HQPlayer - just installed, and I'm impressed

[wilburthegoose]

I decided to audition HQPlayer, and I must admit that I'm impressed right now.  It's upsampling a CD to 192/24.

It upsamples music (I don't yet know how), and the resulting audio is almost stunning.   I'm controlling it with Roon, but that's not required.   I'm feeding the upsampled data to my XMC-1.   A well-engineered CD sounds almost like DVD-Audio.

Am I getting tricked by the placebo effect, or does this software really work?

[qdtjni]

I think Placebo but I may be wrong. ;-)

XMC-1 with or without Dirac or any other DSP functions?

[KeithL]

HQPlayer upsamples like anything else - it uses interpolation.
That's the technical terminology for "making up new data to fill in the spaces between the data you have".
(It's basically the same process as when a TV "upsamples to 4k from HD"....  )
However, HQPlayer offers you a whole bunch of different options about how to do it... and each may sound a tiny bit different.

Since you cannot CREATE new data, the result is NOT any more accurate than what you started with.
Therefore, what you have is STILL not the same as if you had a real 24/192k recording.
(Just as "an upsampled 4k picture" is not the same as a real 4k picture.)

In fact, you MAY be simply imagining the difference, but it's also possible that you are in fact actually hearing a difference.

This could happen for two reasons:

1)
A specific DAC might work better at one sample rate than at another (independent from the "quality" of the signal itself).
So, with exactly equal signals, a specific DAC might just produce a cleaner output with one or the other sample rate.
However, you should be aware of two things on that score:
INTERNALLY, there are huge benefits to upsampling/oversampling - which is why virtually all modern DAC chips already do oversample internally (so they're just doing it a second time).
EXTERNALLY, the circuitry in some DACs may work better at 192k, but the circuitry in some other DACs may work worse (remember that the DAC is already oversampling internally anyway).

2)
ALL conversions involve creating new data by some form or interpolation, then filtering it.
This inevitable results in some CHANGE from the original; this change may be audible, and you may like the way it sounds.
Therefore, they're giving you a bunch of options, each of which slightly alters the sound.... which isn't saying that any of them actually make it more accurate (or "better").

Finally, with HQPlayer, there is an option to use a special DAC which doesn't oversample internally, and have HQPlayer do it instead.
This lets you choose from a lot of alternative methods rather than simply use the one built into your DAC.
However, since people who design DACs usually know what they're doing, the one they designed into the DAC probably works pretty well.

In short, think of it as "one more way to adjust the sound, like a tone control" .

Dec 27, 2016 11:12:53 GMT -5 wilburthegoose said:

I decided to audition HQPlayer, and I must admit that I'm impressed right now.  It's upsampling a CD to 192/24.

It upsamples music (I don't yet know how), and the resulting audio is almost stunning.   I'm controlling it with Roon, but that's not required.   I'm feeding the upsampled data to my XMC-1.   A well-engineered CD sounds almost like DVD-Audio.

Am I getting tricked by the placebo effect, or does this software really work?

[qdtjni]

While we're kind of on the topic, is the XMC-1 when used with Dirac resampling everything to 48 kHz or just signals with higher then 48 kHz?

[wilburthegoose]

I'm sending the audio from HQPlayer via HDMI to my XMC-1. The XMC-1 is my DAC, and it's showing as a 192/24 input on the XMC-1 control panel.

[qdtjni]

Are you using Reference Stereo mode or something else?

[vneal]

It cannot sound better than what you put in.

[copperpipe]

Dec 27, 2016 12:23:47 GMT -5 vneal said:

It cannot sound better than what you put in.

Sure it can; for example, filters to reduce artifacts or distortion. Audacity has a nice feature useful for removing pop's and clicks from records. Just as another example. And the TV example was also used, how sometimes upscaling the resolution can look better on a higher resolution screen than the original video on a smaller resolution (but same physical size) display.

[wilburthegoose]

Dec 27, 2016 12:23:47 GMT -5 vneal said:

It cannot sound better than what you put in.

Sure can - we're talking about math and algorithms.  It's all just data these days.

