DTS DCH Driver for Realtek HDA [DTS:X APO4 + DTS Interactive]

[Ferather]

Note that PAM can be analogue or digital, in the case of PAM X which goes direct to speaker, it can be considered as both at the same time (to human ears).
PAM X is still digital because in audio each colour is a digit value, 0 - 255 (256 total values) which are digits, and a range of digits.

However, the value does not need to specifically be a digit, it can be a direct command.
Either way RGB optical is NOT binary, binary is two values 0 or 1.

----

PCM (pulse code modulation) uses binary, which is why compression exists (to use less code), well you can not beat, nor compress 1 bit values.

In basic calculator form, 256 x 256 would be 3 bits linear, with a 1 bit result if the range was high enough (65,536).

 

[Ferather]

Modern optical transmitters and receivers can do 300+ Tbits/s (Tera), the universal RGB cable could transmit video at that rate, with a maximum value range of ~16.78 million.
Using copper and PAM (not PAM X), HDMI 2.2 is rated at 96 Gbits/s, no where near, in addition to get longer ranges, optical HDMI cables are needed.

In terms of an optical CPU with the same rate (or even audio for that matter), it would be 300+ THz, ~16.78 million bit.

----

A flexible enough optical cable can be wrapped around a Wi-Fi router, and then a magnet, and get 0% interference (plastic is an insulator).

----



Basic illustration for the input and output points for an audio core, PAM X based.


For an emitter, we [also] transcode colour value to power value.

 

[Ferather]

Reminder, its possible to group 2 bit pulses into 1 bit time, and get ~16.78 x ~16.78 in values, in this case 300 Tbits/s would still be 300, but 150 Tbits/s effective.

----

Its also possible for companies like DTS or Dolby to use PAM X and the available colour range differently based on their own format.
However, in general the number assigned to each colour would need to be standardised and [given a version].

Note that you can not compress 1 bit, you can however increase the rate of bits (bit rate).

 

[Ferather]

In terms of a decibel specification there isn't one, the bit value range specifies the max bit, for example 256 bit (which would be 128 up [+] and 128 down [-]).
A microphone could be using for example 6v over the 256 bit steps, and a speaker 12v over the 256 bit steps (256 volt steps).

In the analogue sense 6v would produce a lower power signal than 12v (hence amps).
All speakers have their own sensitivity (and room size) in addition.

----

Also note a speaker can still be 12v max with a larger bit range, and volt steps, for example 512, same max power output.

Note that an off pulse is still a value, and in with audio, 0v, also off.

----

The DMAS main unit can change bit values (for example +2,-2) to amplify the output, without using power, and logically as not to clip.
Given all programming is working as intended the only point clipping should ever apply is the mic, due to its limits.

 

[Ferather]

Here is a basic 8 bit illustration of a microphone to speaker, with isolated power, which can be any power specification as long as 8 volt steps are used to represent each bit.
One base signal that can be any given voltage at either end, without using power to transmit or power amplifier, can be processed (digital, SoC, PAM X).



Faster and more accurate than any current method. For audio, bit value (infinite) = bit depth, bit rate = sample rate.
In a single linear lane, 10 channels at 200 Mbits/s requires 2 Gbits/s, plus sample header bits.

 

[Ferather]

Reminder: If you like the DTS processing used in this pack, buy a Sound Unbound license, not only will it work with the pack (in full), but also fund DTS who made it.

I will test the Realtek update in the next week or so.

----

You could imagine, a DTS DMAS (DMAS with added DTS processing), with DTS certified DMAS speakers, would knock the current market out of its socks.




---

You can also imagine what would happen if AMD used PAM X to produce next gen optical CPU's and GPU's (non binary).

 

[Ferather]

DTS DCH 6.0.9800.1 (2025), is available to download (see post 1). Several updates and changes, if DTS does not work for you, use the previous pack (alternative files no longer work for me).

----

I have not updated the provided apps, but I did test they still work. If you want a newer version see the .txt file. The current Realtek app has a typo, front instead of rear.
Also make sure to restart the computer if you decide to unlock all app features available, see the .txt file provided in the driver pack for instructions.

----

Updated the DTS service to a newer version that supports 'External Speakers' as the default, not 'Internal Speakers' (latest versions sets default 'Internal').
Improved the license unlock for the DTS controller, no license key is required, however you do still need a preset (provided).
Added a fix for cases where the Realtek secondary task is not added to the device startup.

