1 (edited by kray 2019-07-07 07:01:04)

Topic: Up/oversampling on final output?

New here and 2 days in to my ADI-2 DAC and was searching but not sure I could find the answer. My other R2R DAC (and several other I've had) will up all signals to the max the DAC outputs too, i.e. 384khz for my R2R DAC.

Does the ADI-2 upsample the input a a max rate? or does it just apply filtering and outputs the same rate? i.e. passthrough?

for instance since the ADI-2 supports max rate of 768khz would it convert a 16/44 feed to 768 on output?

2

Re: Up/oversampling on final output?

There is no upsamling in our units. Usually you can do so in the player software, which might also give you many more options to fine-tune the upsampling than a hardware solution could offer.

Regards
Matthias Carstens
RME

Re: Up/oversampling on final output?

MC wrote:

There is no upsamling in our units. Usually you can do so in the player software, which might also give you many more options to fine-tune the upsampling than a hardware solution could offer.

Actually that's really good to hear. the ADI-2 does a great job feeding it 16/44, and I prefer that, and very nice to know whatever I feed it, it will stay at that rate, so upsampling in software would be the only stage of upsampling, no added one that can't be avoided.

4

Re: Up/oversampling on final output?

BTW, it is kHz, not MHz.

Regards
Matthias Carstens
RME

Re: Up/oversampling on final output?

MC wrote:

BTW, it is kHz, not MHz.

LOL.. ah yes, had my computer hat on at the moment... corrected

6 (edited by Curt962 2019-07-08 16:04:14)

Re: Up/oversampling on final output?

This Text "Rats Out" much of the BS.

https://www.audioholics.com/audio-techn … ital-audio

It's a lengthy read, but just understand that that "Up/Oversampling" can't Magically create something that wasn't there in the first place.  Common sense.

Vintage 2018 ADI-2 DAC. "Classic Edition"
Cables:  Red Ones, and White Ones.
Speakers:  Yes (2)

Re: Up/oversampling on final output?

Curt962 wrote:

This Text "Rats Out" much of the BS.

https://www.audioholics.com/audio-techn … ital-audio

It's a lengthy read, but just understand that that "Up/Oversampling" can't Magically create something that wasn't there in the first place.  Common sense.

Thanks Curt. Totally understand that, and ADI-2 does great job. Nice to know their's options to play around with.

8 (edited by Curt962 2019-07-08 16:33:20)

Re: Up/oversampling on final output?

Oh Kray!

The ADI-2 is a fabulous device, and  has delivered great Music Enjoyment on our end. 

A Fine Unit that quite likely will be here a LONG TIME.   smile

Vintage 2018 ADI-2 DAC. "Classic Edition"
Cables:  Red Ones, and White Ones.
Speakers:  Yes (2)

Re: Up/oversampling on final output?

The same or a very similar question has already been discussed here: https://www.forum.rme-audio.de/viewtopic.php?id=28810.

10 (edited by Curt962 2019-07-08 20:33:00)

Re: Up/oversampling on final output?

Kray.  Good question, and Thanks for Posting. 

I think the Up/Over Sampling confusion dates back to the Mid 80s, where Ad Copywiters had consumers convinced that the "Symphonic Bombast 2000" CD Player  with TWO TIMES (!!) Oversampling was delivering MORE than the others.

Even if it did...knowing that many Home Users have One Speaker next to the Garbage Bin...and the other atop the Kuhlschrank...

Wouldn't matter.  Still going to be an Audio Trainwreck. smile

Can't tell me otherwise.  I've seen some SUPER HIGH END GEAR set up as though done by drunken Chimpanzees...

Vintage 2018 ADI-2 DAC. "Classic Edition"
Cables:  Red Ones, and White Ones.
Speakers:  Yes (2)

11 (edited by orangejello 2019-08-07 19:59:17)

Re: Up/oversampling on final output?

I was looking at the Sound Setting on Windows 10, 1803 for the RME output device.  In the advanced tab I see options for choosing the sample rate conversion.  All of the choices (which are much less that I have seen for other devices) have a bit depth of 32.  On other devices I can see 16 bit, 24 bit and 32 bit depths.

Is something wrong or missing here?

BTW, I did not install any drivers when I hooked up the ADI-DAC2.  And I don't know anything about WASAPI vs ASIO either.  But it surprises me that I can only use 32 bit depth.

Also, does conversion to a higher bit depth have much effect on the CPU load for the conversion?  I would assume that a higher sampling rate would increase the CPU load as a function of the rate that is being set.

Re: Up/oversampling on final output?

Anybody?

13

Re: Up/oversampling on final output?

For what do you need 16 and 24 bit as audio data transfer selection? Both fit into a 32 bit word so you all you need is that...

Regards
Matthias Carstens
RME

Re: Up/oversampling on final output?

