A little test on how loud some players really are

[dfkt]

Despite the specs you can read on manufacturers sites I did a little test on how loud some players really go... the results differ quite a bit from the arbitrary mW specs claimed by the manufacturers.

Especially the Cowon D2 which is claimed to have one of the beefiest amps (74mW) isn't that loud. The other surprise is the Sansa Clip. The chipset maker AMS mentions 120mW of power into 16 Ohm for that specific SOC, but Sandisk seemed to limit the chip quite a lot.

Measurements are made with an Echo AudioFire4 line-in (-10dBV weighted) at nominal input volume, no pre-amp. Test file is pink noise, 16/44.1 WAV at 0dB. All players were used at their maximum volume, no EQ or other enhancements. And no silly EU-crippled firmwares either.

No load:

Cowon X5 (Rockbox +6dB):	-0.4dB
Cowon O2: 			-2.4dB
Cowon D2:			-3.9dB
Cowon X5:			-4.0dB
Cowon X5 (Rockbox 0dB):	 	-4.0dB
Cowon S9:			-5.0dB
Samsung YP-R0:			-5.0dB
Archos 32:			-5.2dB
Sansa Clip+ (Rockbox +6dB):	-6.0dB
Samsung YP-U1:			-6.8dB
Cowon X5 (Line Out):		-7.2dB
Sansa Clip (Rockbox Old +6dB):	-8.1dB
Sansa Clip:			-8.2dB
Sansa Clip+ (Rockbox 0dB):	-11.5dB
Sansa Clip (Rockbox Old 0dB):	-14.2dB
Nintendo DS (Moonshell):	-14.5dB


16 Ohm load:

Cowon X5 (Rockbox +6dB):	-2.5dB
Cowon O2: 			-4.5dB
Cowon X5 (Rockbox 0dB):		-6.7dB
Cowon X5:			-6.8dB
Cowon D2:			-6.8dB
Sansa Clip+ (Rockbox +6dB):	-6.9dB
Cowon S9:			-7.9dB
Samsung YP-R0:			-8.2dB
Sansa Clip:			-8.6dB
Sansa Clip (Rockbox Old +6dB):	-8.6dB
Samsung YP-U1:			-10.1dB
Sansa Clip+ (Rockbox 0dB):	-11.6dB
Sansa Clip (Rockbox Old 0dB):	-14.6dB
Archos 32:			-14.6dB
Nintendo DS (Moonshell):	-29.2dB

The Cowon O2, rated at 64mW, is the loudest player of them all (excluding Rockbox +6dB on the X5). The ancient Cowon X5, rated at 40mW, is actually more or less the same as the 74mW D2. And the Sansa Clip is rather quiet...

[Chop4life]

interesting.......

[saratoga]

Is this just driving the line in or did you add a load impedance too?

[samueljesusfreak]

VERY interesting, thanks. I'm sure the 90% of the DSL volume is straight-up hiss

EDIT: So, none of these players bring any distortion to the sound, even cranked? Or do I have a skewed understanding there?

[saratoga]

Originally Posted by samueljesusfreak

EDIT: So, none of these players bring any distortion to the sound, even cranked?

Depends on how much load you put on them:

http://rmaa.elektrokrishna.com/Compa...2048%20Ohm.htm
http://rmaa.elektrokrishna.com/Compa...ip_-_Loads.htm

Really amazing resource here:

http://rmaa.elektrokrishna.com/

[dfkt]

Originally Posted by saratoga

Is this just driving the line in or did you add a load impedance too?

I was just checking for line-out levels to an amp (nobody in their right mind would use 16 Ohm phones at maximum... I hope). But you're right, with loads there are some differences in the results. Added them to the first post.

Does Rockbox actually overdrive at +6dB, or is that still within limits - or depending on the specific players? IIRC; the THD went through the roof with +6dB at some Cowon D2 Rockbox measurements.

Originally Posted by saratoga

Really amazing resource here:

http://rmaa.elektrokrishna.com/

Thanks.

Originally Posted by samueljesusfreak

EDIT: So, none of these players bring any distortion to the sound, even cranked? Or do I have a skewed understanding there?

I only found the Cowon O2 to have distortion at maximum volume under load, but it's still rather negligible for real-life use: http://anythingbutipod.com/forum/showthread.php?t=37161

[The DarkSide]

The player I own w/the lowest perceived volume is the Zune 120, and the loudest is the Touch. The D2 is right in the middle, and has no distortion at all when maxed. It's the only player I really use for music,.........................period.

[saratoga]

Originally Posted by dfkt

I was just checking for line-out levels to an amp (nobody in their right mind would use 16 Ohm phones at maximum... I hope). But you're right, with loads there are some differences in the results. Added them to the first post.

Without the loads, you're measuring voltage. With the loads, you can calculate the power transfered. However, I'm a little surprised by those numbers. I was 90% sure 0dB in rockbox would give you 1 V, but if you're only getting -3.7dB below -10dBV, then we're really, really off. Are you certain those numbers aren't relative to something other then -10dBV?

Assuming those numbers are right, then the X5 at 0dB in rockbox is actually putting out -16.3dBV or 0.153 volts. At 16 ohms thats 1.5mW. Seems a bit low?