[fbczar]

Dec 27, 2016 11:59:25 GMT -5 KeithL said:

HQPlayer upsamples like anything else - it uses interpolation.
That's the technical terminology for "making up new data to fill in the spaces between the data you have".
(It's basically the same process as when a TV "upsamples to 4k from HD"....  )
However, HQPlayer offers you a whole bunch of different options about how to do it... and each may sound a tiny bit different.

Since you cannot CREATE new data, the result is NOT any more accurate than what you started with.
Therefore, what you have is STILL not the same as if you had a real 24/192k recording.
(Just as "an upsampled 4k picture" is not the same as a real 4k picture.)

In fact, you MAY be simply imagining the difference, but it's also possible that you are in fact actually hearing a difference.

This could happen for two reasons:

1)
A specific DAC might work better at one sample rate than at another (independent from the "quality" of the signal itself).
So, with exactly equal signals, a specific DAC might just produce a cleaner output with one or the other sample rate.
However, you should be aware of two things on that score:
INTERNALLY, there are huge benefits to upsampling/oversampling - which is why virtually all modern DAC chips already do oversample internally (so they're just doing it a second time).
EXTERNALLY, the circuitry in some DACs may work better at 192k, but the circuitry in some other DACs may work worse (remember that the DAC is already oversampling internally anyway).

2)
ALL conversions involve creating new data by some form or interpolation, then filtering it.
This inevitable results in some CHANGE from the original; this change may be audible, and you may like the way it sounds.
Therefore, they're giving you a bunch of options, each of which slightly alters the sound.... which isn't saying that any of them actually make it more accurate (or "better").

Finally, with HQPlayer, there is an option to use a special DAC which doesn't oversample internally, and have HQPlayer do it instead.
This lets you choose from a lot of alternative methods rather than simply use the one built into your DAC.
However, since people who design DACs usually know what they're doing, the one they designed into the DAC probably works pretty well.

In short, think of it as "one more way to adjust the sound, like a tone control" .

Dec 27, 2016 11:12:53 GMT -5 wilburthegoose said:

I decided to audition HQPlayer, and I must admit that I'm impressed right now.  It's upsampling a CD to 192/24.

It upsamples music (I don't yet know how), and the resulting audio is almost stunning.   I'm controlling it with Roon, but that's not required.   I'm feeding the upsampled data to my XMC-1.   A well-engineered CD sounds almost like DVD-Audio.

Am I getting tricked by the placebo effect, or does this software really work?

Keith, Is it not true that the advantage of up-sampling in an app like HQPlayer or Audirvana, at least relative to a delta sigma DAC, is the fact that the DAC does not have to up-sample itself if it is already done by a computer? Basically the DAC has less work to do. Aside from up-sampling both programs offer filters that allow the user to alter the sound to taste relative to a given system. In my experience, with the XMC-1, both programs improve the sound of my system with up-sampling engaged.

[yves]

Remember the fact almost all (with only few exceptions) modern DACs use internal upsampling regardless of whether the data has already been upsampled before it gets passed on to the DAC. The DAC that I have can support 384 kHz sampling frequency, and, in my not so humble opinion, Alexey Lukin's resampling algorithm (iZotope RX 5 Audio Editor) sounds much better to me than HQPlayer.

[fbczar]

Dec 27, 2016 14:23:17 GMT -5 yves said:

Remember the fact almost all (with only few exceptions) modern DACs use internal upsampling regardless of whether the data has already been upsampled before it gets passed on to the DAC. The DAC that I have can support 384 kHz sampling frequency, and, in my not so humble opinion, Alexey Lukin's resampling algorithm (iZotope RX 5 Audio Editor) sounds much better to me than HQPlayer.

I do not think that is correct. A DAC does not repeat the up-sampling process below the sample rate fed to it. If you have a DAC that up-samples to 384KHz and it is fed a 192KHz up-sampled signal from HQPlayer the DAC will  only up-sample one time from 192 to 384. If the up-sampled signal was already at 384KHz the DAC would not up-sample at all. 