Agreed. No need. I was just wondering why this is the only output device that I have seen where other bits depths are not an option. I thought their might be a problem with the driver or the device.  But apparently not. Thanks

Re: Up/oversampling on final output?

MC, as there is no oversampling, what your manual mean by describing some reconstruction filters as "oversampling filters" on page 20 and 57?  Thanks!

16

Re: Up/oversampling on final output?

TomJ wrote:

MC, as there is no oversampling, what your manual mean by describing some reconstruction filters as "oversampling filters" on page 20 and 57?  Thanks!

MC said that there is no upsampling which is not the same as oversampling.

The oversampling is used to do some of the filtering in the digital domain.

Re: Up/oversampling on final output?

yes, understood...  thanks MC -
cheers, Tom

18 (edited by alexander.gruschka 2019-11-11 17:15:08)

Re: Up/oversampling on final output?

jiw wrote:
TomJ wrote:

MC, as there is no oversampling, what your manual mean by describing some reconstruction filters as "oversampling filters" on page 20 and 57?  Thanks!

MC said that there is no upsampling which is not the same as oversampling.

The oversampling is used to do some of the filtering in the digital domain.

Hmm, I'm far from being an expert, but I always thought that upsampling and oversampling are the same thing. The only difference is that upsampling uses a certain value like 96 kHz (for example 44.1 gets upsampled to 96) and oversampling uses multiples (2x, 4x, 8x) of the native value. I think the DAC has 8x oversampling (can MC confirm that?) and that would mean that 44.1 kHz material gets indeed "upsampled" to 352,8 kHz. Am I correct or not?

I am planning to buy the DAC and use it with a CD-transport with 44.1 kHz material only and I want to be sure that it gets the best out of my beloved shiny discs.

19 (edited by jiw 2019-11-11 23:15:47)

Re: Up/oversampling on final output?

alexander.gruschka wrote:

Hmm, I'm far from being an expert, but I always thought that upsampling and oversampling are the same thing. The only difference is that upsampling uses a certain value like 96 kHz (for example 44.1 gets upsampled to 96) and oversampling uses multiples (2x, 4x, 8x) of the native value. I think the DAC has 8x oversampling (can MC confirm that?) and that would mean that 44.1 kHz material gets indeed "upsampled" to 352,8 kHz. Am I correct or not?

I am planning to buy the DAC and use it with a CD-transport with 44.1 kHz material only and I want to be sure that it gets the best out of my beloved shiny discs.

Upsampling is the process of increasing the sample rate of signal. Oversampling is interpolation to a higher sample rate, i.e. it includes a low pass filter corresponding to the passband of the original sample rate. Thus oversampling is one kind of upsampling.

MC's post makes clear what he meant by it.

There is an 8x interpolator on the AK4490 chip used in the DAC (https://www.akm.com/global/en/products/ … /ak4490eq/, bottom). EDIT: However, the new revision uses AK4493 which also uses 8x interpolation (https://www.akm.com/content/dam/documen … tbrief.pdf).

The signal sampled at 44.1 kHz would thus be upsampled to have a sample rate of 352.8 kHz but a low pass filter removes some of the components above 22.05 kHz (the Nyquist-frequency of a signal sampled at 44.1 kHz). The ideal (read: theoretically correct) low pass filter would reject (infinitely attenuate) any components above 22.05 kHz while allowing those below to remain at unchanged amplitude, i.e. the brickwall filter at the Nyquist-frequency. However, this filter is impossible to implement so a compromise between rejection of components above the Nyquist frequency and attenuation of frequencies close to but below the Nyquist frequency must be found. This is where the several filters in the DAC come from.

The perfomance of the DAC's filters at 44.1 kHz is shown in the manual on page 56 (v1.8 EDIT: using AK4490) or 57 (v2.0 EDIT: using AK4493) and here EDIT: for the ADI-2 Pro FS using AK4490
https://1.bp.blogspot.com/-NxWV9PuKdJ4/W6XgqVY1LZI/AAAAAAAARis/BqONTzWt3dApVIcFq6Mgs8tg_vnHThsWACLcBGAs/s1600/Filters%2B1%2B%2526%2B2.png
https://4.bp.blogspot.com/-Tr89baX3fP0/W6Xgukkfr1I/AAAAAAAARiw/G8GTxm2HSMon-XUoVBIzBdD5U2eAo2UIQCLcBGAs/s1600/Filters%2B3%252C%2B4%2B%2526%2B5.png
Source: http://archimago.blogspot.com/2018/09/m … s-dac.html.

Manual v1.8: https://archiv.rme-audio.de/download/adi2dac_e.pdf
Manual v2.0: https://www.archiv.rme-audio.de/download/adi2dacr_e.pdf

Something seems to be in the works with that Short Delay Low Dispersion filter. It's not available with FW 33 on my DAC. More on this here: https://forum.rme-audio.de/viewtopic.php?id=29596.
EDIT: It's for a revised version of the DAC using an AK4493 rather than an AK4490.