Jump to content
IGNORED

iTunes Playing Various Sample Rates Natively On-The-Fly


Van G

Recommended Posts

I find that very difficult to accept, and so does NAD tech support.

 

Was this a reply to my post? But I can't help it, if you find it difficult to accept... :)

 

If people want to keep on using the internal rate converters, that' also fine by me.

 

Have a reference where you are pulling that from?

 

Measurements and hardware internals of all the ones I've seen.

 

Quite many more expensive ones for example are using TI's SRC4392 S/PDIF RX/TX+SRC combo chip programmed to rate-convert to fixed 96 kHz rate. I remember quite many "upsampling DACs" using the particular chip too, configured for 192 kHz, for example MuFi, IIRC. Cheaper devices use cheaper components.

 

Some others do it at the following DSP-processor sitting between S/PDIF RX and DACs.

 

Reason is that the internal software measures and calculates cross over and eq filters only for one fixed rate.

 

Older ones used 48 kHz internally, now many do 96 kHz and newest ones 192 kHz. For example my little bit older top-of-the-line Yamaha does 96 kHz and drops off all cross overs and eq's and enters "straight"-mode with 176.4/192 input rates because it doesn't down convert and doesn't have filters for those rates.

 

Edit: corrected the part number...

 

Signalyst - Developer of HQPlayer

Pulse & Fidelity - Software Defined Amplifiers

Link to comment

Well, I simply don't know, but I am sure someone else here will.

 

Regardless, I think your statement a bit disingenuous and still don't buy it, at least not as a general case. LFE, perhaps.

 

But if I am playing a 24/96k sample, the DAC is seeing the 96k sample rate - it is not dithered down to 48k before the DAC sees it.

 

-Paul

 

 

 

 

 

 

 

 

Anyone who considers protocol unimportant has never dealt with a cat DAC.

Robert A. Heinlein

Link to comment

Sorry - I don't buy it without more solid documentation

 

Naturally, you don't have to. I can have my laughs here with myself regarding AVRs and bit perfect data.

 

Well, I simply don't know, but I am sure someone else here will.

 

And the knowledge I have doesn't count? :)

 

I've studied quite a bunch of different devices. Not all, but many.

 

But if I am playing a 24/96k sample, the DAC is seeing the 96k sample rate - it is not dithered down to 48k before the DAC sees it.

 

Yes, with 96k sampling rate probably yes, unless the AVR is TrueHD/HDMA capable in which case it is possibly 192k. If you feed it with 44.1, 88.2 or 176.4 material and use crossovers(sub)/eq it likely gets resampled to one of those 48/96/192 rates used together with video.

 

Dithering doesn't have anything to do with sampling rate.

 

 

Signalyst - Developer of HQPlayer

Pulse & Fidelity - Software Defined Amplifiers

Link to comment

And the knowledge I have doesn't count? :)

 

Depends, from my point of view, you may be talking about something else.

 

I think you are assuming video, not audio. Even then, I think that does not apply to most AVR's built in the past five years. (Okay, quality AVR's. The $99 Insignia model is a different animal.)

 

Most decent AVR's strip the sound off HDMI and treat it as a pure audio signal anyway, and I do not believe a 44.1 sample rate is upsampled to 48k automatically, invisibly, and without any notice.

 

In plain old stereo mode, I don't think it ever does any upsampling or dithering unless for some reason, one would tell it to do so.

 

If you are talking specifically and in isolation about a digital crossover, then yes, I can see what you are saying.

 

but, you appear to be saying that any audio signal sent to an AVR at 24/88.2khz from a digital interface via S/PDIF is going to be downsampled and dithered to 16/48kz before it is played. This is at best, a misleading statement.

 

 

 

Anyone who considers protocol unimportant has never dealt with a cat DAC.

Robert A. Heinlein

Link to comment

Been reading on supplying the AVR via dig coaxial and have come across that only audio being supplied in 2CH PCM audio can be played to extended zones.

 

If I'm connecting the Win 7 to AVR via the mobo's dig coax, does that not mean the track is decoded on the PC and sent to the AVR?

 

 

I'll do a little searching but if you have a layman's answer/explanation please share.

