Using ASIO with Amazon HD ... a partial solution?

[JanesJr1]

I have not been able to make it work with my D10s

View attachment 190045
, did you do anything special? I'm using the V5.30 driver and W10, but it never autoswitched with previous driver versions either.

You can get it to change the reported resolution with different source material. But as I figured out in later forum conversations, it's still being sourced at the bit rate you select in Windows, and then re-sampling. Just set things in Windows at 24/96 or something like that and don't worry about it. I personally don't think there is any practical sound quality benefit from trying to match higher resolution sources without resampling, even if you could on Amazon. Follow the forum thread through and you should find postings on this. The Amazon telemetry is misleading as to what it really is.

 

[DavidBarajas]

You can get it to change the reported resolution with different source material. But as I figured out in later forum conversations, it's still being sourced at the bit rate you select in Windows, and then re-sampling. Just set things in Windows at 24/96 or something like that and don't worry about it. I personally don't think there is any practical sound quality benefit from trying to match higher resolution sources without resampling, even if you could on Amazon. Follow the forum thread through and you should find postings on this. The Amazon telemetry is misleading as to what it really is.

Hi! I know that, I was replying to Mikett who said it autoswitched on his Topping D90SE (same driver as my D10s, unless he was using the Microsoft driver), which is how exclusive mode is supposed to behave, is it not? I would probably agree with you about the practicality of it, but why Amazon fails at implementing exclusive mode the proper way is still of interest to me.

 

[JanesJr1]

Hi! I know that, I was replying to Mikett who said it autoswitched on his Topping D90SE (same driver as my D10s, unless he was using the Microsoft driver), which is how exclusive mode is supposed to behave, is it not? I would probably agree with you about the practicality of it, but why Amazon fails at implementing exclusive mode the proper way is still of interest to me.

Yes, I don't know why they don't do it the proper way, and I wonder more that they finessed their telemetry to imply something different. But on the plus side, at least they are doing better than Spotify at lossless streaming, and at a good price, esp. for Prime members.

I find their catalog pretty darn good, but their search and personal library management interfaces need a lot of work.

 

[gerG]

I just spent the morning on this annoying bug. I believe that all exclusive mode does is set the volume to 0db. The Amazon app does use the shared mode setting as the device capability. However, despite what the app claims it is streaming, the output to the DAC is always taken from the shared mode setting. That hints to me that the resampling is done in Windows. I suspect that neither asio nor wasapi are at play here. Caveate: I have not installed any drivers, so running on the windows built in stuff.
I got the same result with Wondows 10 pro and Windows 11 home on different but relatively new laptops. DACs were the Gustard X18 and the SMSL M500 mkii.
I spent a good bit of time on the phone with Amazon support. I would honestly like to see this fixed. I am on the second "escallation". They are supposed to call me back. I have no doubt that at some level they know exactly what is going on here. We shall see if they call me back.

 

[JanesJr1]

I just spent the morning on this annoying bug. I believe that all exclusive mode does is set the volume to 0db. The Amazon app does use the shared mode setting as the device capability. However, despite what the app claims it is streaming, the output to the DAC is always taken from the shared mode setting. That hints to me that the resampling is done in Windows. I suspect that neither asio nor wasapi are at play here. Caveate: I have not installed any drivers, so running on the windows built in stuff.
I got the same result with Wondows 10 pro and Windows 11 home on different but relatively new laptops. DACs were the Gustard X18 and the SMSL M500 mkii.
I spent a good bit of time on the phone with Amazon support. I would honestly like to see this fixed. I am on the second "escallation". They are supposed to call me back. I have no doubt that at some level they know exactly what is going on here. We shall see if they call me back.

You are correct. The Amazon telemetry is at best misleading. I didn't have separate telemetry on sampling rate from my basic DAC but later found that the Amazon presentation involves enough semantic sleight-of-hand to be outright deceptive. Exclusive mode is possible but there is no autoswitching except by means of re-sampling the source to the shared mode resolution.

The thing that surprised me the most is that Amazon would have both the guile to finesse the facts so much, and a sufficient corporate reason to do so. However, I personally doubt that any of it makes much difference: I can find no practical sound quality benefit from seeking un-resampled, ultra-high resolution. With Amazon, the price is right and the catalog is large; the real things to fret about are the search and library interfaces.

