Hi, recently some family members recieved Vorbis capable DAPs, namely a Samsung U2 and an iRiver T10. Of course i started to do some basic testing being curious what the results will be.

I started with the Samsung and Lancer SSE AoTuV b5 -q1.5 :
the U2 skipped all the files, however when i tried some Vorbis 1.1.2 encoded (-q1.5) ones, they worked.
(This is my father's player and he has no idea what ogg vorbis, so it will play just mp3s)

Second: The iRiver, with the same codec LancerSSE AoTuV b5 -q1.5:
i converted to ogg 8 albums (from FLAC), 5 of them played on the T10. The songs from the 3 remaining albums either didn't play at all or played only the first 1-5 seconds. Now the tricky part: all the 3 albums were by the band Sentenced Again, encoding with 1.1.2 solved the problem.
(This is my bro's dap and he would like to use as low bitrate as possible, so Vorbis is a clear preference)

So what's going on ? Is the AuTuV output format slightly different (what seems to me very hard to believe) ?
Is iRiver only for hip-hoppers, metalheads should look for another bands ? (I always thought iPod is the hip-hop gadget )

Any comments appreciated

did you try non-SSE builds of aoTuV?

iRivers tend to only support oggvorbis with bitrates between 96 to 225 kbps

No. But... how can the output format be different ? Although different encoders were used, the file format should remain the same, not ?
edit: nevermind, i found the aswer in the wiki

edit2: T10 seems to accept q-1 (45 kbps) too.
However those Sentenced albums doesn't work even with non-lancer aotuv, only with the 1.1.2. What can be so special about them ?

What Firmware version does your U2 use? I own one myself and never had a problem playing Ogg Vorbis, and I encode with Lancer's all the time. If you could post a sample that won't play on your U2, I'll test it on mine. I'm currently using version 1.350 of the korean firmware (found at http://cczclan.com/yp-55.asp)

I would try aotuv r1 (or maybe even older) and see if the problem remains for that band.

aotuv r1 makes the same problem, and today i found out that from the album Tiamat-Prey only two tracks play, the rest is skipped by the T10. I wonder what can be so special about these albums, maybe some VorbisComment tags or something ?

@MedO: The Samsung's firmware is 1.352 (should be the same as yours, just EU version). Currently i have no access to that player, i'll try to test further next week.

I think my player had this firmware too when I bought it, but I switched firmwares a few times because I wanted to try if the FM receiver that's present in the US version is only disabled in software . No success though, trying to use the receiver just crashes the device. But as I said, I never had a problem with ogg playback on the device, so I'd like to test this myself.

I just bought a SAMSUNG Z5 and it works fine with q3 - q9 (AoTuV b5). I would say it sounds much better than my old Nano with ALAC (however, no ABX). The new firmware even promises gapless playback ... unfortunately little clicks between two more quit ambient tracks (e.g.) are audible.
Actually no bad times Vorbis!!!

It seems that a specific player fails in play of the OggVorbis file containing large VorbisComment.
It's better not to use padding, when you use such a player.

I just tested my player (YP-U2) through with two different firmwares (korean 1.350 and european 1.355), which seemed to make no difference at all. I only tested one file though so it's not really saying all that much. I encoded "The Bloodhound Gang - Mope" at q-2,q-1,q0,q2,q4,q6 and q10, using Lancer's 20061103. The first two files didn't start, but everything above that played through the whole file. Q10 was "clicking" in a few places, probably the decoder couldn't keep up (no big surprise, the file is 471kbps average). I reencoded the q-2-File with aotuv b5, and it still didn't play. This aligns perfectly with the datasheet so far (they claim the device plays -q0 to -q9 files).

What is the difficulty with low-bitrate vorbis files? And is there a way (e.g. some encoder setting) to make them more "flawed implementation friendly"?

Edit: Added player type so you won't have to skim up to find it.

I own a iRiver T10 and I can tell you libvorbis 1.1 with Q0 works as good as Q1. On average 1 track per album is rejected. The ones that don't play I encode at Q1 or Q2. Q0 is 64kbps, Q1 is 80kbps and Q2 is 96kbps (nominal bitrates).

I encode in Q0 then play each file (just to see if it plays the first 1 or 2 seconds) and for the ones that don't play I use higher quality.

The manual sais the player supports Q1-Q10, so I was using Q1 before. I had about the same number of files rejected, 1 per album. I used Q2 for the rest and I never found a rejected file at that quality. Now I reencoded everything at Q0 and just kept about 10 songs over all in the previous quality.

I even got a CD with voice which is not stereo and I used -q0 --downmix and I got 48kbps nominal bitrate and it plays, it rejected only 1 track out of 17.

The firmware is 1.70, the player came with 1.60 ad I upgraded.

Looking at the list of my Vorbis files, i can't see connection between the VorbisComment size and the playabilty on the iRiver T10. Bitrate doesn't seem to be the issue too, because i was able play a 45 kbps file encoded with q -2. (I tried only one track though)

OK, finally i had time to play around with the Samsung U2. Stripping the VorbisComment from the files made them perfectly playable, so in the case of the Sammie the tag size was the issue.
I think i will exchange my Philips DAP for my father's U2...

I just noticed there is a new Firmware (2.16) for the U2 around. Installed and tested it, and downgraded again immediately: I didn't see any noticable improvements, but the ogg support seems to be broken. When playing the Fight Club soundtrack (encoded with aotuv b4 -q6) for checking, I got reproducable nasty-sounding crackling in several places (high-bitrate?) which is not present with the old firmware.

Edit: Just ran my ogg q-level test again with the new firmware, the result is exactly the same as last time, i.e. still doesn't play low bitrate files.

To the people who still use the problem player.

I did a trial to deal with this problem in the encoder side.
Please use new "venc".

It should have an effect on some players at least.

Thanks Aoyumi