Audirvana Plus 3 (official thread)

[Indydan]

I read that direct mode does not work with Mac OS Sierra with version 2.6 of Audirvana. Has this been fixed in version 3.05?

 

Can someone please confirm that direct mode in version 3.05 works with Sierra? 

 

Many thanks! 

[Indydan]

2 hours ago, jhwalker said:

 

Direct mode (i.e., Audirvana's custom code for bypassing the Mac OS) does not work any more - it hasn't for a couple of years (since Apple turned on system integrity protection on El Capitan in 2015).  Integer mode (which is the official method Apple allows for the shortest possible path to audio hardware) continues to work.

 

Since direct mode was developed originally as a workaround for Apple removing integer access a couple years earlier, direct mode is no longer necessary, though some people keep trying to resurrect it   I suspect contravening macOS security to patch in direct mode is not a priority for Damien - he can certainly speak for himself, though.

From reading the manual, I thought direct mode had to function for integer mode to work as well. So integer mode works even though direct mode doesn't? 

 

I imagine that the sound quality difference between direct mode VS integer mode must be minimal.

[Indydan]

2 hours ago, Jud said:

The story on direct mode as far as I understand:

 

- Direct mode was developed when Apple broke integer mode back in (IIRC) the Lion days.  It is no longer necessary for integer mode.

 

- It is essentially a driver, one that is even more stripped-down than Apple's own, which some people think contributes to better sound.

 

- Drivers interface with the kernel.  As of El Capitan and forward, Apple's System Integrity Protection (SIP) won't allow installation of such drivers without specific Apple permission, which Apple has not granted.  Damien has requested here several times for people to ask Apple (respectfully) to give permission for the Direct Mode driver in Audirvana Plus to be installed on MacOS, and he has also asked himself as a developer, but so far permission has not been forthcoming.

 

- There is a workaround to allow the installation, which is to patch the kernel, but people should not do this unless they understand the steps very precisely, have a printed copy of the instructions to hand, and have an up-to-date backup ready to install if something goes wrong.  This is not meant to scare people but to try to ensure against complaints of "I did just about what I thought you said and now my computer doesn't work!" from people who don't take these simple precautions.

Thanks for the information. I'll most likely update to Sierra (from Yosemite) since I know I won't be losing sound quality because of no direct mode. 

 

 

[Indydan]

I've had Audirvana 3 for about 10 days now. I downloaded the trial version and bought a license after 2 days. 

 

Previously, I had been using Bitperfect software with iTunes as the interface (Bitperfect does the processing, iTunes is only used for its interface and controls) for two years. I controlled my Mac Mini headless with the Apple remote app.

 

I compared both solutions: Bitperfect vs Audirvana on a good selection of music. Bitperfect sounds good, but Audirvana is better. On every piece of music I tried with both, each time the music played with Audirvana was better. Audirvana sounds smoother, less dry and brittle than Bitperfect. The sound with Bitperfect was also a little more closed in and close to the speakers.

 

It's not that Bitperfect sounds bad, it is just that Audirvana sounds better. I can't fault the sound quality of Bitperfect; especially for the price. But, the sound quality improvements with Audirvana are worth it for me. 

 

Also, I pushed the adjustments in Bitperfect to get the best sound quality. It did not always operate smoothly because of this. I would press to start an album, it would play a few seconds, then stop. It kind of stuttered every time I started a new playlist. This does not happen with Audirvana. It takes a second or two for it to load a playlist, then it starts playing without stuttering or stopping.

 

I also MUCH prefer the A+ app for iPad than Apple remote. The Apple remote gets the job done, but is somewhat rudimentary.

 

I am very happy to have discovered this fine product! 

[Indydan]

1 hour ago, Indydan said:

Thanks for the information. I'll most likely update to Sierra (from Yosemite) since I know I won't be losing sound quality because of no direct mode. 

 

 

 

I just updated my main iMac to Sierra (not the Mac I use for music) and I have been trying Audirvana on it. When I disable direct mode, the option of Mode 1 and Mode 2 for integer mode disappears. The integer mode box is still checked, but the option for mode 1 or mode 2 is gone. 

 

I am not sure what this means... 

 

On my Mac mini for music, running Yosemite, I have direct mode enabled as well as integer mode. I listened to both mode 1 and mode 2 in integer mode. I prefer mode 2. 

[Indydan]

34 minutes ago, audiocanyon said:

 

If you replaced the Sierra IOAudioFamily.kext file with the El Capitan  IOAudioFamily.kext file, I'm not sure why your integer mode choices disappear.  The Mode 1 and Mode 2 choices are still available to me on Sierra after doing the file replacement.  I prefer Mode 1 on my system.

I did not change the kext file. I was hoping to not have to do that.

[Indydan]

19 minutes ago, RunHomeSlow said:

 

Have you plugged the DAC into that MAC ?