At the very least, Amazon is doing 100% better than Spotify at lossless streaming. Tidal has the MQA tar-baby. Qobuz has a much better interface than Amazon but seemingly a much more limited classical catalog (a personal preference for me).

 

[gerG]

You are correct. The Amazon telemetry is at best misleading. I didn't have separate telemetry on sampling rate from my basic DAC but later found that the Amazon presentation involves enough semantic sleight-of-hand to be outright deceptive. Exclusive mode is possible but there is no autoswitching except by means of re-sampling the source to the shared mode resolution.

The thing that surprised me the most is that Amazon would have both the guile to finesse the facts so much, and a sufficient corporate reason to do so. However, I personally doubt that any of it makes much difference: I can find no practical sound quality benefit from seeking un-resampled, ultra-high resolution. With Amazon, the price is right and the catalog is large; the real things to fret about are the search and library interfaces.

At the very least, Amazon is doing 100% better than Spotify at lossless streaming. Tidal has the MQA tar-baby. Qobuz has a much better interface than Amazon but seemingly a much more limited classical catalog (a personal preference for me).

It was very devious. Amazon never claimed that "Exclusive Mode" did anything more than give their client control of the external audio device. I never found any reference to asio, wasapi, or anything similar. Everything else was conjecture and assumption based on the status quo.
What really wrinkles my undies is that they certainly have the resources to resolve this issue almost overnight (second day at most). Why the hell are they letting this linger? I mean, windows is a fairly common operating system. Within that subgroup there are probably more than a couple of dozen users streaming to an external DAC. What do they have to gain from burrying this issue?

 

[JanesJr1]

It was very devious. Amazon never claimed that "Exclusive Mode" did anything more than give their client control of the external audio device. I never found any reference to asio, wasapi, or anything similar. Everything else was conjecture and assumption based on the status quo.
What really wrinkles my undies is that they certainly have the resources to resolve this issue almost overnight (second day at most). Why the hell are they letting this linger? I mean, windows is a fairly common operating system. Within that subgroup there are probably more than a couple of dozen users streaming to an external DAC. What do they have to gain from burrying this issue?

I noticed recently that Qobuz reduced their subscription to $10.63/month if paid in one annual payment. I was using savings on Amazon Prime to subsidize a second subscription to Idagio, which specializes in a deep classical catalog with good search features. But I might pay the extra $2.63/month and switch to Qobuz for my general-interest listening, now that their price is down.

 

[JanesJr1]

You are correct. The Amazon telemetry is at best misleading. I didn't have separate telemetry on sampling rate from my basic DAC but later found that the Amazon presentation involves enough semantic sleight-of-hand to be outright deceptive. Exclusive mode is possible but there is no autoswitching except by means of re-sampling the source to the shared mode resolution.

The thing that surprised me the most is that Amazon would have both the guile to finesse the facts so much, and a sufficient corporate reason to do so. However, I personally doubt that any of it makes much difference: I can find no practical sound quality benefit from seeking un-resampled, ultra-high resolution. With Amazon, the price is right and the catalog is large; the real things to fret about are the search and library interfaces.

At the very least, Amazon is doing 100% better than Spotify at lossless streaming. Tidal has the MQA tar-baby. Qobuz has a much better interface than Amazon but seemingly a much more limited classical catalog (a personal preference for me).

PS on exclusive mode: my XMOS driver confirms exclusive mode if I toggle the right set-up items in Windows and Amazon; and the 0db volume I'm pretty sure confirms that the driver is handling volume. But as we discussed, this doesn't also translate into correct autoswitching, etc.

 

[gerG]

For the sake of thoroughness, I installed the latest Gustard driver. No change overall. However Gustard software tells me that the stream is arriving at 24/96 (the limit that I specified) while the Amazon client is claiming that the stream is departing at 16/44.1 sharp. Apparently the stream takes a detour through some interstitial dark zone in the nether reaches of my PC, and gains an awful lot of weight along the way!