I use  Audirvana which contains iZotope on my iMac. The sound is fantastic, but it does take considerable work to set the filters and such for the best sound. Of course you can use the default settings if you wish. HQPlayer is an amazing program with a great set of very complex filters you can choose from. The two programs definitely sound different, but both are exceptional.

If you use Audirvana do you use custom filters and up-sampling?

[klinemj]

Dec 27, 2016 13:35:19 GMT -5 wilburthegoose said:

Dec 27, 2016 12:23:47 GMT -5 vneal said:

It cannot sound better than what you put in.

Sure can - we're talking about math and algorithms.  It's all just data these days.

Agree...my Geek DAC makes this very clear. Each of its 4 filters sound different from each other.

Mark

[KeithL]

To answer both your questions......

Yes, in theory, by upsampling ahead of time, you would reduce the amount of processing required to be done by the DAC. However, while this may intuitively seem to be an advantage, it usually is not. While it may intuitively seem as if the DAC should "sound better because it isn't working as hard", that simply isn't usually the case. The computers we are all familiar with are general purpose devices, which means that it's easy to ask more than they can deliver, which can have negative consequences. Ask your computer to decode a video that requires more processing power than it has available and you'll see dropped frames and pixellation. However, because DAC chips are dedicated devices, they are never called upon to deliver beyond their capabilities. And so, since they're always operating within their capabilities, reducing the processing needs further really doesn't make any difference. While it's possible for this to happen in a custom-programmed DAC, if it was poorly designed, it simply doesn't happen in commercial DAC chips.

This same logic is applied - equally incorrectly - to various colored magic markers, cleaning products, and sticky things - claimed to "reduce the error rate on CDs". The reality is that, after the error correction mechanisms inherent in CD players do their job, the typical final bit-error-rate for virtually all CDs is zero - and you can't reduce zero to a lower number no matter what you do. (Of the 500 or so CDs I've ripped, precisely TWO failed to be 100.0% perfect; one of those had a bad scratch, which resulted in a few data errors, all within a few seconds, on a single track; the other had a manufacturing flaw that was present in multiple copies - and so was totally beyond proper correction.)

There's also the possibility that the oversampling algorithm employed by the computer might do a BETTER job than the one in the DAC chip. While it CANNOT improve the original data, or produce new data that is deterministically closer to the original, it might conceivable do less damage to the data than the algorithm used by the DAC chip itself. (If the chip itself is running at the same rate, doubling the sample rate externally, then doubling it again inside the chip, MIGHT produce a better result than simply quadrupling it inside the chip. I've never seen any proof that this actually has happened in real life, but it's theoretically possible. Note that this would depend on the specifics of the DAC chip used, and of how it was configured.... it would almost certainly be different for different chips and different configurations. This is the same argument for using a DAC chip that doesn't oversample at all, and doing the oversampling on your computer.)

Your second question is more reasonable. Because every upsampling process involves calculating new samples, and applying filters, every one has at least the potential to change the sound... and many of them do. Therefore, each of those options may well sound different, and there's the distinct possibility that you'll simply prefer the way one or the other sounds. HOWEVER, someone else suggested that the upsampling process might act to correct or remove some sort of distortion, and so actually improve the original. The problem there is that, with no reason to expect distortion to be present to begin with, there's no way to predictively correct any distortion that might be present. While some processors, including the MQA encoder, CLAIM to actively detect and correct some types of distortion, that claim is yet to be demonstrated effectively, and is FAR beyond the capabilities of an upsampling algorithm. It seems far more likely that each setting is causing more or less random changes, and that some people happen to like the way some of those changes sound with some program material.

PLEASE NOTE that this isn't a problem - any more than turning up the Treble by 3 dB - if that setting happens to sound better on your system.
(Just don't read more into the entire process than belongs there.)

Dec 27, 2016 13:36:29 GMT -5 fbczar said:

Dec 27, 2016 11:59:25 GMT -5 KeithL said:

HQPlayer upsamples like anything else - it uses interpolation.
That's the technical terminology for "making up new data to fill in the spaces between the data you have".
(It's basically the same process as when a TV "upsamples to 4k from HD"....  )
However, HQPlayer offers you a whole bunch of different options about how to do it... and each may sound a tiny bit different.