 

Van G

Link to comment

I do not believe a 44.1 sample rate is upsampled to 48k automatically, invisibly, and without any notice.

 

No, it could be resampled to 48, 96 or 192, depending on device. Why would they tell you about it? Believing is not same as knowing.

 

Most normal consumers just don't care what happens inside the box as long as data goes in and audio comes out.

 

Most decent AVR's strip the sound off HDMI and treat it as a pure audio signal anyway

 

It's already ruined if it's coming over HDMI since it's such a crappy spec for audio to begin with. And no, AVRs can do a lot of DSP on the incoming data stream.

 

but, you appear to be saying that any audio signal sent to an AVR at 24/88.2khz from a digital interface via S/PDIF is going to be downsampled and dithered to 16/48kz before it is played. This is at best, a misleading statement.

 

No, I said it is typically converted to 48, 96 or 192 depending on device when cross-overs (for sub and/or main) and/or eq is in use.

 

You are free to think it as misleading based on your beliefs. I made the statement based on my studies on bunch of AVRs, including looking at schematics and ripping those apart.

 

Edit: And most likely that "resampling" (interpolation) is still better algorithmically than what mid-priced DAC chip manages on it's interpolation stage.

 

Signalyst - Developer of HQPlayer

Pulse & Fidelity - Software Defined Amplifiers

Link to comment

@Miska, if I'm following what you are saying right, is this not just the same resampling that 97% of DACs do at some point prior to the final "analogue" conversion stage?

 

Eloise

 

Eloise

---

...in my opinion / experience...

While I agree "Everything may matter" working out what actually affects the sound is a trickier thing.

And I agree "Trust your ears" but equally don't allow them to fool you - trust them with a bit of skepticism.

keep your mind open... But mind your brain doesn't fall out.

Link to comment

Then be clear and give facts rather than expecting people to take your word about statements that are controversial.

 

I gave you the benefit of the doubt, but when you make statements like:

 

Naturally, you don't have to. I can have my laughs here with myself regarding AVRs and bit perfect data.

 

you're credibility is seriously compromised.

 

 

Most normal consumers just don't care what happens inside the box as long as data goes in and audio comes out.

 

Do you know of any normal consumers here? I do not.

 

It's already ruined if it's coming over HDMI since it's such a crappy spec for audio to begin with. And no, AVRs can do a lot of DSP on the incoming data stream.

 

That's your opinion, but it sure doesn't make it a fact. HDMI transmits the same information as S/PDIF, and in a more robust manner. In fact, I would argue it is better. And yes, that is just my opinion as well.

 

No, I said it is typically converted to 48, 96 or 192 depending on device when cross-overs (for sub and/or main) and/or eq is in use.

 

Assuming this is true, that bit may make sense. Of course, what happens if tone controls are defeated and it doesn't run through a crossover? As Eloise pointed out, how is this different from the analog conversion stage of most DACs? Or pick a specific DAC- how about a WaveLength Proton?

 

 

 

 

Anyone who considers protocol unimportant has never dealt with a cat DAC.

Robert A. Heinlein

Link to comment

And in other news, I figured out the 2CH PCM stream required by the Denon to play in all zones is in fact what the media server will pass to the AVR.

 

Assuming there won't be a quality difference between getting the music to the AVR through dig coaxial or using DLNA across ethernet?

 

Van G

Link to comment

I would have edited / deleted that if I had thought twice about it, even though I wasn't "hot" when I wrote it. :)

 

Yes, PCM audio is pretty much what all the audio "formats" deliver.

 

What that note really means is that you cannot play 5.1 audio in the second zone - only stereo.

 

-Paul

 

 

Anyone who considers protocol unimportant has never dealt with a cat DAC.

Robert A. Heinlein

Link to comment

if I'm following what you are saying right, is this not just the same resampling that 97% of DACs do at some point prior to the final "analogue" conversion stage?

 

Well, sort of, it is done one way or the other. It's just about the work split. Some of the work still remains at the DAC, but it's sonic impact is smaller.

 

Reason for resampling in AVRs is to get all audio signals represented at one fixed sampling rate in order to simplify creation of cross-over filters for subwoofer and possibly main channels, as well as any (room) eq settings, delay/level settings for speakers and possible "dsp effects". And partially to accommodate algorithms for A/V sync with the video.

