Latest Thread Images

SONY NW-WM1Z / WM1A
Status
Not open for further replies.
Jan 4, 2020 at 2:04 AM Post #31,066 of 45,723
On why 30years older DAC chips sound better than modern high speed Delta shitgema dacs, this is well explained by the audiophile legend, the late Ken Ishiwata of Marantz.(5min 45sec of video)
 
Last edited:
Jan 4, 2020 at 3:15 AM Post #31,067 of 45,723
I've read the document on the signatute of @Sonywalkmanuser , pretty interesting read of the S-MASTER, the anti-jitter precision in attoseconds, wow, that is even lower than these femto clocks... and the desktop version of S-Master has high voltage (or current) FETs in the output secrion, making the desktop amp capable of outputting 75W per channel for powering speakers.

I think sony can implement a mobile version of these Switching FETs to put more piwer in a walkman f.e 1W or 800mW
 
Jan 4, 2020 at 6:41 AM Post #31,069 of 45,723
Jan 4, 2020 at 7:06 AM Post #31,071 of 45,723
aceedburn said:
There isn’t such a function on the WM1A.
Click to expand...

yes there is and it is mentioned how several times in this thread but I can’t remember exactly how, something about pressing the power and volume buttons simultanious
 
Jan 4, 2020 at 7:29 AM Post #31,073 of 45,723
simon740 said:
yes I tried this...But I dont know what Im doing wrong..
Click to expand...

https://www.head-fi.org/threads/sony-nw-wm1z-wm1a.815841/page-1253#post-14080788

screen off/ vol up/ vol down/ hold power till screen lights up... connect WM-1A to computer look in Walkman storage (not SD)

screenie should be in that folder

I just did it

screenshot_20200104_072933786.png
 
Jan 4, 2020 at 7:30 AM Post #31,074 of 45,723
gerelmx1986 said:
I've read the document on the signatute of @Sonywalkmanuser , pretty interesting read of the S-MASTER, the anti-jitter precision in attoseconds, wow, that is even lower than these femto clocks... and the desktop version of S-Master has high voltage (or current) FETs in the output secrion, making the desktop amp capable of outputting 75W per channel for powering speakers.

I think sony can implement a mobile version of these Switching FETs to put more piwer in a walkman f.e 1W or 800mW
Click to expand...

Sony really have alot of technical prowess that most other smaller companies do not have. Maybe exceptions include Samsung (AKG) and Harmon International(JBL), THX and Dolby.

From this article, it seems that Sony DSEE HX AI is not just a simple automatic mode selection but more like performing a continuous real time adjustment to the oversampling mode. And Sony actually uses music from their record label to allow their machine AI to analyze big data and to learn from it.

https://www.sony.jp/feature/products/dseehx/

sony
検索
Product information / store
General support and inquiries
My Sony
Corporate Customers

New `` DSEE HX '' that approaches high-resolution sound quality further by deep neural network
"DSEE HX" brings high-resolution quality realism to CDs and compressed sound sources by predicting and restoring information inherent in music data. This technology has evolved further in the fall of 2018 by incorporating AI technology. The details of the evolution will be introduced by the engineers involved in the development.


Solving DSEE HX Challenges with Advanced AI Technology

The new “DSEE HX” is
a three-dimensional sound that adds “feeling of height” realized only by the Sony Group.


Continuous playback time is the same as before while performing high resolution processing in real time

Solving DSEE HX Challenges with Advanced AI Technology
Before we talk about the new and improved DSEE HX, can you first explain what DSEE HX is all about?
Chinen: From Sony Walkmans, wireless speakers, AV amplifiers, etc. released after 2013, Sony's proprietary functions are used to upscale CD and compressed sound sources to high-resolution sound sources equivalent to “high-resolution”. is. In such a sound source, the frequency spectrum of the sound is impaired in an unnatural manner at high frequencies, but this function predicts and restores it from the frequency spectrum of the bass. Three years later, in 2016, a new operation mode switching is supported. By selecting the operation mode (5 types of standard / female vocal / male vocal / percussion / strings) according to the type of song, the user will be able to obtain the effect of upscaling more effectively. I am.

