[ameliavega27]

 
try increasing the 'stereo width'
toggle the treble up slightly.

Tried it and sounds great! Thanks!
 
But why are some tracks just skips and plays the next song? When I tried to repeat the song, it skips again on the same part. How can I fix that?

 

[hakuzen]

  Aaah , I see..... Thanks A Ton . So the stock firmware gives you a slightly clearer percieved sound?
 
So , other than the reduced battery life and occasional instability, are there any other quirks using rockbox?

well, i don't know. stock firmware seems to do some highs roll-off. so for highs, i would prefer rockbox version.
i can NOT affirm stock fw is clearer (resolution) than rockbox version, based on the data i collected.
anyway, you get a dual boot when installing rockbox, so you can use whichever you like.
i'd use rockbox for 44.1 and 48KHz. it allows many useful features to compensate an slight worse sound quality, if any (i'm not sure stock fw is better).
for higher source resolutions, i'd boot with stock firmware.

 

[stellarelephant]

Hakuzen,
Fascinating. Thanks for the measurements! To me it looks like the ripples in the Rockbox frequency response are so slight that they would be inaudible to most. Still, flatter is better in my opinion. The high treble roll-off of the stock firmware is a surprise. I use EQ to correct my headphone response so I'll continue using Rockbox for 16-bit music regardless. But I agree that the stock firmware excels for getting the awesome dynamic range of 24-bit. I wonder what other factors may be causing some to notice differences in sound quality between firmwares, since your measurements make it look like Rockbox should be the one to sound more "airy" --opposite of what some hear?!?

 

[hakuzen]

Hakuzen,
Fascinating. Thanks for the measurements! To me it looks like the ripples in the Rockbox frequency response are so slight that they would be inaudible to most. Still, flatter is better in my opinion. The high treble roll-off of the stock firmware is a surprise. I use EQ to correct my headphone response so I'll continue using Rockbox for 16-bit music regardless. But I agree that the stock firmware excels for getting the awesome dynamic range of 24-bit. I wonder what other factors may be causing some to notice differences in sound quality between firmwares, since your measurements make it look like Rockbox should be the one to sound more "airy" --opposite of what some hear?!?

you are welcome. that is exactly what i thought, rockbox should sound airier and brighter, not darker, if any audible difference, after looking the frequency response graph.
the ripples have lower amplitude than 0.05dB (completely inaudible). and the highs roll-off of the stock version should be practically inaudible too (near -0.2dB at 13KHz).
so tonally, at 16bit 48KHz (or 44.1KHz), i can't believe rockbox sounds darker than stock.
 
about cleanness and resolution, we should look for noise and distortion differences, i guess, but got near identical results on this.
 
i should try to compare when playing higher resolutions. maybe the people who state a clear difference between stock and rockbox sound are using higher resolution sources. or maybe i'm missing something.

 

[stellarelephant]

I'm interested in your noise and distortion measurements too. How are you doing these measurements?

 

[saratoga]

I'm interested in your noise and distortion measurements too. How are you doing these measurements?

 
Those are done in RMAA using a cable splitter to plug in the headphones and the sound card line in at the same time.  

 

[hakuzen]

i've learned in this amazing thread:
http://www.head-fi.org/t/800208/the-rmaa-rightmark-audio-analyzer-source-and-audio-device-measurement-thread-measurements-in-post-3-tutorial-in-post-2
 
thanks to @HiFiChris, @Samueru Sama (who has got infinite patience teaching me electronic and audio basics), @castleofargh, @bartzky, etc., i'm stepping into this world.
 
rest of data i got in RMAA: (but this info depends very much of the conditions and gear used when you record; i don't know if it is accurate enough)
 

 
Noise level:                                                          Dynamic range:
 
 
THD:                                                                                   Intermodulation distortion + noise:
 
 
Stereo crosstalk:                                                  Intermodulation distortion + noise (swept frequencies):
 
 
Frequency response (swept sine):                           Frequency response (swept sine), more info, both are also identical
 
 
THD (swept frequencies):

 

[stellarelephant]

Hi Saratoga. Gotta pay homage: thanks for Rockbox! Since we have your attention, could you speculate on what may be causing a perceived difference in sound, since these measurements seem to show only negligible differences?