 

For example the Yamaha I have allows bypassing this stage with "straight" mode, but then it drops off subwoofers etc. for stereo listening.

 

Now let's assume this resampling is done with target rate of 96 kHz, this also eases up design of the following analog stages etc. And since it runs the DACs at that rate, the audio is "upsampled" to 96 kHz already. What follows is the standard interpolation practically every DAC these days does. So in a way it is similar to "upsampling DACs" but allows bunch of other processing too.

 

If the resampling is done using the TI chip I referred to, it is better quality most (any?) of the algorithms built into DAC chips. So it is not necessarily a bad thing to do. If one wants to bypass or minimize the impact of this stage, it is best to feed the AVR at it's native internal sampling rate. Rate converter of OS-X CoreAudio and that TI chip are on roughly same quality level. Biggest difference is that the TI chip is also ASRC (unlike CoreAudio one, AFAIK) used to compensate for jitter and I'm not sure how much I like it as approach. But smaller the incoming jitter, better the result as usual.

 

Depending on AVR's level it may have one of the good SRC chips or algorithms, or some worse one. For cheaper ones, the rate converter of CoreAudio would probably match or exceed the AVR one in terms of quality... If it has cheaper SRC approach, it probably also has cheaper DACs with even worse internal interpolation filter quality...

 

 

Signalyst - Developer of HQPlayer

Pulse & Fidelity - Software Defined Amplifiers

Link to comment

Assuming there won't be a quality difference between getting the music to the AVR through dig coaxial or using DLNA across ethernet?

 

Depending on particular hardware pieces, ethernet could be better, since it is asynchronous in a way that audio is clocked by the AVR in this case. Rest depends on how it is actually internally implemented.

 

Signalyst - Developer of HQPlayer

Pulse & Fidelity - Software Defined Amplifiers

Link to comment

Do you know of any normal consumers here? I do not.

 

AVRs are most likely NOT designed for readers of this forum, but for average consumers.

 

That's your opinion, but it sure doesn't make it a fact. HDMI transmits the same information as S/PDIF, and in a more robust manner. In fact, I would argue it is better.

 

Except that it is bound to video resolution and clock on the HDMI instead of having a proper independent audio clock. And it is sent as bursts without any better reliability than S/PDIF. Receiver side must reconstruct the audio clock from the video clock sent over the HDMI cable. And the audio you can send over HDMI depends on the video resolution, because the data bandwidth available for the audio is tied to the video bandwidth.

 

HDMI is just a DVI on steroids with audio slammed on top as an afterthought.

 

I've read through the HDMI spec couple of times, but I wrote this out of what I remember, if I made any errors, please refer to specific chapter in the spec for corrections... :)

 

As Eloise pointed out, how is this different from the analog conversion stage of most DACs?

 

It's not, so the same thing can be done already on computer. That was my point. And CoreAudio is equal or better on this than most DAC chips or those SRC chips/algos commonly used on AVRs.

 

In short: My recommendation for those using AVR or similar; try setting output format to 96/24 in Audio/MIDI Setup and keep it there. And give it open minded listen.

 

Signalyst - Developer of HQPlayer

Pulse & Fidelity - Software Defined Amplifiers

Link to comment

AVRs are most likely NOT designed for readers of this forum, but for average consumers.

I've been called many things in my days, but average hasn't been among them. (grin)

 

Obviously we disagree on that - unless you restrict it to what you can buy in Best Buy or online with Walmart. Amps are simpler, and in general, you spend less to get the same quality of sound with an amp- but sometimes an AVR is the right choice. Sometimes it is even the best choice.

 

Certainly the low end AVR I have here (A NAD T747) isn't going to compete with a $5000 amp. It does, however, sound as good as a NAD BEE325 Amp, which is what it replaced.

 

Personally, I will put that little amp up against any

Anyone who considers protocol unimportant has never dealt with a cat DAC.

Robert A. Heinlein

Link to comment

Paul & Miska, you both seem current, wondering if you could help determine if one AVR/Pre is better than the other in my price bracket?