Impendence detection is working for the first time for me. Note that the preset is tuned for 'True Sound', not adding unoriginal flavour.

----

Note that SPDIF and HDMI are not officially supported form factors, however I have modified the check system to work fully on multichannel, and default mode only when in stereo.
Under normal conditions you will not have control of the app when in stereo mode with an unsupported form factor, but as mentioned you get the default mode.

Some Realtek devices have a secondary form factor on SPDIF, which is 'External Speakers', the app will be available to use in stereo mode.
Also note that SPDIF has been programmed with a multichannel EFX (DTS, 12+ ch), and not the stereo EFX version (DTS).

----

 

[Ferather]

Note that it may a take a little while before Realtek-DTS configuration settles down (most likely Windows), after installing the drivers from fresh.
For me analogue response feels instant, although I have 'Stereo Mix' enabled, muted, and set listen to: default device.

Highest execution time for the HDA bus is [0.002607 ms], using SPDIF Optical out.



CPU includes EAPO upmixing stereo to 6 ch (own preset).

 

[Ferather]

The motherboard HDA is an AMD Starship/Matisse HD Audio Controller, the one in my AMD GPU is an AMD Navi 31 HDMI/DP Controller.
Both are recommendable, and yes you can run this pack on an AMD GPU, as its HDA the same as your Realtek.

Three names I can suggest: Asus, AMD, DTS, with a minimum of an ALC1200 (ideally 1220+).

----

> If you enjoy the DTS processing, buy a Sound Unbound license if not already, you are entitled to all future updates, and fund DTS.
> If you own a Sound Unbound license you can download and update the DTS apps (both) from the Windows store.

----

Also note that the above CPU image is also an AMD, and in addition its 32 bit float processing, all apo's and 6 channels.
Below is an image of Equalizer APO, a slightly older screenshot, regardless it measures 0.0ms.



I am running a customized, minimized version of EAPO, no panel (personal reasons).

----

I have changed the RAM and GPU since but still find this amusing: https://valid.x86.fr/z741zq

I tell my friends its a platinum edition just for giggles.

====

To answer a question, yes its possible to do 16.77+ million bit RGB processing, but not in power form, 14v into 16.77+ million, not likely.
32 bit float 'dynamic range' 1528 dB, 1528 /6 = 254.66 (255 bit), I started the DMAS at 256 bit, 256 power steps.

Note that its power step and not dB with the DMAS, bit 256 = max power, not max decibels.



Position to position, and position (bit) processing.

 

[Ferather]

For me the overall audio quality is better with the newer version, plus all the Realtek fixes. I noticed added USB support in the newer DTS service.
Note that you can use my other pack as a DTS file installer, if for some reason you need alternative files, but the newer driver.

Its likely you will need the service in the other pack, but you can try the OEM files first (replace the files in APO4x).
Since the endpoint is already setup by the driver you only need to change installed files.

 

[Ferather]

Might not be too long before we get RGB Optical audio-data and the DMAS, CQRobot TCS34725FN. Optical PAM X should be possible soon, if not already possible.

RGB: 1 bit data stream = 16.77+ million values at [X] data rate | 2 bit = [16.77+ M x 16.77+ M] values at 50% of [X] data rate ([1 bit value], 2 pulse).
Note that RGB values could be converted to basic 2 value binary if needed, but the data rate will not be as good.

If you include IR and UV, you should get [R x G x B x I x U] total values, 1 bit (1 pulse).
Also, 400 Mbits/s in data would be 400 M linear pulses per second.

----

Reminder, the total bit value (colour) range does not effect the required bit rate, it stays the same. Infinite value 1 bit, [X] pulse technology.
A speaker manufacturer will not be able to put 14v into 16.77+ M steps, don't expect more than 256 [bit value] speakers.

 

[Ferather]

A 24 bit binary stream (NRZ), 24 bits = 1 value, at 400 Mbits/s data rate, would be 9600 Mbits/s data rate in PAM X, 1 bit = 1 value.
In PAM X audio this could be 24 channels (23 + 1) at 400 Mbits/s data rate each, with the same pulse rate as NRZ.



8 channels, 2 samples, 18 total 1 bit linear pulses. Bit time and sample as one master clock.