 

[JanesJr1]

For the sake of thoroughness, I installed the latest Gustard driver. No change overall. However Gustard software tells me that the stream is arriving at 24/96 (the limit that I specified) while the Amazon client is claiming that the stream is departing at 16/44.1 sharp. Apparently the stream takes a detour through some interstitial dark zone in the nether reaches of my PC, and gains an awful lot of weight along the way!

Yup.

 

[xenoson]

...and then re-sampling. Just set things in Windows at 24/96 or something like that and don't worry about it. I personally don't think there is any practical sound quality benefit from trying to match higher resolution sources without resampling...

One major drawback is cpu usage. The amazon music application is based on chromium browser and that makes it a resource hog anyway. Poking around with Process Explorer you can find a lot of load on libavfilter threads, I suspect they are responsible for upsampling. Playing 44.1 kHz 16 bit material on 192 kHz 24 bit triples cpu load for me and causes glitches. It doesn't really help in that regard that chrome always opens the sound device in full multichannel mode meaning that all channels except LR play empty. Comparing that to neglectable load with foobar 2000 playing such material, using a full blown browser as an audio player is a giant mess. Based on avfilter action and studying the logs I would suspect that playing all music at 44.1 kHz 16 bit works without resampling because ultra hd material is also available as "AudioQuality: HD44" on the server.
Higher than that is nice to have but honestly beyond human perception. Gaps and glitches on the other hand are highly annoying and amazon shouldn't interpret "exclusive mode" in a way that nothing else is to be done on the machine because a music player grabs two cpu cores.

 

[JanesJr1]

One major drawback is cpu usage. The amazon music application is based on chromium browser and that makes it a resource hog anyway. Poking around with Process Explorer you can find a lot of load on libavfilter threads, I suspect they are responsible for upsampling. Playing 44.1 kHz 16 bit material on 192 kHz 24 bit triples cpu load for me and causes glitches. It doesn't really help in that regard that chrome always opens the sound device in full multichannel mode meaning that all channels except LR play empty. Comparing that to neglectable load with foobar 2000 playing such material, using a full blown browser as an audio player is a giant mess. Based on avfilter action and studying the logs I would suspect that playing all music at 44.1 kHz 16 bit works without resampling because ultra hd material is also available as "AudioQuality: HD44" on the server.
Higher than that is nice to have but honestly beyond human perception. Gaps and glitches on the other hand are highly annoying and amazon shouldn't interpret "exclusive mode" in a way that nothing else is to be done on the machine because a music player grabs two cpu cores.

I have a response that mostly agrees with your assessment, although I am not technically savvy enough to work out the details. On the one hand, given that Amazon UltraHD selections are mostly 24/96, I set Windows at 24/96 and at least those selections don't impose resampling. On the other hand, 24/96 imposes a resampling load for CD quality material. Further, I've noticed that when I set Windows at a resolution other than 24/96, my very inefficient DCA headphones are more likely to show signs of overload, which I tentatively attribute to poor network bandwidth at busy times of the day and/or resampling burden. That last part is a little mysterious to me because I have trouble parsing out how it would affect the headphones. It seems to me that a network bandwidth bottleneck might induce bass attenuation or pumping that would resemble a power shortfall to current-hungry headphones. I'm less sure that a resampling-related cpu bottleneck could also do that, but maybe it could. Anyway, the 24/96 being an easier burden during busy work hours on my business ISP than, say, 24/48, is an intriguing footprint that I didn't expect. I mean, would a cpu bottleneck influence headphone sound as bass attenuation? ... as opposed, say, to causing slow PC response to user commands, GUI instability, slow web-surfing or degraded video streaming?

 

[xenoson]

I mean, would a cpu bottleneck influence headphone sound as bass attenuation?

I would rule out bass attenuation. Most often some samples get skipped in the playback stream causing an audible click, pop, chirp or short silence, depending on what exactly happened.

poor network bandwidth at busy times of the day

I never observed the network as a bottleneck. The download is shown with the dark grey progress bar, it is always fast and ahead of playback, but this could probably vary with bad wifi.

The second big CPU churn is the Widevine DRM decryption. It seems to work on chunks and gets triggered every couple of seconds. There used to be a bug with that about a year ago that caused a playback glitch every time it triggered even with CD quality playback. I reported that several times to the service and it got attention and works better now.