 

Denon 3311ci - $806

Yamaha RX-A2000 - $875

Marantz AV7005 - $1100

Yamaha RX-A3000 - $1200

Denon 4311ci - $1150

 

Since I'm not using the amp portion of the I rationalized that I didn't need to buy the higher models of Denon or Yamaha since the features I was going to use were all in the lower model. There is the argument that the higher end models are built in Japan (vs. China) which I'm not sure I put value in.

 

The Marantz is at the absolute top end of my price range for this purchase but I'm not sure it is justified considering the spec seems similar to the Denon models, only in a pre-amp format.

 

Am I over thinking this?

 

Van G

Link to comment

AVR/Pre is better than the other in my price bracket

 

Hard to say how much you would gain. Maybe the biggest gain would be balanced outputs with pre/pro.

 

Marantz AV7005

 

I consider this fairly decent version. AFAIK, it doesn't support DSD over HDMI though...

 

In Europe Onkyo has PR-SC5508 which I find interesting also in terms of DSD.

 

 

Signalyst - Developer of HQPlayer

Pulse & Fidelity - Software Defined Amplifiers

Link to comment

Perhaps a little. :)

 

Out of that list, I would probably go with the Denon 3311CI, though that is a mighty nice price you have there for it! (It's more in the $1200 range I thought.)

 

I don't remember the reason you don't want to use a built in amplifier, but if that is really the case, the input specs of any of those receivers are not that much different than consumer grade machines, like the Pioneer VSX 1020-K, which is in the under $400 range.

 

I know that sounds like blasphemy. ;)

 

Me? I would go with something like one of the NAD receivers, they always sound good, and have very conservative power ratings. :)

 

-Paul

 

 

Anyone who considers protocol unimportant has never dealt with a cat DAC.

Robert A. Heinlein

Link to comment

I have one of the Emotiva UPA-7 (unbalnaced outs) amp on order which from what I've read will give be more headroom across all channels at 125 W as opposed to the AVR amp's which are rated at 125 W driving 2 channels but take a plunge when you're talking about 4 or 7 channels.

 

What I want from a pre or AVR this time around is the network programming/features, quality audio and video processing, and good sound.

 

I know the amp is a better investment than the AVR/Pre and that the latter depreciates very quickly.

 

I think of this as a stepping stone and would like to put dollars in the right place so if I can get by spending less on an AVR becuase it has less power but equal processing/features, then I'm interested.

 

I would also consider spending more on the Emotiva Amp (XPA series) if that money is worth spending.

 

Van G

Link to comment

That sounds like clear thinking to me.

 

Here is what I would do, I would print out the specs for a NAD T175 A/V Pre-amp.

Then compare that to the lowest end AVR that would meet your needs. (The little Pioneer is what I just used.) Make a list of the differences between them that are

most important to you, and rank them.

 

Then do a point by point comparison against your list of choices. When I did just that, they all came out pretty good, but with my list, I would buy the Pioneer.

 

Actually, I would use a credit card and buy the T175, but that's just because I know not much less is going to satisfy me, and why waste the time and effort installing, tearing down, and returning / reselling product when I know I am going to wind up spending $3k and buying a T175 anyway?

 

At least, that is what my wife says. :)

 

-Paul

 

 

 

Anyone who considers protocol unimportant has never dealt with a cat DAC.

Robert A. Heinlein

Link to comment

I think the Denon 3311ci works for me in the mid term:

- network features and iOS/Android App to control AVR

- extended zones can play 2Ch PCM while main watchs TV or Movie

- low enough price point that selling down the road won't be a substantial bottom line hit

 

The Marantz and NAD pre's don't make work for my situation although being better products.

 

Likely, my next two steps will be:

- speaker upgrade in extended zones

- changing extended zones to dedicated kits with DAC, Music Server

 

Thanks for all the help to date.

 

Van G

Link to comment

I'm being told that the audio performance of the 4311 vs the 3311 is substantial due to the fact that the 4311 uses the Audyssey MultEQ XT32.

 

From my understanding this is a room calibration tool that doesn't apply to 2 Channel playback.

 

If I look at Miska's point regarding the AVR mucking the file, room eq is actually a detriment to fidelity.

