring! I installed the 3/1 build yesterday and things are improved over the previous build I was using, which could have been prior to your fix. However, I still find audio quality with the 5G/Rockbox lacking compared to the 5G running Apple's firmware. Rockbox is harsher and seems to clip occasionally on both FLAC and 256K VBR LAME MP3 files. I'm also finding that some themes cause the audio to pause and stutter. The comparison was made with no sound modifications and no EQ on a 30GB 5G running Apple's 1/10 firmware.

Anyway, I'm not criticizing what you folks are doing. To the contrary, I'm GRATEFUL that Rockbox is finally getting ported to the iPod and I appreciate all of the hard work that goes into it's continued development. Things are still going to be lacking and I have no problem with that. I'm a patient guy.

 

[RotAtoR]

Quote:

Originally Posted by greenleaves except that seeking in FLAC remains partially implemented.

Yes, right now with FLAC Rockbox only seeks to the nearest point in the seek table, which is placed every 10 seconds in the file by default I believe. What remains to be implemented after seeking to the nearest seek table entry is to decode to the actual seeked position. So seeking in FLAC works but not at full resolution.

 

[greenleaves]

Quote:

Originally Posted by linuxstb Firstly, are you using a recent build of Rockbox? I fixed the initial "mono and distorted" audio bug in Rockbox on the 5G on the 20 February, and since then no-one has complained about the audio quality. If your audio problems remain when using a current version, can you please provide more detail - preferably as a bug report on the Rockbox bug tracker: http://www.rockbox.org/bugs/ Thanks.

Welcome, linuxstb. It's great to see one of the Rockbox developers on the Head-Fi forum. First of all, many thanks for all the great work on Rockbox. Appreciate it very much.

According to one of the posters that said "The 5G lineout is variable under Rockbox. In other words the iPod's volume control affects the lineout volume." This raises a concern of the quality of the lineout because the signals still go through the iPod internal amp where we want to bypass for an external headphone amp. Is that the case?

In the case of the optical lineout in iRiver H-120, the volume control is bypassed. It would be nice to see a similar implementation in the iPod lineout. I am sure many of us here are keen to see such to happen too.

Thanks again ...

 

[linuxstb]

Quote:

Originally Posted by dpippel Thank you for inquiring! I installed the 3/1 build yesterday and things are improved over the previous build I was using, which could have been prior to your fix. However, I still find audio quality with the 5G/Rockbox lacking compared to the 5G running Apple's firmware. Rockbox is harsher and seems to clip occasionally on both FLAC and 256K VBR LAME MP3 files. I'm also finding that some themes cause the audio to pause and stutter. The comparison was made with no sound modifications and no EQ on a 30GB 5G running Apple's 1/10 firmware.

I'm not worried about the occasional pauses and stutters - this is simply because under certain circumstances Rockbox isn't decoding the audio fast enough to maintain real-time playback. This will improve as the various parts of Rockbox are optimised.

Please don't take this the wrong way, but saying that Rockbox is "harsher and seems to clip" doesn't help me debug... Does it definitely clip? If so, at what volume are you listening at, and are you listening via the line-out or headphone-out?

Also, are you able to perform a comparison of the two firmwares playing back the same WAV file? This would help rule out any codec-related bugs.

Does anyone else have a view on the current audio quality of Rockbox?

It wouldn't surprise me at all if there were bugs, but the more information you can give, the more chance we have of tracking them down.

Thanks.

 

[linuxstb]

Quote:

Originally Posted by greenleaves According to one of the posters that said "The 5G lineout is variable under Rockbox. In other words the iPod's volume control affects the lineout volume." This raises a concern of the quality of the lineout because the signals still go through the iPod internal amp where we want to bypass for an external headphone amp. Is that the case? In the case of the optical lineout in iRiver H-120, the volume control is bypassed. It would be nice to see a similar implementation in the iPod lineout. I am sure many of us here are keen to see such to happen too.

We don't have a datasheet for the specific DAC in the 5G, but so far it has appeared to be identical to the WM8983 - which has a datasheet is available here:

http://www.wolfsonmicro.com/uploads/.../en/WM8983.pdf

If you look at the Audio Signal Paths diagram on page 12, you will see how the line-out (LOUT2/ROUT2) are connected to DACs.

You will see that it is not possible to bypass LOUT2VOL and ROUT2VOL - all we can do is set the gain to 0dB, which is what you can do yourself by setting the overall Rockbox volume to 0dB.