====

In terms of PAM X memory (Optical RAM), as normal, colour value to volt value, since volt value can be held.
You need variable voltage 1 bit banks, or similar, still linear, returning the same colour.

It should be possible to also do CPU to CPU bridge, potentially transistor to transistor for processing.
A full spectrum RGB CPU can input 3 bits, [4096] [x] [4096], and output 1 bit 16,777,216.

 

[Ferather]

Looking into 64-bit computing as it is now, it should be that a PAM X register would be 16.77+ million to the power of [X] bit used, for example 16.77+ m ^256.
It should also be that larger word length can be be done within a 1 bit pulse, opposed to 16 hexadecimal for only 64 bit word length.

The same rules will apply, anything using 64 pulses to 64 bits, will be 64x faster using the same rate in PAM X.
In addition you could also run 64 single routes at the same data rate as 64 x 1 bit length.

Calculator needs an upgrade.....

====

Binary's days are numbered (free pun), in favour of superposition.

 

[Ferather]

Here is a link to a 2016 paper on photon superposition, with PAM X we are using three base colours, but its still superposition in linear bit-pulse format.
In terms of audio its microphone position to speaker position without using power (as such), until the end point, the emitter.

Yes it's possible to produce a USB-C to DMAS connection via inline optical transcoder (PAM X).
Ideally the optical components should be inside the device opposed to inline.



USB X: 1 optical core, 2 copper in one cable.
2 optical cores for up-down devices.

 

[Ferather]

A DMAS could be classed as quantum audio, although quantum just tends to overcomplicate things, I prefer superposition or position-to-position.

Also make sure all standards have a version for both OEM's and consumers to follow, example DMAS 1.0, 1.1, 1.2, so on.
Secondary example: DMAS 1.0 - 256 bit values, X bitrate | DMAS 1.1 - 1024 bit values, X bitrate.

Quantum DSP would be an acceptable name for the processor SoC.

----



Increased accuracy with the same wave, no DAC.

 

[Ferather]

Based on RGB optical broadband, each fibre can carry 400 Tera bits per second, or more, so now you can imagine a USB solution for every possible device.
If all bits where superposition bits, opposed to other methods, this number will dramatically increase, same number of pulses.

For example, 1 value = 4 bit length, would be 4x 400 TB faster | 1 value = 8 bit length, 8x 400 TB faster.

Now try and do the same with a copper cable and power stages. LOL.

----

To simplify [bit value] is the [range] of [superposition values] both produced and measured, and in addition as required.
It is also possible to breakdown colour superposition into 3 direct values, range 0-255, for 3D bits.

 

[Ferather]

Post 1 updated, just the post its self, not any new files. Cleaning up the post in general for better reading.
Also note while you can update the Ultra app via the store, it wont be for this version.

====

An RGB optical module with relevant sensors should not cost more than £15-£20 retail (much less in parts).
Meaning it is the most cost effective way to utilize superposition, and PAM X (digital).

A DMAS, portable (without power supply), should also be cost effective, although the DSP will add to it.
Higher bitrate, pulse rate, will also come at an additional cost, relevant to parts used.

----

Note that increasing the base range is more desirable than increasing bit pulses to produce more values.
For example RGBI would be 256x256x256x256 (256^4) = 4,294,967,296 bit values.

====

USB X as a video connection will support some silly resolutions, and also supply the monitor power.

DMAS headphones, with built in transcoder, would be optical digital headphones.
Either direct connection, connected to a DSP, or has built in DSP.

 

[Ferather]

In terms of PAM X resolution, the more bit values and bits per second, the better (Sine, not square).
The basic 8 bit illustration, in reality would produce terrible sound, very square output.

We achieve more resolution by utilizing more [original] information.
More positions, for better position-to-position signal.




Superposition audio CPU, DSP, SoC.

----

Be sure to use 32 bit float with your media player.
Added a note to the bottom of post 1.

APO's in effect are software DSP's using system.

 

[Ferather]

In terms of bit switches (power), the superposition value is also the switches address or ID value, with audio its a wave position value.

----

ID1 = if
ID2 = then
ID3 = end
ID4 = print

local ID6 = "Pie"

1 6 2 4(6) 3

Switches which are programmed as a variable will need to be set via input before any code is executed.
Reserved bit values to represent alphanumericals and other characters for input.

