Find out if there's something wrong with your computer source setup
Apr 30, 2004 at 7:36 AM Post #61 of 104
Wodgy, that may be true for chips (after c. mid 90s or so), but it surely is not true for software implementations
smily_headphones1.gif
 
Apr 30, 2004 at 8:19 AM Post #62 of 104
Quote:

Originally Posted by halcyon
Wodgy, that may be true for chips (after c. mid 90s or so), but it surely is not true for software implementations
smily_headphones1.gif



Not true. What do you think the SSRC window is doing? It's a windowed FIR filter. It would be nowhere near as processor intensive if they didn't filter at all.
 
Apr 30, 2004 at 8:31 AM Post #63 of 104
Wodgy or Halcyon,

if you don't want to upsample, what is better:

Keep the Resampler (SSRC) in DSP Manager and set it to 44.1 kHz or deactive it?

I use M-Audio Transit line out, ASIO, 24bit fixed-point. Previously I upsampled to 88.2 kHz, however in this case the testing sound isn't clean even the latency of the Transit is set to "very high".

Apart of that, Advanced Limiter and Dither cause strange noises when turned on.

Thanks for reply.
 
Apr 30, 2004 at 8:42 AM Post #64 of 104
With the Transit, I'd deactivate the resampler in DSP manager. If you're mostly listening to CDs and don't use any special DSPs or crossfeed or the equalizer, set the output to 16-bit fixed-point and turn off dither.

If you do use DSPs or things like that, set the output to 24-bit and turn on dither.
 
Apr 30, 2004 at 8:46 AM Post #65 of 104
Quote:

Originally Posted by Wodgy
Don't get me wrong, you definitely want slow mode. All I was saying was that even slow mode involves tradeoffs.


With a Chaintech AV-710, AD832AN powered CMoy, and HD 280s, I cannot tell any difference whatsoever with SSRC turned on or off, nor with Slow mode on or off. Nor does the Advanced Limiter change it any. There is 0 distortion as far as I can tell. No warbling (and I know what it sounds like, courtesy of the SB Live! residing one PCI slot below the Chaintech) in the least, nothing abnormal. Just a piercing 20KHz tone that causes headaches no matter what the volume level... yeesh. Don't do that more than is ever neccesary.

And Permonic, if you don't want to upsample, just remove it. I don't think SSRC would be doing anything from from 44.1 to 44.1, but it'd have to at least look at the file, and the more components in a signal path, the more chance for corruption.

That being said, I still am using 24/96 upsampling. Why? I really have no idea. Perhaps because I can't tell a difference, but with future, better equipment, I may be able to. It's not degrading the sound any, anyway.

One [stupid] question: why can't this be used with speakers? I was under the impression that tweeters were specifically designed to handle high frequencies. Why could a midrange driver in headphones handle it, but not a tweeter?


(-:Stephonovich:)
 
Apr 30, 2004 at 9:06 AM Post #66 of 104
If you can avoid 24/96 upsampling with SSRC, I would. As Halcyon points out, if you have high-enough resolution gear, subtle artifacts seem to be introduced. With the Chaintech card though, I dunno, 24/96 upsampling may be better. I'm not up to date on what drivers to use with that card to prevent it from resampling internally.

The danger is the tweeters overheating. Some cheaper tweeters are designed to handle high frequencies for short durations, but not continuously. With this sound sample, people may be tempted to turn it up, sending unusual amounts of energy to the tweeters. As far as I know, it's not the voice coils themselves that overheat, but the actual tweeter driver and the fluid coolant around it. Headphone drivers are made of different materials and I've never heard of one overheating. With huge amounts of power you can melt the voice coils, but that is a rarity; the drivers would be clipping far before.
 
Apr 30, 2004 at 9:12 AM Post #67 of 104
With the newest VIA drivers on the AV-710 and 2 channel hi-rez mode turned on, foobar (on kernel streaming) wouldn't play any files unless I resampled to 96000 specifically. Maybe it's 96k internally?
 
Apr 30, 2004 at 9:12 AM Post #68 of 104
Ok, thanks guys for the advice.

I can tell you that with ASIO and upsampling to 24/88.1 or 96 kHz I hear distortion. With 48 kHz it's Ok.

With Directsound there's no distortion in the test sound. Is it due to the fact that ASIO bypasses the Kmixer (I have Win XP) and Directsound not?
 
Apr 30, 2004 at 9:22 AM Post #69 of 104
Quote:

Originally Posted by Permonic
Ok, thanks guys for the advice.

I can tell you that with ASIO and upsampling to 24/88.1 or 96 kHz I hear distortion. With 48 kHz it's Ok.

With Directsound there's no distortion in the test sound. Is it due to the fact that ASIO bypasses the Kmixer (I have Win XP) and Directsound not?