With 24/192 music there is a different file on the server for UHD192, UHD96, UHD48 and HD44 as they call it in the logs. Meaning there is no on-the-fly downsampling during playback and hopefully the downsampling of those files was done correctly.

They really have it inside out with upsampling because in exclusive it would be possible to set the correct mode for every song. Resampling is only necessary for shared mode WASAPI. But I have the impression that the additional higher CPU load in exclusive mode could also be some other bug because upsampling is active in shared mode and the load is lower there.
Anyway, I try to avoid settings that involve resampling and it would be so much more convenient with a correct implementation of exclusive WASAPI.

 

[12Many]

AHA!

I think I have figured out how to enable the wasapi exclusive/bypass feature of Amazon music HD. It's not obvious! This feature enables you to play Amazon Music HD selections at "HD" or "Ultra HD" resolution (for tracks so-labelled) and auto-switches the Amazon playback bit depth/sampling rate to the native resolution of the source recording, without upsampling or downsampling. "Exclusive Mode" does this by bypassing the regular Windows 10 audio mixer which otherwise will constrain playback to lower resolutions, or in some cases, force upsampling to higher resolutions, depending on Windows settings.

1. First, in Windows 10 you have to go to Settings/Sound and select your DAC or other output device. Then for that device, choose Device Properties/Additional Device Properties and select the Advanced tab at the top. In the Advanced tab, check the two boxes enabling exclusive mode. ALSO: on the top of the same tab menu, choose your device's maximum bit depth/sampling rate (e.g. 24/192) for the "shared-mode" Windows default sampling rate. Even though you will be in exclusive mode, not shared mode, Amazon interprets that value as the maximum for the device in Amazon's auto-switching of the sampling rate in exclusive mode. (Perhaps not its intended use! But perhaps a practical ceiling rate for Amazon's auto-switching if it is not set too low.) So set it to 24/192 or to something that is your device's true maximum. Amazon will not oversample a lower-resolution recording up to that ceiling; it will just auto-switch to the correct rate. But if you set the default/shared-mode rate too low (e.g. 16/44.1), Amazon will downsample a higher-resolution recording to that rate in playback.

2. In Amazon Music HD, open the settings menu (little head in a circle in upper right corner), Under the "Playback" heading, toggle exclusive mode to On to allow exclusive playback.

3. Then when actually playing your first music selection of a listening session in Amazon Music HD (and apparently ONLY in that screen), click the little speaker image/icon at the lower right corner to display the "Available Devices" list. As a result of the prior step, there should be not only a list of available devices, but also an "Exclusive Mode" toggle below it. Make sure the correct device is checked and toggle the Exclusive Mode on. I believe you have to do that last step EACH time you open Amazon Music HD or change playback devices ... the setting is not "set and forget", and I imagine this is so that when you leave your listening session, Windows audio will work normally for other Windows apps, instead of being bypassed.

If you do all of the above, the "volume" slider at the bottom of the Available Devices menu grays out and is disabled, seemingly validating the wasapi exclusive-mode bypass of the Windows mixer. Or you can click the little "HD" or "Ultra-HD" indicator on the bottom left of the Amazon playback menu for that selection to see the playback rate as well as the recording's native sampling rate and the "device limit" set in step 1 above. (Thanks DIYAUDIONUT for that tip!)

The above set of steps is not explained clearly anywhere that I have seen ...

This is very valuable information. This seems to work on my system. Not 100% sure, but based on my limited ability to monitor, it is matching the PC output to the resolution of the song. And, it sound much better since I have clicked on the small speaker and moved the slider to exclusive. It sounds so much better that there must be more going on beyond just the resolution change from CD quality to high res. JanesJr1 is my new hero at this point.

 

[JanesJr1]

This is very valuable information. This seems to work on my system. Not 100% sure, but based on my limited ability to monitor, it is matching the PC output to the resolution of the song. And, it sound much better since I have clicked on the small speaker and moved the slider to exclusive. It sounds so much better that there must be more going on beyond just the resolution change from CD quality to high res. JanesJr1 is my new hero at this point.