Yuki Yamamoto, Sony R & D Center
Elemental Technology Development Division

Sony R & D Center
Elemental Technology Development Division
Toru Chinen

Upscaling with DSEE HX (TM) (image)

Despite its widespread use, many people have not yet switched to high-resolution audio sources, and there are many songs that do not have high-resolution audio sources in the first place, so this is a very effective function. And it is said that the new "DSEE HX" incorporating AI technology will be installed in the audio product autumn / winter model of 2018, how exactly is it evolving?
Chinen: After the launch of "DSEE HX" in 2013, we received many opinions from users who purchased it and professional engineers involved in music production. In response to this, the 2016 operating mode switching support was realized. The 2018 update includes fundamental improvements to make it work more effectively.

That is, within one song, the operation mode automatically switches according to the scene of the song. With the conventional `` DSEE HX '', the voice of the vocal sounds beautiful in the `` Female Vocal '' mode, but there is a problem that in voiceless scenes such as interludes, the effect cannot be maximized. did. The new "DSEE HX" is challenging a method of automatically performing optimal upscaling while analyzing the contents of songs in real time.

Please explain why you decided to use AI (Artificial Intelligence) there.
Chinen: In the world of search, the approach called deep neural network (DNN) is attracting attention. This is the state-of-the-art AI technology that analyzes more digital data to generate more useful search results. I thought this could be used to identify songs and analyze scene changes.

Solving DSEE HX Challenges with Advanced AI Technology

The new “DSEE HX” is a three-dimensional sound that adds “feeling of height”
realized only by the Sony Group
Next, please tell us how you built AI for DSEE HX.
Chinen: Building a good DNN requires a process of “learning” with large amounts of data and appropriately “evaluating” the results. And on both sides, the power of the Sony Group, which has a music label, was very helpful. Sony Music Entertainment uses many high-resolution audio sources to create optimal algorithms. As a technical issue, it was very simple. We upscaled a CD sound source and a compressed sound source, and worked as a numerical target to see how close it was to a high-resolution sound source.

What kind of ingenuity is needed to make a good DNN?
Yamamoto: In order to create an optimal DNN, not only the number of songs but also the breadth of genres is important. In addition, regarding the evaluation of the upscaling effect, not only numerical comparison (objective evaluation), but also subjective evaluation by experts such as mixing engineers and mastering engineers. In addition to just driving in the numbers, the parts that do not appear in the numbers are actually adjusted based on the valuable opinions of professionals involved in making such sounds. This is something that only the Sony Group could do. I am confident that it is unique to Sony with its content and know-how.

With the new "DSEE HX", it is said that the operation mode is automatically switched according to the scene of the song within one song, please tell me a little more about this. First, is it correct to understand that users no longer need to specify the mode according to the tune to be played as before?
Yamamoto: Yes, exactly. In addition, the mode is not as simple as the conventional five operation modes, and the optimal upscaling is performed every time for various scenes. It may be too much to say "there are endless modes", but the vocal scene is not as simple as vocal mode.

Can you tell us a little more clearly how it can be improved?
Chinen: So far, "DSEE HX" had the effect of complementing the "expansion" and "depth" of the sound, but the sound actually has another element of "height". I knew that if I did it well, I could get a more three-dimensional effect in playing the music.

So how do you get a sense of height? What is important here is the dynamics of the sound, specifically the onset of the sound of the percussion instrument. In the update two years ago, we prepared a mode called `` percussion '' to improve this even a little, but as a countermeasure, this time in the scene where vocals are singing, maximize the voice There was a problem that the sound quality could not be improved.

In that respect, the latest `` DSEE HX '', which can apply the optimal up-conversion according to the scene, makes the vocal voice beautifully echo according to the scene of the song, and at the same time, the rise of percussion sound is also very good, You can feel the height more firmly. When it comes to high res, it's easy to talk about the width of the sampling frequency of the sound, but the dynamics of the sound largely depends on this rising part. By reproducing this beautifully, I am confident that the sound will have a three-dimensional effect. This is where you want to listen and compare.

