Do Kx drivers overcome this limitation?

seskanda

It's explained in this link:
How much is 24bit/96khz on all Audigy a fake? - Hydrogenaudio Forums

Basically, whenever you apply effects via the DSP to audio of ANY bitrate (i.e. 24bit/96Khz) it converts it to 16-bit, 48Khz. This happens on ALL Audigy-Audigy 4 series cards, i believe. Is this a limit of ONLY Audigy-based cards or is it of the EMU10K2 DSP itself?

Thanks in advance.

Maddogg6

Not that I know for absolute certain... but I am fairly confident in saying is a DSP limitation... see, 24/96 is *all* handled by an additional device (A2's have the P16V) that essentially 'bypasses' the DSP all together... so at the very least - its converted to 48 Khz when introduced into the kX DSP - tho I thought I seen some send >16 bit thru the DSP... ASIO in 20/48 does not work in my case - so if this is true or not - it would be useless for me, unless its usable with kX ASIO... (tho would be interesting to know exactly what happens with bit depths in that case)

So yes, I can feed 24/96 into the DSP - for certain - the effects are 48Khz as the DSP is locked at the sample frequency.
edit: but keep in mind - the original 24/96 is mixed at the outputs... so, I am thinking the only SRC is what is affected in DSP... tho I am not certain.

seskanda

So, is there any way i could convert the audio to 24/96, or ANY quality i need?

Doug W

If you are talking about converting the final wav file(s) to 24/96, R8Brain by Voxengo, (freeware), will convert to whatever you need.

Voxengo

Maddogg6

Im not sure what the point would be... up converting to 24/96 wont add any detail lost when recording at 16/48... ??

TravelRec.

Upconverting is nonsense, thatīs true!

__________________
Always take a look behind.

seskanda

That's true, it'll just bump up the size of your project quite a lot. So, if upsampling can't add the quality of 24/96 to 16/44.1(48K) audio, what can?

I'm curious, according to that link i posted, Audigy 2 cards and above playback ANY audio from 2 speakers at 24bit/192Khz, right? Does that mean it has to resample audio at another quality (e.g. 16/44, 24/96, etc.)? Or does it ONLY resample when you apply effects to the audio via the DSP? It also says there that "resampling still introduces noise." How noisy is it? Can it actually damage/distort the audio?

Russ

I am not an expert, but:

I do not think that anything is going to give you 24/96 quality from a 16/48 source.

Anytime that you go from one sample rate to another, re-sampling would need to be done, otherwise it would play at the wrong speed. Also, the 10k1/10k2 DSP still operates at 48 kHz, so any audio entering the DSP will still be re-sampled to 48 kHz, thus you could only get higher quality when bypassing the DSP.

I think that depends on the re-sampling algorithm used and who you ask, but generally re-sampling involves inserting zeros between samples, skipping samples, or a combination of both, so obviously you are going to lose some quality in the process (i.e. either you are throwing away information, or you are guessing at what the missing information might be). You can get good results, but you are not going to gain quality.

Maddogg6

When my card plays back with kx Wave HQ - it will bypass the kX DSP via the P16V device - and avoid forced SRC to 16/48. But with no kX DSP - it cant add effects.
UNLESS - I rout Wave HQ playback into the DSP with the P16V plugin - but the DSP effects are still only 48 Khz - and mixed AFTER the DSP with dry/un-affected Wave HQ - so, the dry is same (22-192Khz), but DSP effects are 48Khz.
Or so thats my understanding - I never bother with WaveHQ - its not fully implemented in kX and kinda useless imho.

seskanda

Fair enough, it's not easy to tell the difference between 16/44 and 24/96 anyway.


Does this happen during playback, recording, or both?


Does audio always enter the DSP, or ONLY when you apply effects to it via AudioHQ or KX DSP?



So, CL drivers don't get good results, then? Are Kx drivers any better at resampling?

Russ

Both (assuming that you are going from one sample rate to another in both cases).

On 10k1 models, all audio goes through the DSP. With 10k2 models you can bypass the DSP with WaveHQ, otherwise all audio goes through the DSP.

Again, not an expert, but I think the re-sampling is done in the hardware of the card, so it would be the same for both.

seskanda

So, if i were to just play a .wav file at 44.1k and another at 96k in a media player, resampling would occur, right? What about bit depth? If my Audigy 2 ZS card is locked at 24-bit, would there issues if i playback audio at 16-bit or 32-bit in my sequencer or im some media player?


Can you bypass the DSP in CL drivers by NOT applying effects to audio from AudioHQ, and instead use a VST/DX plugin?


Hmm...that's good to know. Because in that link i posted, it said the resampling of an X-fi card is somehow superior to its predecessors. Is that true? If yes, how does it's resampling fare with other brand name, top-notch sound cards?

Russ

Yes.

Yes, some type of conversion would be needed to change bit-depth as well, since the range of possible values is different for different bit depths.

Sorry, I have no idea what is possible with 10k2 cards with the CL driver (I do not have a 10k2 card, and it is not possible to bypass the DSP with 10k1 card models).

I do not know much about the X-FI, but I think it has native support for multiple sample rates, and has better re-sampling capabilities (any info I could give would be just info found on the internet, that you could find yourself).

alexs3d2

Using CL drivers, the only way to bypass the DSP is when using Creative's DVD Audio Player. Unter all other circumstances the card will resample at 48kHz.

As for the X-Fi, the DSP still works at 48 kHz, but it does resampling through its SSRC. This delivers upto 130+dB of SNR vs. the 80 of the previous architectures. In fact, the majority of the chips processing power is used for this conversion.

The X-Fi also allows another mode, where the DSP may be used on non 48-kHz content, by splitting the incoming signal. I guess that this means it may do upto 192kHz, since it has 4 effect slots, but then you only get 1 effect.

You may also completely disable the DSP and allow for the sound to pass without any conversion. (Music Creation mode??)