 

There is a $350 price difference between the 4311 and 3311.

 

Worth the expense?

 

Van G

Link to comment

From my understanding this is a room calibration tool that doesn't apply to 2 Channel playback.

 

It should be possible to use it also for stereo playback. At least it has been the case for all AVRs I've used.

 

If I look at Miska's point regarding the AVR mucking the file, room eq is actually a detriment to fidelity.

 

No, not at all! I'm all for using room eq. What I mean is that blindly running after bit perfectness may actually be detrimental to fidelity... In many cases choosing processing to be done where it can be done at best possible quality makes sense.

 

In case of AVRs, it often makes sense to feed it data at it's native rate - at the rate it uses for the room eq and other processing. Then it doesn't have to apply conversion and the conversion applied at the computer side has potential of being higher quality. AVRs are usually optimized for sampling rates used together with video (48/96/192), so using those rates may be beneficial.

 

What I'm strongly opposing is that what ever hardware is used to receive-and-DA the data would perform the best when fed at what ever happens to be the data stored at computer. I could even go as far as claiming this not being true and claim that any particular piece of hardware has one single rate and word length it performs best at(*). Processing can improve fidelity since after all, fidelity is realized as the audio that enters listener's ears, not as bits at some intermediate stage.

 

 

*) For example, in most cases the same analog filter stage after the DAC is used regardless of the input sampling rate. For optimum results, there would need to be different stages for 44.1, 48, 88.2, 96, 176.4 and 192 kHz. Having six switchable analog stages is more complex and expensive, than having six different parameters for software in digital domain.

 

 

Signalyst - Developer of HQPlayer

Pulse & Fidelity - Software Defined Amplifiers

Link to comment

Okay, so assuming I go with Denon 3311 or 4311, stream audio content to it through the Win 7 PC box via Dig Coax, and use iTunes as the Media Server/Reneder:

 

- how do I find out what the native resolution of the Denon is?

 

- how to best process the audio on the PC side that is not the native resolution for the Denon, ie: Denon is 96/24 native and selected music file is 44.1/16?

 

This is all in the context of using 'Remote App' to control iTunes from extended zone.

 

Van G

Link to comment

- how do I find out what the native resolution of the Denon is?

 

That's a bit harder without measurement hardware. As rough rule of thumb; for ~$1000 TrueHD/dtsMA capable AVRs it is likely 192 kHz. For cheapest ones and non-HD AVRs it is likely 96 kHz. But in any case it is likely to be either of the two and any conversion between these two is likely to have smaller sonic impact than conversion from 44.1 to 96.

 

- how to best process the audio on the PC side that is not the native resolution for the Denon, ie: Denon is 96/24 native and selected music file is 44.1/16?

 

Windows 7 can do the conversion by setting 96/24 or 192/24 as output rate. Some of the other player applications can also do better than Win7 on the conversion front. I don't now remember if QuickTime had it's own conversion settings on Windows. Mac OS X has also fairly good conversion since 10.5.

 

And finally it makes sense to listen and compare how it sounds compared to conversion made by the AVR. In the end it depends on particular AVR's capabilities and one's sound preferences.

 

 

Signalyst - Developer of HQPlayer

Pulse & Fidelity - Software Defined Amplifiers

Link to comment

The 3311ci has Audyssey MutliEqu XT support built into it.

 

This will allow you to digitally adjust the speaker levels, distance, and equalization for the speakers, which is something that makes surround sound work much better. This adjustment usually applies, no matter what the listening mode is (stereo, DTS, True HD, etc.)

 

Plain stereo playback can benefit from this kind of equalization, but - here's the kicker - it is analogous to putting a delay circuit, a volume control, and a graphic equalizer in line for each speaker.

 

Long and short, in my opinion, at this level, it isn't worth the extra $$. Get some experience with the 3311ci and if you find it makes that much difference, re-evalaute your choices. I suspect you will wind up defeating some or all of the Audyssey equalization for straight stereo playback though. :)

 

YMMV - Paul

 

 

 

 

 

 

 

 

Anyone who considers protocol unimportant has never dealt with a cat DAC.

Robert A. Heinlein

Link to comment

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now



×
×
  • Create New...