09-04-2009 05:27 AM - edited 09-04-2009 07:35 AM
It looks like it did get improved over the original V1 clip. I measured a 1000 Hz tone as playing back at 1007 Hz, so +0.7%. Some others measured the same percent error by various means.
All the players I had available were consistent with your findings... battery powered ones on the order of 10x closer than (original) clip, and AC powered ones (CD and DVD players) another leap closer yet.
Message Edited by donp on 09-04-2009 10:35 AM
09-04-2009 09:21 AM
donp's results are from the v1 clip. User daytona955 also did some tests on the fuze v1 here. Both have the pitch bug. Any ipod would need to be tested in a similar way to what others and I have done.
09-04-2009 11:03 AM
In the interest of accuracy...yes, the Fuze did have it, but it was fixed with a firmware update....so Sansa does know how to fix it.
iPod Touch 5G 32GB, Touch 4G 32GB, Clip Sport 8GB.
Rockbox-> Clip Zip 4GB, iPod Nano 2G 4GB, iPod 5.5G 80GB
2012 Nexus 7 32GB, Asus MeMoPad 8 16+64GB, LG Optimus G Pro, Nokia Lumia 900, Nokia Lumia 520
09-04-2009 12:30 PM - edited 09-04-2009 12:32 PM
There's a thread under Fuze about this problem http://forums.sandisk.com/sansa/board/message?board.id=sansafuse&thread.id=11835&view=by_date_ascend... talk of Sansa's fix is around page 14.
As long as the sample frequency was 44 khz it didn't matter whether the format was ogg, flac, or mp3.
I recall that it was a lot better for 48 kHz so after a round of denial and "it doesn't matter" posts, a bunch were saying to just reencode your whole collection to 48 kHz. I don't know that anyone tested lower sample rates.
The clock frequency affecting pitch isn't the frequency of the crystal itself, but one derived from it using a phase locked loop. On the fuze, Sansa was able to choose a different PLL frequency with just a firmware change.
To be fair to those who rave about the clip's sound, the usual quantitative measures are things like frequency response, S/N ratio, and distortion, maybe time delay/stability between channels. Qualitiative might include soundstaging. It probably didn't occor to anyone to check the pitch (because, after all, "nobody" gets that wrong once you're past mechanical playback devices).
Message Edited by donp on 09-04-2009 03:32 PM
09-04-2009 01:40 PM
sansafix's comment in that thread (message here) shows the improvement, but I didn't think they implemented it in any fuze firmware.
Later in that same thread, mp3geek references from Wikipedia one study that showed the just disconcernable difference threshold of human hearing to be around 0.3% pitch error. This would suggest that for the majority of people, the clip+, at 0.247% pitch error, will qualitatively go unnoticed as having a pitch error. On the flip side, from an instrument tuning perspective, .247% pitch error is not acceptable, and gtune's vst plugin shows that the 1000hz tone on the clip+ is 'out of tune'.
09-04-2009 02:23 PM
Thanks for posting that link.
It has a lot of useful info about this problem and shows that a lot of other people are concerned about it.
It's a shame Sandisk won't respond to requests for info about a fix for the Clip.
I'm now sorry I recommended the Clip to my friends before I discovered this problem.
I've now had to admit I was mistaken and advise them to avoid Sansa Products until Sansa get their act together.
It's clear now that the Clip could never be sold in a music store with it's playback speed/pitch problem.
It's to bad, otherwise it would be a great musicians buddy.
Yes, it was an oversight to assume that the playback speed/pitch would be accurate, based on the fact that all other contemporary products are and no company in their right mind would release a digital audio product that wasn't.
Having said that.
Playback speed accuracy is a serious sound quality metric to me.
And so I stand by my statement that until the speed/pitch problem is fixed, the Clip and possibly the Clip + cannot be said to have excellent sound quality.
Come on Sansa software dev team, has Sansa management cut off your ?
Your customers are on this forum begging for info!
Where is your voice?
09-04-2009 03:36 PM
Hmmm... as the originator of this thread, I'm interested to see that there are reports of there being pitch bugs with the v1 Clip.
I have 2 v1 Clips -- a 2GB and a 4GB and don't hear problems with them (and hear no differences when running Rockbox on either of them). It was an 8GB v2 Clip that played back audibly too slowly to my ears.