And yes its possible to have an optical USB X keyboard.

----

In terms of a generic superposition bit switch CPU, it can handshake a code base (word), set variables then run.
A dedicated superposition audio DSP will be using position-to-position (PAM X), not code form.

 

[Ferather]

How RGB LEDs work and how to control colour | CircuitBread (note PWM is digital, superposition output).

Hexadecimal colour values – Finalsite Support (some extra notes, binary based system).
1185 Colour Sensor Datasheet (RGB sensor, 16.7 million values, binary based).

Reminder: Data Delivery Rates 4.5 Million Times Faster (2024).

----

PAM X: Set power (x3) > Pulse > Set power (x3) > Pulse. Each pulse is a bit value, superposition value.
Position-to-position, no binary. Bit value to bit switch. X number of values, switches.

Example: 1024 bit value audio, 1024 bit switches (power steps).

----

40 channels of audio at 400 Mbits/s would be 16 Gbits/s serial (ignoring start-end bits).
Note that is 400 Mbits/s in PAM X, a rather large number if it where binary.

The 400 Mbits/s is also the sample rate, bit value range does not impact bit rate.




PCM (binary): 16.384 Tbits/s.

 

[Ferather]

PAM X, at 48KHz (48 Kbits/s), 40 channels would be 1.920 Mbits/s. That sounds great, and it is, but 48KHz is not enough audio positions.

The best way to describe a DAC is pseudo (generated) samples, to output a sinewave, which needs a pre-amp.
Generated gap filling and rounding is NOT original sound, though it might be close.

There is no way for a DAC to produce 100% true sound if pseudo values are being generated, hence why DAC's sound different.

True Sound needs only original information, and outputs original information, no DAC or amp.

----

PAM X at the right bitrate will already have enough information, the only thing that needs to happen is transcoding to power.
100% of the original data direct to speaker, without change (unless the user has changed it, added bass).

----

Why is it a transcoder and not an analogue converter? Because its digital-to-digital (square), and position-to-position values.
The values remain the same at either end of the power format, light > power > light, and visa versa.

A sound emitter can be only in 1 position at a specific time, not more.

 

[Ferather]

Here is a general idea of what a DMAS would be like: ECdesigns PowerDAC-SX Stereo HIFI Integrated DAC — Audiophile Heaven
The DMAS design uses RGB optical in and out, but the above system is closer than any current AVR.

The general idea of the DMAS is that different models, brands, will sound the same, its the speaker you change.

----



8 channels 2 samples example input, single channel routing example output.
In this case full white has been reserved, as start-end bits.

No colour (off pulse, or fully off), specifies 0v in the speaker (do nothing).

----

If you need to calculate bit rate, its channels x sample rate, 1 bit word length.



400Mhz = 400 Mbit, 1 bit cycles.
400 Tbit = 400THz.

----

When it comes to binary and word length, just say no.

 

[Ferather]

Input start-end bits are calculated as +1 channel, there is a stream end cap, its the same reserved full white, but it only occurs once.
Channels are in linear order by default, so no extra code or stream structure is needed for multichannel.

The speaker stream has no reserved bits, and is the result of 1 channel per start-end.
The sample rate, bit rate, is also the master bit clock.

----

Position-to-position technology, position processing, to speaker position. Using superposition. Lol.

 

[Ferather]

Found the first image to be interesting then I read the whole page: How to Plot Sine Wave in Matlab | Delft Stack


Position (plot) in digital form, simple 8 bit example.

====

I wont be posting driver updates very much. The only thing to update is the Realtek files, nothing else.

If you need the alternative DTS files, you can install them using my other DTS pack.
Move the OEM files to APO4x, install then restart the computer.

Ignore the service stop error when installing, its stopped by dependency (added feature).


Nothing left to optimize, 2024-2025 DTS files don't suit all devices (unsigned files here).

 

[Lassi]

​

Ferather, thank you for your great work on these drivers.

I just installed new PC with Gigabyte X870E Aorus Master and Windows 11. I installed DTS DCH 6.0.9800.1 drivers and those are working fine in 5.1 speaker configuration through SPDIF. There is only problem with 'Stereo Mix' stream, because that prevents PC to go to sleep mode. Any possibility to fix that?