By the way, this is a simple question, but what level of compressed sound source does DSEE HX support? For example, can it reproduce beautifully 96kbps sound sources that were created in the MP3 era 20 years ago?
Yamamoto: Of course. There is no doubt that the better the original sound quality, the closer to the high-resolution sound quality, but "DSEE HX" can firmly improve the sound quality of old, high-compression sound sources. This effect is further improved with the latest DSEE HX. "DSEE HX" allows you to enjoy all your existing sound sources with higher sound quality.

The new “DSEE HX” is good for a three-dimensional sound with the added “feeling of height” realized only by the Sony Group.

Continuous playback time is the same as before
while performing high quality sound processing in real time
Please tell us about the technical difficulties in developing the new DSEE HX.
Yamamoto: Actually, DNN is becoming very popular in the home appliance and IT world, but services using general AI are not in the device in front of us, but in the cloud (servers on the Internet). Most of them send data, process the data with the abundant machine power, and return it. However, this method cannot be used with "DSEE HX", which needs to analyze music data in real time. On the other hand, portable devices such as a Walkman do not have that much processing power and affect battery consumption. During the development of the DSEE HX, it was difficult to reduce power consumption while maximizing the performance of DNN.

If you try to make full use of AI, you will inevitably increase power consumption.
Chinen: With "DSEE HX", which continuously analyzes the music scene in real time, that is an inevitable side effect. Therefore, we are pursuing low power consumption by accumulating fine ideas when mounting it on products.

Please tell us specifically what kind of “ingenuity” you did.
Yamamoto: In the early stages of development, neglecting power consumption, the "ultimate version" aims to achieve the highest sound quality by incorporating the most advanced AI technology learned in dissertations and the Sony Group's knowledge and plenty of knowledge. Create. Of course, this cannot be integrated directly into the product. The battery would run out in a matter of minutes if left untouched (laughs). Then, I thought about how to reduce the power consumption to a level that can be used on mobile devices.

The first step was to change the amount of information to be processed. In the "Ultimate Version", there was a place where the performance was improved by inputting extremely detailed information, but it was found that even if it was limited to some extent, the effect was hardly reduced.

In addition, most of the latest AI papers that we referenced during development focused on images, so there was room for reduction. Images and audio have very different signal characteristics, so different approaches must be taken. I noticed that I was able to reduce power consumption even further by making the DNN specialized for audio.

Chinen: Eventually, even with the new DSEE HX built-in, we were able to achieve the same continuous playback time as the previous model (in the case of the Walkman NW-A50 series).

The ability to use server-level AI processing on mobile devices such as a Walkman without increasing the size of the built-in battery seems to be an important enhancement of the new DSEE HX. Finally, do you have a message for your readers?
Yamamoto: The new "DSEE HX" greatly enhances sound quality and usability, while maintaining continuous playback time. We pride ourselves on those who have never had a DSEE HX-equipped player, as well as those who have existing DSEE HX-equipped equipment.

Chinen: One of the major improvements is that not only has the sound quality improved, but also the need to switch modes has been eliminated. I think that the effect can be felt especially by those who throw in songs of various genres to the player and listen while zapping. Again, the "three-dimensional effect" of the sound has greatly increased, so please check the effect at stores and other places.
Click to expand...
 
Last edited:
Jan 4, 2020 at 7:41 AM Post #31,075 of 45,723
Sonywalkmanuser said:
Sony really have alot of technical prowess that most other smaller companies do not have. Maybe exceptions include Samsung (AKG) and Harmon International(JBL), THX and Dolby.

From this article, it seems that Sony DSEE HX AI is not just a simple automatic mode selection but more like performing a continuous real time adjustment to the oversampling mode. And Sony actually uses music from their record label to allow their machine AI to analyze big data and to learn from it.

https://www.sony.jp/feature/products/dseehx/
Click to expand...


Id love to see this new dseehx in wm1a and 1z with a new update that would be so cool. So until they release the new successor we have some goodies to play with!
 