Since you cannot CREATE new data, the result is NOT any more accurate than what you started with.
Therefore, what you have is STILL not the same as if you had a real 24/192k recording.
(Just as "an upsampled 4k picture" is not the same as a real 4k picture.)

In fact, you MAY be simply imagining the difference, but it's also possible that you are in fact actually hearing a difference.

This could happen for two reasons:

1)
A specific DAC might work better at one sample rate than at another (independent from the "quality" of the signal itself).
So, with exactly equal signals, a specific DAC might just produce a cleaner output with one or the other sample rate.
However, you should be aware of two things on that score:
INTERNALLY, there are huge benefits to upsampling/oversampling - which is why virtually all modern DAC chips already do oversample internally (so they're just doing it a second time).
EXTERNALLY, the circuitry in some DACs may work better at 192k, but the circuitry in some other DACs may work worse (remember that the DAC is already oversampling internally anyway).

2)
ALL conversions involve creating new data by some form or interpolation, then filtering it.
This inevitable results in some CHANGE from the original; this change may be audible, and you may like the way it sounds.
Therefore, they're giving you a bunch of options, each of which slightly alters the sound.... which isn't saying that any of them actually make it more accurate (or "better").

Finally, with HQPlayer, there is an option to use a special DAC which doesn't oversample internally, and have HQPlayer do it instead.
This lets you choose from a lot of alternative methods rather than simply use the one built into your DAC.
However, since people who design DACs usually know what they're doing, the one they designed into the DAC probably works pretty well.

In short, think of it as "one more way to adjust the sound, like a tone control" .

Keith, Is it not true that the advantage of up-sampling in an app like HQPlayer or Audirvana, at least relative to a delta sigma DAC, is the fact that the DAC does not have to up-sample itself if it is already done by a computer? Basically the DAC has less work to do. Aside from up-sampling both programs offer filters that allow the user to alter the sound to taste relative to a given system. In my experience, with the XMC-1, both programs improve the sound of my system with up-sampling engaged.

[KeithL]

As is often the case, the reality is a bit more complicated......

You are correct that any given DAC has a limit on the top frequency limit it can use internally.
However, that may not be the same as the highest sample rate it will accept as an input.
For example, the AD1955 DAC we use in a lot of our equipment has an oversampling limit (internally) of 384k.
But the maximum input it can accept is 192k.
It can be programmed in terms of what input data rates are oversampled, and by what multiples.
(So you CANNOT send in 384k to avoid oversampling, but you could configure it not to oversample a 192k input.)
This is all "conceptual" because the delta-sigma modulator itself operates at narrower bit depths and higher frequencies (internally).

Some DAC chips also use more complicated schemes... for example, Sabre DACs directly decimate and resample to the modulator stream frequency.
(So, whatever you feed into a Sabre DAC might end up as a 1-bit data stream at somewhere around 40 Mhz.)

In the end the math works out the same.

However, what I said remains true, virtually all modern DACs oversample on their own.
And, by doing your own oversampling externally, you're simply separating the process into multiple steps, giving yourself the opportunity to fiddle with the sound in more places.
If your goal is optimum ACCURACY, then there's little proof that the filters used by HQPlayer are more accurate than the ones used internally by most DAC chips.
And, if you're looking for VARIETY, then it gives you lots of options to play with.

And, as I said, another alternative is to use a true NON-oversampling DAC and let HQPlayer do all the oversampling for real....
However, some non-oversampling DACs are very expensive; and some of them are downright awful; and some are both expensive and awful.
(And there are a few cheap ones that aren't all that bad..... )
The catch is that the ones I've heard, including a few somewhat expensive ones, actually don't end up sounding any better than a "regular" DAC.....

Dec 27, 2016 14:46:01 GMT -5 fbczar said:

Dec 27, 2016 14:23:17 GMT -5 yves said:

Remember the fact almost all (with only few exceptions) modern DACs use internal upsampling regardless of whether the data has already been upsampled before it gets passed on to the DAC. The DAC that I have can support 384 kHz sampling frequency, and, in my not so humble opinion, Alexey Lukin's resampling algorithm (iZotope RX 5 Audio Editor) sounds much better to me than HQPlayer.