If anyone is interested in helping make best use of this DAC, please read the datasheet, and let me know if you can spot any improvements which can be made to the way the DAC is initialised and used in Rockbox. The relevant Rockbox code can be seen here:

http://www.rockbox.org/viewcvs.cgi/f....c?view=markup

The audio driver for the 5g is still a work-in-progress, so please be gentle with me if you spot any mistakes... It's very possible that audio quality can be improved by initialising the DAC differently - I just haven't had the time to properly investigate it.

Thanks.

 

[muckshot]

Just want to chime in with some huge thanks for the Rockbox developers, nice to see you here at Head-fi... you guys kick some serious ass in my books

 

[swmtnbiker]

Quote:

Originally Posted by linuxstb I'm not worried about the occasional pauses and stutters - this is simply because under certain circumstances Rockbox isn't decoding the audio fast enough to maintain real-time playback. This will improve as the various parts of Rockbox are optimised.

Yep.

Quote:

Originally Posted by linuxstb Please don't take this the wrong way, but saying that Rockbox is "harsher and seems to clip" doesn't help me debug... Does it definitely clip? If so, at what volume are you listening at, and are you listening via the line-out or headphone-out?

Ha!! You're an engineer looking for absolutes on an audiophile forum where people routinely debate the sonic differences of the *wires* that connect their sources/amps/headphones/etc. That's a good one!

Hehe.

In all seriousness though, I would *love* to help you in your debugging effort. Unfortunately I just don't have the time to become part of the iPod Rockbox development project right now. So all I can do is give you my general impressions of the differences I'm perceiving between the Rockbox and Apple firmware. Hopefully some of the members here who DO have the spare time will be able to offer you some more substantial input.

 

[greenleaves]

Quote:

Originally Posted by linuxstb We don't have a datasheet for the specific DAC in the 5G, but so far it has appeared to be identical to the WM8983 - which has a datasheet is available here: http://www.wolfsonmicro.com/uploads/.../en/WM8983.pdf If you look at the Audio Signal Paths diagram on page 12, you will see how the line-out (LOUT2/ROUT2) are connected to DACs. You will see that it is not possible to bypass LOUT2VOL and ROUT2VOL - all we can do is set the gain to 0dB, which is what you can do yourself by setting the overall Rockbox volume to 0dB.

Thanks for the information. I did a few checks to confirm in the case of the iPod firmware that the volume control is actually bypassed in the line-out. It is not true, however, in the case of Rockbox.

Either the iPod firmware automatically turns off the volume control in the line-out or the actual DAC in 5G is different from what the WM8983 datasheet has shown or perhaps some other possibilities that could prop up. Surely further investigation is needed.

 

[K2Grey]

My thinking is that the default firmware just sets the gain to 0 for the lineout. After all, the firmware is software; it cannot physically bypass anything in the DAC chip.

 

[linuxstb]

Quote:

Originally Posted by greenleaves Thanks for the information. I did a few checks to confirm in the case of the iPod firmware that the volume control is actually bypassed in the line-out. It is not true, however, in the case of Rockbox. Either the iPod firmware automatically turns off the volume control in the line-out or the actual DAC in 5G is different from what the WM8983 datasheet has shown or perhaps some other possibilities that could prop up. Surely further investigation is needed.

The volume control for the headphone out and line out are controlled independently. The Apple firmware appears to fix the line out at 0dB and the user volume setting is only applied to the headphone out, and Rockbox adjusts both line-out and headphone-out together.

But I'm interested in feedback - should Rockbox fix the line-out at 0dB (as the Apple firmware appears to), or does anyone find it useful to be able to adjust the gain on the line-out?

 

[greenleaves]

Quote:

Originally Posted by linuxstb The volume control for the headphone out and line out are controlled independently. The Apple firmware appears to fix the line out at 0dB and the user volume setting is only applied to the headphone out, and Rockbox adjusts both line-out and headphone-out together. But I'm interested in feedback - should Rockbox fix the line-out at 0dB (as the Apple firmware appears to), or does anyone find it useful to be able to adjust the gain on the line-out?

The main reason for using the line-out is so that we can have clean signals directly coming from the DAC, bypassing the iPod internal amp, and all fine ampification is done by the external amp. Bypassing the internal amp for external control is the crux of the concern for all audiophiles in their setup.

I suggest that Rockbox should fix the line-out at 0db by default. It would be a nice option in the Rockbox setting should the user decide to use the volume control for the line-out.