Jan 4, 2020 at 9:39 AM Post #31,077 of 45,723
captblaze said:
https://www.head-fi.org/threads/sony-nw-wm1z-wm1a.815841/page-1253#post-14080788

screen off/ vol up/ vol down/ hold power till screen lights up... connect WM-1A to computer look in Walkman storage (not SD)

screenie should be in that folder

I just did it

Click to expand...

simon740 said:
Thank you...I will try later
Click to expand...

That may work. Still the way I do it is different. You have the screen on, as you may want to capture the title scrolling across. You hit “volume down” then “volume up” then power. When you know you’ve held the power button long enough, it will have a message appear that asks if you want to power down; you then push “no”....and you have a screen capture with the screen on the whole time.

Found it here.
https://community.sony.co.uk/t5/portable-audio/how-to-take-screenshot-walkman-nw-zx300/td-p/2453777
 
Last edited:
Jan 4, 2020 at 5:08 PM Post #31,079 of 45,723
Sonywalkmanuser said:
Found another whitepaper from Sony on how they eliminate jitter and how conversion is done in detail. Starts from page 5.

https://docs.sony.com/release/SCD1_TWP.pdf
Click to expand...

Jitters! There is no one who understands and take measurements to negate jitters as much as Sony. Not only Jitters but also Errors. These Errors come from many different technical reasons.

Noises is error
Timing accuracy, phase noises, power supply noises, the pulsations and switching of the power supplies ripples ...etc.

Yes, Chord is one of the many companies to Argue that with switching power supply, the ripples and noises are out of human hearing range. But then what is Noise shaping and noise filterings in the DSP?

They are a techniques to add noises similar to Dithering, in order to allow the algorithms to eliminate errors better, and to reproduce a more accurate dynamic, amplitudes of the original sin Waves. They are white noises, and they will be oversampled, multiplied and amplified ways much that you will need Low Bit Filter 35K, 45Khz, 65Kz.... etc

So, tell me again...between Sony who takes Noises as a form of Errors VS someone who thinks as long as the noises is out of human hearing range, it is no concern ?

Anyways, I already mentioned this, while in Walkman, the usages of Polyester Multi-Filmed Layer Capacitors are necessary, the DMP Z1 are also full of it. Why ? Because they are self dampening and Minimize the Mechanical Vibrations when a current or voltage is running through it....such Mechanism is called Piezoelectric. That is how Crystal Oscilliators becomes a Clocking devices. Unwanted-Mechanical vibrations can add in unwanted values of electricity. It would be meaningless to have a precisely clocked Data retrieved but then having a lot of noises and errors down the chains.

There was a period when I started getting to know digital music. I bought in the ideas from A&K about their Femto Clocks etc....then other people also follow in their advertising. It is important that a precise clock is needed in digital worlds. However, what is more important is the phase noises which is hugely correlated to operating temperatures, and the surrounding components of everything else that may generate it own noises, for example, the Printed Board Mask (Hence it is Glass Epoxy), the Surface mounting capacitors (PML and minimize Ceramic caps), the SMD Electrolytic Capacitors, the Composite Wax Epoxy around these SMD capacitors to help dampen the vibrations.

This is why I always call out to Sony “welcome back! The king of Digital music”. Once again, the DMP Z1 isn’t just an engineering marvel, but it is a dare, a challenge toward many other manufacturers designers on “how many details and aspect of different bugs that could degrade the sound and is being polished by Sony. Not only that but DSEE-HX AI , DSD Remastering, Vinyl Processor, Walkman OS, are another saying that Not only Hardware is important but Software and Algorithms as well”
 
Last edited:
Jan 4, 2020 at 5:33 PM Post #31,080 of 45,723
Found another good whitepaper on how Sony S-master works, very intriguing for those interested:
https://www.docs.sony.com/release/strda9000es_twp.pdf

For those who don't understand what is jitter and the effects of jitter, watch this video and you will get a better understanding:
 
Last edited:
Status
Not open for further replies.

Users who are viewing this thread

Total: 9 (members: 0, guests: 9)
Back
Top