I do not think that is correct. A DAC does not repeat the up-sampling process below the sample rate fed to it. If you have a DAC that up-samples to 384KHz and it is fed a 192KHz up-sampled signal from HQPlayer the DAC will  only up-sample one time from 192 to 384. If the up-sampled signal was already at 384KHz the DAC would not up-sample at all. 

I use  Audirvana which contains iZotope on my iMac. The sound is fantastic, but it does take considerable work to set the filters and such for the best sound. Of course you can use the default settings if you wish. HQPlayer is an amazing program with a great set of very complex filters you can choose from. The two programs definitely sound different, but both are exceptional.

If you use Audirvana do you use custom filters and up-sampling?

[KeithL]

NOTE: FOR THOSE OF YOU WHO WANT TO PLAY WITH SAMPLE RATES

HQPlayer is cool, but it doesn't let you save files (it only converts "on the fly").

The converter in Izotope RX is excellent (it isn't free, but they do have a 30 day free trial).

Or you could try Voxengo R8Brain (there are both a free version and a pro version).
You can get both here:    www.voxengo.com/product/r8brain/

AND, FOR THOSE OF YOU WHO LIKE THIS SORT OF STUFF.......
It's a bit out of date, but here's a performance comparison of a whole lot of sample rate converters...
src.infinitewave.ca/

[wilburthegoose]

@keithl - Thanks - I really appreciate your posts!

Q: How does the DAC in the XMC-1 compare to some of the pricier options out there? I happen to think it works great for me, but there are obviously others who think the $ spent is worth it for a standalone DAC.

[fbczar]

Dec 27, 2016 16:02:48 GMT -5 KeithL said:

NOTE: FOR THOSE OF YOU WHO WANT TO PLAY WITH SAMPLE RATES

HQPlayer is cool, but it doesn't let you save files (it only converts "on the fly").

The converter in Izotope RX is excellent (it isn't free, but they do have a 30 day free trial).

Or you could try Voxengo R8Brain (there are both a free version and a pro version).
You can get both here:    www.voxengo.com/product/r8brain/

AND, FOR THOSE OF YOU WHO LIKE THIS SORT OF STUFF.......
It's a bit out of date, but here's a performance comparison of a whole lot of sample rate converters...
src.infinitewave.ca/

Thanks Keith for taking the time to answer. I have read that great DACs benefit less or not at all from HQPLayer and the like while DACs some would consider average may benefit more. It does seem almost universally agreed upon that sample rate converters that do not convert on the fly give better results. The AKM DAC in the upcoming RMC-1, may be one of those that would not benefit from up-sampling. Would you agree?

[KeithL]

The DACs in the XMC-1 are actually quite good.
I haven't actually done an A/B comparison between them and the DC-1, but I would say they are definitely in the same class.
(I've owned three DACs in the $1k range, and I'm not convinced they were better than the XMC-1 or the DC-1.)

I've got to say that, once you get past the level of the XMC-1 or the DC-1, while I hear tiny differences between different DACs, I'm not totally convinced that any of them are really significantly better.
(And, to put it bluntly, I'm going to have to be VERY convinced that the DAC itself really is better, and that I can hear the differences on the recordings I listen to, before I consider spending $1k or more on a DAC.)

Dec 27, 2016 16:15:23 GMT -5 wilburthegoose said:

@keithl - Thanks - I really appreciate your posts!

Q: How does the DAC in the XMC-1 compare to some of the pricier options out there? I happen to think it works great for me, but there are obviously others who think the $ spent is worth it for a standalone DAC.

[novisnick]

Dec 27, 2016 16:15:23 GMT -5 wilburthegoose said:

@keithl - Thanks - I really appreciate your posts!

Q: How does the DAC in the XMC-1 compare to some of the pricier options out there? I happen to think it works great for me, but there are obviously others who think the $ spent is worth it for a standalone DAC.

Great question!