No, I was justing testing it functionnaly, not for sound.  But, when I leave direct mode on, integer mode still offers modes 1 and 2 (but Audirvana won't play with direct mode on). If I uncheck the direct mode box, the choice of modes 1 and 2 for integer mode disappears.

 

With direct mode off, Integer mode seems to still be available, but locked into default mode (mode 1?).

[Indydan]

7 hours ago, MemoryPlayer said:

1) Yes, why not if I paid for a downgrade in sound quality?

2) Seriously, I like to enjoy when I pay for 3.90, 39.00, 390.00 or 3,900.00.

3) Yes with some softwares, but not with A+ 3.0!

It was in other words, a waste of money. 

 

The software comes with a FREE 15 day trial period. You were not forced to purchase this product, nor were you told falsehoods about what this product does.

 

If you used the free trial to test the software, and then bought it, then you made an informed purchase decision.

 

If you did not test the software using the FREE trial, and bought it blindly, you made an uninformed purchase decision. 

 

Either way, you are old enough to have a credit card and or a Paypal account. Thus, you are an adult who supposedly has the capacity to make decisions. You made one when you bought the product. 

 

Man up and ask yourself if your integrity is worth 39$

[Indydan]

I have started experimenting with the Izotope upsampling. From this thread:

 

https://www.computeraudiophile.com/forums/topic/29516-audirvana-how-to-select-izotope-settings/

 

I used one of copy_of_a's settings, because I liked the way he described the sound it produces. Here is a copy paste of the settings:

 

"proven" settings if you don't care about aliasing/artefacts above the nyquist limit (for whatever reason):

 

low pre-ringing "extreme" setting:

• phantastic instrumental seperation

• great transients

• super weak (actualy 'no') filter quality

steepness: 3

cutoff: 1.3

Anti-Aliasing: 50

pre-ringing: 0.36

 

Multiples of 2X upsampling. Everything is upsampled by powers of 2 to 176.4 or 192.

 

I used those settings, but ended up changing the cutoff to 1.0 and pre-ringing to 1. From what I understand, these settings will allow aliasing above the Nyquist limit. I sort of went with the rationale (as I understand it) of NOS DAC makers, that aliasing artefacts over the Nyquist limit will be filtered out by our ears (so no strong filter needed in DAC or software). 

 

I have listened to a lot of music with these settings. It sounds great to me! I am very sensitive and dislike brightness and brittleness in sound. If their are artefacts above the Nyquist frequency, I am not hearing them. The separation of instruments is better. Cymbals sound really natural (my brother plays the drums) with these settings. Shouldn't the aliasing above the Nyquist limit be producing downsides to the sound? I don't hear any.

 

Another question I have; my DAC ( Rega DAC R) has a choice of 3 filters. The filtering cannot be turned off. I have it on filter 1, which is a Linear phase soft knee filter at higher resolutions. would the DACs filtering be cleaning up the aliasing that the software is not filtering? If so, then maybe that is why I am not hearing anything bad in the music?

[Indydan]

On 3/23/2017 at 9:07 AM, Jud said:

So if you have a 44.1kHz file and you choose to upsample in A+ to the highest rate the DAC accepts or the nearest 2x rate (let's say 176.4 or 192kHz for a DAC that doesn't take DSD input), you've avoided a couple rounds of doubling.  But more important than that, you have the opportunity, for the initial and most critical stage of upsampling, to employ different and possibly "better" filtering (for some value of "better," whether it be particular measurements, your listening enjoyment, or both) in software than the filters programmed into the firmware in your DAC chip.

 

Wouldn't this purpose be defeated if the DAC has a filter (Linear, minimum phase, etc) you cannot turn off? The signal would be filtered twice?

 

[Indydan]

7 hours ago, Jud said:

 

Nope. 

 

In your situation, it may *reduce* the number of times filtering is applied.

 

It looks like the chip in the DAC-R may be limited to 192KHz, but I'm not certain of that.

 

Internally the DAC chip upsamples in "rounds" of doubling, so a 44.1KHz input will be doubled once, to 88.1KHz, then again to 176.4KHz, before being sent on to sigma-delta modulation.  On the other hand, if you upsample to 44.1KHz to 176.4 or 192KHz in A+, it happens in a single "round," and is then sent to the DAC-R, where (if it's correct that it uses 176.4/192 internal rates) it would bypass the DAC-R's internal upsampling and go through to the sigma-delta modulation.  So two rounds of upsampling internally versus one in A+.

 

Thanks Jud! That's good to know.

[Indydan]

2 hours ago, RunHomeSlow said:

 

i see that you didn't quote him till the end...

- cutoff 1.3 requires to reduce the gain by 1db to avoid too high intersample peaks. So if you set the gain by default for instance to -3db you should set it to -4db with this setting.

 

That is why I left the cutoff at 1.

