I just got an iRiver iFP-790 with 256MB and found out the hard way that it won't play Vorbis files with an actual bitrate less than 96kbps. This makes absolutely no sense, since the VBR files I've been playing with it have frames with bitrates all over the place.

So that leaves me

1. Encoding everything individually, tweaking the -q value for each song to get as close to 96kbps as possible without going under, takes forever but gives the best results;

2. Using ABR mode in the vicinity of 100kbps, which I'd rather not;

3. Using WMA, which I've always had issues with that wouldn't go away even at 160kbps, such as an irritating image steering effect similar to Pro-Logic decoding. (problem probably exists between keyboard and chair, but it's Microsoft anyway, which is reason enough to avoid it. )

4. Spoofing this this thing into thinking the file is > 96kbps somehow by toying around with the vorbis file itself, not likely to work since I can't find the exact bits to twiddle with a hex editor, and would be moot if it computes the real bitrate based on file size and sample count or some other means.

5. Sending feedback to iRiver.

I'd really like to be able to do something like #4 at the expense of having non-standard files provided #5 doesn't get results (probably won't....)

I'm happy with playback quality and especially happy with its low power consumption, since it uses a single AA at a time and doesn't requre the expense associated with Li-ion replacement when those crap out.

Any other ideas?

Iriver responds eventually.. witness that there is ogg support at all.

But is is frustrating to correspond..
User: Is there some good reason why I can't play q0 or q1 files?
Answer: Use a conversion program to make them into q3.

User: I would most want to play q0 files on a flash player, since (a) storage is limited, and (b) The shock resistant nature of a flash player is suited for use in a gym, where I wouldn't be listening critically.

Answer: You can play q=0 files on the hard drive model.

iriver players that support Vorbis can support it starting from q0, but not below. I've prepared an example 96 kbps file (use "save as..") encoded to q0,06 using aoTuVb3. The first part has the average bitrate about 40 kbps, so 96 kbps limit is imo unjustified.

godzilla525, can you please test the file on your player (to see if it's played back correctly)? My other question: Is the volume of Ogg Vorbis files significantly lower than volume of MP3 files as it was the case with earlier models (iFP-300)?

The sample file plays identically on both foobar2000 v0.8.3 and the iFP-790.

The firmware on the iFP is v1.28b UMS (US version) dated 11/29/04.

I don't notice a difference in volume level between Ogg and MP3, but I'll toss a 1kHz sine wave into each at 128kbps CBR and check the output.

1kHz 20sec sine generated with foobar2000 (tone://1000,20),

oggdropXPd v1.7.11 libvorbis 1.1.0 --> 304mV @ full volume, R-ch no load
lame 3.90.3 --alt-preset cbr 128 --> 566mV @ full volume, R-ch no load

Yup... that's a big difference...

You could use the -m option to set the minimum allowable bitrate when encoding.

eg. oggenc -q 4 -m 96 song.wav

Yeah - about 5.4 dB. Thank you for the test. It seems that the newest iFP players suffer from the same problems with Ogg Vorbis support as the older ones.


In case of iriver players this is not necessary (btw I couldn't ever get the bitrate management to work properly) - as my test shown up the bitrate can go as low as 40 kbps and the file is played back without any problems...

I think that with iRiver players, the problem is not the local bitrate, but the frame size.
Lower vorbis modes are using bigger window sizes with the current encoders. This is a problem (memory/computations) for players. If you encode at higher settings, and the bitrate is going low, that should not be a problem for the player.

Would that explain why the q-0.01 and below are not supported?

I have found some kind of replacement for tremor decoder which speeds up decoding process. I wonder if it is used in irivers and does it have something to do with q0 limit/low volume problem. Unfortunately I cannot download it (the server is down?).
I've also found a suggestion to iriver on their old website. Piotr mentions "13 bit sampling level resolution" - does it have something to do with "13-bit correct output" mentioned by Johannes Sandvall?
Thank you in advance for any information.

I have the IFP-799 for testing, and I've decided to return it.
I had a IFP-395 and I loved.
Instead this time I got a Jetaudio G3 1 Gig that can play Vorbis files up to q9.
I tried q2 and it sound transparent to me. Vorbis version (aoTuVB3_ArcherB10).

I also liked q0.54, while is not transparent it sounds much better than WMA for the same bit rate which sold me right away to the Vorbis format.
So I'm going to mass encode my collection to Ogg Vorbis at q2.
Is too bad that Iriver is not supporting Vorbis at low bitrates, especially for the flash players.

Cowon (aka Jetaudio a Korean firm as well) have had a good reputation in the past for high fidelity sound, but their players only supported MP3 files. Now they have really take care of that issue with this new generation of players.
It seems that Iriver is resting too much in their laurels, and Cowon seems to be cashing in that oprtunity.
So my advice unfortunately for you, is to get a Jetaudio G3 or the soon to be released I5 or stick with the IFP and MP3 files.
The only thing better on that IFP vs the G3 is the MP3 FM, Voice and Line-In encoding quality.
Hope this help.

I did some checking around on iRiver's worldwide support page... I thought DonP was being sarcastic at first, but their page does indeed mention converting files and shows a compatibility table with the HDD players supporting 32 to 500 kbps!

On their Taiwan forum (found via Google), someone claims the device that does all the work in the iFP-700 series is a Philips PNX0101, however, the spec sheet reveals that to be a USB 1.1 device, whereas the 790 is USB 2. The PNX0102 is identical to the '01 except for USB 2 support and 8Mbit program flash, twice that of the '01, so the chip in the 790 is probably the '02.

I cannot verify this as

1. My 00 Phillips screwdriver has gone MIA and none of my others fit, and
2. I don't think it would be a very good idea to void the warranty so soon.

In any case, both chips have 32KB ROM and 64KB SRAM onboard, in addition to 8KB of unified cache. The core is ARM7TDMI operating at 60MHz. Philips also has a PNX0105 designed for HDD players with a built in IDE interface, faster ARM926EJ 140MHz core w/MMU, and larger 16KB data/16KB instruction cache.

It is possible that Vorbis support, judging from the differences in chips designed for flash or HDD, might be hardware limited in the flash players, but I sincerely hope that's not the case.

In the meantime I've been using -q 3 which sounds very good most of the time and results in files that aren't that much larger than -q 2... though I would like to be able to use smaller files with it at some point in the future.