You're making life harder for yourself than you need it to be. Start out with no upsampling and WaveOut. If you don't get distortion, you should stick with this. Believe it or not, the Sonica is bit-perfect with just basic waveout (I've verified this personally using a DTS decoder and PCM encoded 44.1kHz DTS files), and the Transit is mostly similar hardware to the Sonica but with different drivers.

If that didn't work, try no upsampling and DirectSound. If you don't get distortion, then stick with this.

Only if those two don't work, start fiddling with ASIO or Kernel Streaming and upsampling.
 
Apr 30, 2004 at 9:32 AM Post #70 of 104
Quote:

Originally Posted by Wodgy
You're making life harder for yourself than you need it to be. Start out with no upsampling and WaveOut. If you don't get distortion, you should stick with this. Believe it or not, the Sonica is bit-perfect with just basic waveout (I've verified this personally using a DTS decoder and PCM encoded 44.1kHz DTS files), and the Transit is mostly similar hardware to the Sonica but with different drivers.

If that didn't work, try no upsampling and DirectSound. If you don't get distortion, then stick with this.

Only if those two don't work, start fiddling with ASIO or Kernel Streaming and upsampling.


Wodgy, big thanks. I will follow your advice.

Just FYI, I have both Sonica and Transit, Sonica use AKM AK 4353 DAC while Transit has AKM 4584 Codec.
 
Apr 30, 2004 at 9:36 AM Post #71 of 104
Quote:

Originally Posted by CingKrab
With the newest VIA drivers on the AV-710 and 2 channel hi-rez mode turned on, foobar (on kernel streaming) wouldn't play any files unless I resampled to 96000 specifically. Maybe it's 96k internally?


I wouldn't know; I'm using the 1.43d drivers, as per a recommendation given in a recent thread that unfortunately, I can't find. The reason was that in the latest drivers, the volume control would occasionally mess up, or not adjust it equally among the left and right channels. Since there were no sonic improvements made, there's no reason to go with the latest. If you can find the thread (would be in Sources), there's a link to some place that still has them. In any case, no, I don't think it resamples to 96K internally. And I can tell you that with Kernel Streaming, I can get it play resampling from 48K to 96K. Using DirectSound or waveOut, of course, it plays anything.

Wodgy, the reason I'm using upsampling is because practically everyone who has one has recommended it. Granted, I may be yet one more sheep in a blind flock, but I figured go with the masses. My setup isn't extremely revealing, in any case, so I'm fine. But perhaps tomorrow I'll setup an an ABX with DiskWriter files that have been resampled to 24/96...

(-:Stephonovich:)
 
Apr 30, 2004 at 10:17 AM Post #72 of 104
I have a problem with my M-Audio Revo.

Using ASIO output I can only go down to 32bit fixed-point output and 44.1kHz. Using KernelS I can use 16bit fixed-point output and 44.1kHz.

The problem is that when using KernelS no other application can use the card, using ASIO output solves that problem but then I need to use 32bit output. Is it the M-Audio drivers that force the ASIO ouput to be 32bit, or is it the ASIO plug-in (0.38a SSE) in foobar2k?
 
Apr 30, 2004 at 11:02 AM Post #73 of 104
Quote:

Originally Posted by Wodgy
My personal recommendation for people with good, non-resampling hardware like the M-Audio cards is simply not to use upsampling.


Ehhm... not to be rude, but I didn't ask. Nor are you the one listening to and enjoying my setup, nor are you here judging what upsampling to 24/96 sounds like with my setup, and if it is/isn't an improvement over 44.1/16.

Sheesh... this board is getting HA-like (or maybe it's just this thread). Then in other threads, we have people replacing op-amps on soundcards and pronouncing huge sonic improvements... frankly, it's just weird. A person has to use their own brain/ears... I don't take anyone's advice on this matter.
Quote:

Originally Posted by Stephonovich
So take off Slow Mode. It does diddly crap anyway. Even the author says it's not noticeable unless you're doing 8000-48000 or something similar.

(-:Stephonovich:)


Nope, I'm using slow mode. In fact, I'm not changing a thing on my setup, I like it the way it sounds.
 
Apr 30, 2004 at 11:19 AM Post #75 of 104
Yes... the author of Foobar is a hardcore "ABX geek" (which means heavily biased toward "everything sounds the same") and one should keep that in mind. He may also have poor hearing -- being able to hear subtle sonic improvements with upsampling is not a prerequisite to be able to code audio playback programs.

Fwiw, he's also a total arsehole (I'd use stronger language, but it's not allowed).
 
You must log in or register to reply here.

Users who are viewing this thread

Total: 2 (members: 0, guests: 2)
Back
Top