Originally Posted by dfkt

Does Rockbox actually overdrive at +6dB, or is that still within limits - or depending on the specific players? IIRC; the THD went through the roof with +6dB at some Cowon D2 Rockbox measurements.

I think you're +6dB above whatever the player's DAC defines as line level. I thought the standard was 0dBV == line level, but apparently not. Either way, the player probably doesn't have enough voltage to go past whatever it calls 0dB, so you'll start to clip quite badly (hence the THD).

[d_headshot]

I thought rockbox wasn't even supported for the clip?

[The DarkSide]

Originally Posted by d_headshot

I thought rockbox wasn't even supported for the clip?

A number of people here @ABI have been testing it for awhile. Seems they've been getting some great results too.

[dfkt]

Originally Posted by saratoga

Without the loads, you're measuring voltage. With the loads, you can calculate the power transfered. However, I'm a little surprised by those numbers. I was 90% sure 0dB in rockbox would give you 1 V, but if you're only getting -3.7dB below -10dBV, then we're really, really off. Are you certain those numbers aren't relative to something other then -10dBV?

Assuming those numbers are right, then the X5 at 0dB in rockbox is actually putting out -16.3dBV or 0.153 volts. At 16 ohms thats 1.5mW. Seems a bit low?

To make matters even more confusing, I've realized I did something stupid by using the auto-sense inputs on my sound card. Now I've used the proper weighted line-ins on the back of the card, and the results are even quieter (about 0.3-0.7dB average). I'm no EE (I'm an SE), I don't know much about the backgrounds (thanks for explaining, by the way) - but I can assure you that I used the inputs at -10dBV (not +4dBu), turned up to the maximum (AudioFire specs). The card runs over ASIO, bypassing all WDM stuff... Well, at least the differences between the players is a somewhat meaningful result.

Maybe pink noise is not a good test signal for these measurements?

Here's the new results (right column) vs. the ones from yesterday (left column):

No load:

Cowon X5 (Rockbox +6dB):	-0.0dB		-0.4dB
Cowon O2: 			-2.6dB		-2.4dB
Cowon X5:			-3.3dB		-4.0dB
Cowon D2:			-3.4dB		-3.9dB
Cowon X5 (Rockbox 0dB):		-3.7dB		-4.0dB
Cowon S9:			-4.3dB		-5.0dB
Samsung YP-U1:			-6.1dB		-6.8dB
Cowon X5 (Line Out):				-7.2dB
Sansa Clip (Rockbox +6dB):	-7.9dB		-8.1dB
Sansa Clip:			-9.1dB		-8.2dB
Nintendo DS (Moonshell):	-13.6dB		-14.5dB
Sansa Clip (Rockbox 0dB):	-14.3dB		-14.2dB


16 Ohm:

Cowon X5 (Rockbox +6dB):	-1.8dB		-2.5dB
Cowon O2: 			-4.3dB		-4.5dB
Cowon X5:			-4.5dB		-6.8dB
Cowon X5 (Rockbox 0dB):		-6.3dB		-6.7dB
Cowon D2:			-6.4dB		-6.8dB
Cowon S9:			-7.2dB		-7.9dB
Sansa Clip:			-8.0dB		-8.6dB
Sansa Clip (Rockbox +6dB):	-8.0dB		-8.6dB
Samsung YP-U1:			-9.9dB		-10.1dB
Sansa Clip (Rockbox 0dB):	-14.5dB		-14.6dB
Nintendo DS (Moonshell):	-28.4dB		-29.2dB
Cowon X5 (Line Out):				-32.6dB

[bascom]

Just so I understand, the players with the smallest negative numbers are the loudest, and the ones with the biggest negative numbers are the quietest, right? For example, the X5 and O2 are loud and the Clip is quiet?

[dfkt]

Exactly, loudest player on top of the list, quietest on bottom.

[saratoga]

Very interesting results. It seems the manufactorer's specs aren't worth very much then. I'll see if I can reproduce them using a calibrated oscilloscope this weekend if I get some free time in the lab.

On a side note: thought about testing any amps? We've got pretty detailed specs for the Fiios. Might be neat to see if they're made up too.

[dfkt]

I did test some amps performance with RMAA, but can't really test them for loudness - they're all too loud and clip when fed into the -10dBV input. I could only test them for relative volume compared to each other.

By the way, I tested some players again with white noise instead of pink. I thought the additional high frequency power was negligible, but white noise gives quite a bit louder results. Cowon O2 for example gives -2.4dB instead of -4.5dB, Sansa Clip gives -6.2dB instead of -8.6 under load.

[Albinoni]

Is the Creative Zen Vision M, Zen Touch and Zen Nomad Xtra louder than the Cowons. I have all those 3 above and their quite a loud player.

Interesting to see what the MW on the Zens are?

[dfkt]

Sorry, I don't have a single Creative product, and I will never get one.

[Albinoni]

Originally Posted by dfkt

Sorry, I don't have a single Creative product, and I will never get one.

Thats fair enough as everyone is entitled to their own opinion. I dont have a single Ipod and prob will never get one either, but I do love my Creatives and Cowons. Well thats my own personal opinion.

[ajai_dev]

I was about to req. one for ZEN and ZEN XFi.. ^^

Oh well my zen seems louder than my Sansa View and quieter than my S9

[nodirectionhome]

how loud are the modern Cowons, i9, J3 etc. ? about the same as the S9?