I'm sorry to say this is an old post and I think subsequent events showed that while the Amazon telemetry about playback bitrate and so on appears correct, it is in fact misleading. I don't remember all the details now but I can remember being rather amazed that Amazon would care enough to finesse the presentation of data to appear to be one thing while doing another. When I originally worked this out, I did not have a DAC with a readout screen that could validate the info presented in the Amazon telemetry. I later did that in conjunction with another ASR member and it proved that Amazon consistently re-sampled up or down to either 16/44.1 or 24/96 (or something like that), depending on your windows sound advanced properties, and that matching source and DAC resolution, if it occurs, is accidental. I finally concluded that it probably doesn't make much difference to sound quality compared to a lot of other things and let go of the issue.

 

[12Many]

I'm sorry to say this is an old post and I think subsequent events showed that while the Amazon telemetry about playback bitrate and so on appears correct, it is in fact misleading. I don't remember all the details now but I can remember being rather amazed that Amazon would care enough to finesse the presentation of data to appear to be one thing while doing another. When I originally worked this out, I did not have a DAC with a readout screen that could validate the info presented in the Amazon telemetry. I later did that in conjunction with another ASR member and it proved that Amazon consistently re-sampled up or down to either 16/44.1 or 24/96 (or something like that), depending on your windows sound advanced properties, and that matching source and DAC resolution, if it occurs, is accidental. I finally concluded that it probably doesn't make much difference to sound quality compared to a lot of other things and let go of the issue.

I trust your test results and I have some conflicting data from the info I get from the software. That being said, when I set my computer to 25/96 and give the exclusive control to the requesting app in windows sound settings, and then toggle the exclusive mode slider in Amazon, there is a significant improvement in sound quality in exclusive mode which is noticeable by me and my wife. I don't know what is causing it because I feel it is more than just a jump from cd quality to high res (which is often a very small improvement). So, while I can not confirm or prove what is going on, the sound quality improvement is significant. The exclusive slider in the app does something. Since your test results show it is not giving true exclusive control, I wonder what it does.

 

[Propheticus]

If it's truly acting like an exclusive feed, it could be bypassing any number of 'enhancements' on your pc. Perhaps you have some Dolby Atmos or other virtual surround DSP that came with your pc? Some laptops even have a keyboard function key to invoke the config panel to enable/disable this.

 

[12Many]

If it's truly acting like an exclusive feed, it could be bypassing any number of 'enhancements' on your pc. Perhaps you have some Dolby Atmos or other virtual surround DSP that came with your pc? Some laptops even have a keyboard function key to invoke the config panel to enable/disable this.

That could be. I have all the processing and enhancement functions turned off in the Sound portion of the control panel that I can find, but I could be missing something. When I get more time I am going to try qobuz for a year and I believe it truly allows for exclusive control, although I have not tired it.

 

[JanesJr1]

I trust your test results and I have some conflicting data from the info I get from the software. That being said, when I set my computer to 25/96 and give the exclusive control to the requesting app in windows sound settings, and then toggle the exclusive mode slider in Amazon, there is a significant improvement in sound quality in exclusive mode which is noticeable by me and my wife. I don't know what is causing it because I feel it is more than just a jump from cd quality to high res (which is often a very small improvement). So, while I can not confirm or prove what is going on, the sound quality improvement is significant. The exclusive slider in the app does something. Since your test results show it is not giving true exclusive control, I wonder what it does.

You probably know this, but exclusive mode could provide such a benefit by avoiding windows-based sound processes that are noisy or distorted. I am not clear that exclusive mode is not truly exclusive mode in Amazon, in the sense of isolating the Amazon playback from windows-based processes. In other words, Amazon's handling of autoswitching may have nothing to do with exclusive mode. I suspect the autoswitching limitation is an Amazon platform limitation on the signal bitstream whether exclusive or not. If you have better technical insight, let me know.

In my case, exclusive mode adds nothing to sound quality on three of my PC's, but precludes my use of EQ software which has profound benefits to sound quality for me. I do not know why exclusive mode would improve sound quality for you except by quashing noisy or distorted processes on the Windows side. Did you suppress the "enable audio enhancements" option in the windows sound advanced properties? Are enhancements like Dolby Atmos or other such OEM processes automatically turned on during PC startup? Does the same benefit of improved sound quality appear if you use exclusive mode on a different machine with similar settings in Windows sound advanced properties?