 

- Too, as outlined above, IMHO settings with such a low filter steepness are better used in conjunction with an appropriate high quality lowpass (highcut) filter in the AU plugin section of A+ if you want to upsample on the fly (and if you do so better set Anti-Aliasing to 50). But this goes for all the settings listed in this second section...

 

It sounds great with a low steepness. I know that theoretically, there is aliasing above the Nyquist limit, but I cannot hear it.  

 

it is where he lost me... there is always something not easy to understand or one setting somewhere missing

 

Since i use upsampling DSD128 for now, i had choose from that thread, his settings:

16 -  200,000 - 1.00 - 200 - 0,72  for my DAC upsampling to DSD128, Filter C, -3db

 

[Indydan]

7 hours ago, Jud said:

 

Nope. 

 

In your situation, it may *reduce* the number of times filtering is applied.

 

It looks like the chip in the DAC-R may be limited to 192KHz, but I'm not certain of that.

 

Internally the DAC chip upsamples in "rounds" of doubling, so a 44.1KHz input will be doubled once, to 88.1KHz, then again to 176.4KHz, before being sent on to sigma-delta modulation.  On the other hand, if you upsample to 44.1KHz to 176.4 or 192KHz in A+, it happens in a single "round," and is then sent to the DAC-R, where (if it's correct that it uses 176.4/192 internal rates) it would bypass the DAC-R's internal upsampling and go through to the sigma-delta modulation.  So two rounds of upsampling internally versus one in A+.

 

There are two WM8742 chips in the Rega. There is a chart on page 25 of sampling frequency and clock speed. Is it possible to find if the limit of the DAC chip in the Rega is 192 with this chart?

 

 

 

WM8742_v4.3.pdf

[Indydan]

3 hours ago, Jud said:

 

You may think you are very sensitive to artifacts, but if you were, that would be unusual among most people.  We humans are quite famously insensitive to this sort of thing.  Yes, ultrasonic frequencies themselves may be filtered out by our lack of ability to hear them, but the problem is intermodulation: these ultrasonics interact among themselves and with audible frequencies to produce new frequencies within the audible range that aren't in the recording, i.e., intermodulation distortion.  NOS DACs have quite high amounts of it.  To a lesser extent you may be getting some of that too with those settings.  (The fact that NOS DACs have easily measurable intermodulation distortion and aren't rejected out of hand by listeners shows just how insensitive most people are to such distortion, even though it's in the audible range.)

 

Edit: I should qualify this - NOS DACs have high intermodulation distortion, *particularly when fed 44.1/48KHz rates*.  Oversampling should help.

 

I should clarify. I am very sensitive to brightness, brittleness and edginess. Usually this is because of the equipment used. I tend to prefer soft dome tweeters VS metallic tweeters (I am not saying metallic tweeters all sound bad). I also prefer a warm sound versus a colder sound. 

 

It is quite possible, and very probable I cannot hear the artifacts you mention. Just to be safe, I augmented the steepness from 3 to 12, and the anti aliasing from 50 to 85. 

 

This is still a weak filter. In either case, I cannot hear anything wrong in the music, and it sounds better than using no upsampling. I can't claim to hear a difference though between setting the steepness from 3 to 12, and anti aliasing from 50-85.

 

If most people cannot hear these artifacts, why do most DAC makers install strong filters? Sure they measure better, but if humans cannot hear these artifacts, why should we care? There seem to be benefits to using a weak filter.

[Indydan]

7 hours ago, Jud said:

 

That's the chart I looked at when saying that's what I'd guess, but I can't be certain.  I'm no expert in these things and don't know whether there were revisions I didn't see, different ways of implementing the chip, etc.

 

Edit: The easy way is just to look at what A+ says about your DAC capability.  Assuming it's telling the truth about itself, the DAC advertises its capability to A+ so that the player won't send it anything it can't handle.

 

Do you mean this panel in A+?

 

A+ screen.tiff

 

 

 

[Indydan]

7 hours ago, Jud said:

 

Do you know it's the "weak" aspect of the filter you like?  Aside from the Ayre paper (it's always dicey forming ideas from manufacturer white papers, which are part technical and part marketing material), could it be that you like the pre-ringing setting pushing some of that energy into post-ringing and possibly creating an ever so slight reverb/"presence" effect?  More important, how many different "strong" filters designed by good filter designers have you heard, so that you might have some idea if you liked those better?  What you get with A+ isn't so much different filters as it is one filter (albeit certainly one of the very best commercially available) with several almost infinitely adjustable parameters.

 

Those of us who aren't filter designers ourselves don't get a tremendous amount of listening experience with all the different sorts of filters that are available, so it's really difficult to come to sweeping conclusions that have any solidity on the basis of that limited experience.

 

You're right. What I am perceiving as great sound might be due to other factors. It could quite simply be the better upsampling by Izotope-Audirvana, versus the upsampling of my DAC.

 

I will experiment further with the settings before drawing conclusions.