 

[vapman]

Guys take a closer look at the graphs.  It is not just the treble: on Rockbox the response is almost exacte on the 0dB line. But on the stock FW chart, there is a very minor dip before 50Hz and it starts picking up above 0dB after about 500-600Hz. It also cuts off the high treble a bit earlier than stock firmware.

 

[j0p3Y]

Hakuzen,
Fascinating. Thanks for the measurements! To me it looks like the ripples in the Rockbox frequency response are so slight that they would be inaudible to most. Still, flatter is better in my opinion. The high treble roll-off of the stock firmware is a surprise. I use EQ to correct my headphone response so I'll continue using Rockbox for 16-bit music regardless. But I agree that the stock firmware excels for getting the awesome dynamic range of 24-bit. I wonder what other factors may be causing some to notice differences in sound quality between firmwares, since your measurements make it look like Rockbox should be the one to sound more "airy" --opposite of what some hear?!?

damn, forgot about the downsampling... I only play 24 bit music, so Rockbox should be a no go for me. Stock also is, since I definitely hear that songs are played slower. Also gapless is something that's missed. Have to abandon X3 for sure.
Any alternatives by any chance?

 

[saratoga]

i've learned in this amazing thread:
http://www.head-fi.org/t/800208/the-rmaa-rightmark-audio-analyzer-source-and-audio-device-measurement-thread-measurements-in-post-3-tutorial-in-post-2

thanks to @HiFiChris
, @Samueru Sama
(who has got infinite patience teaching me electronic and audio basics), @castleofargh
, @bartzky
, etc., i'm stepping into this world.

rest of data i got in RMAA: (but this info depends very much of the conditions and gear used when you record; i don't know if it is accurate enough)




Noise level:                                                          Dynamic range:

 


THD:                                                                                   Intermodulation distortion + noise:

 


Stereo crosstalk:                                                  Intermodulation distortion + noise (swept frequencies):

 


Frequency response (swept sine):                           Frequency response (swept sine), more info, both are also identical

 


THD (swept frequencies):

Those numbers look like your sound card is pretty noisy. Have you tried a loopback of just the sound card to itself?

 

[saratoga]

Hi Saratoga. Gotta pay homage: thanks for Rockbox! Since we have your attention, could you speculate on what may be causing a perceived difference in sound, since these measurements seem to show only negligible differences?

From the rmaa results posted above, I'd guess that people are hearing the pitch error and misinterpreting it. But it's hard to say, I don't have an x3 and haven't personally tested it.

 

[SalfordBoys]

What software do I need to do these cool tests?

 

[hakuzen]

Those numbers look like your sound card is pretty noisy. Have you tried a loopback of just the sound card to itself?

yea, it probably is. i'm using the input of that soundcard (Realtek ALC888, because the other ones i own are even worse) to measure frequency responses (using 100% input volume in them, this could affect too). i guess it's not suitable for noise measures.
enclosed RMAA tests of the soundcard: loopback, and a iBasso DX90 to ALC888 line-in test (if you ever check them, and conclude that soundcard is definitely not suitable for noise measures, please let me know; i lack of knowledge to interpret the graphs):
   
frequency response graphs:
http://www.head-fi.org/t/800208/the-rmaa-rightmark-audio-analyzer-source-and-audio-device-measurement-thread-measurements-in-post-3-tutorial-in-post-2/105#post_13019304
 

  Guys take a closer look at the graphs.  It is not just the treble: on Rockbox the response is almost exacte on the 0dB line. But on the stock FW chart, there is a very minor dip before 50Hz and it starts picking up above 0dB after about 500-600Hz. It also cuts off the high treble a bit earlier than stock firmware.

i don't know if +0.05dB before 1KHz is significant (Cayin C5 amp has a slight ramp from -0.2dB@40Hz to 0dB@1KHz to +0.15dB@10KHz, and people say it's a bit bright colored).
in the frequency response graph, the stock fw cuts off the high treble earlier than rockbox fw.
 

What software do I need to do these cool tests?

http://www.head-fi.org/t/800208/the-rmaa-rightmark-audio-analyzer-source-and-audio-device-measurement-thread-measurements-in-post-3-tutorial-in-post-2

 

[saratoga]

If you want to do distortion or snr measurements with Rmaa you need a good low noise, preferably 24 bit ADC. High end onboard may work but you need to test.

Frequency tests will work on anything so you should be good there.