Topping DX5 II

[Sokel]

after reading your thread I confirmed that this is a bug...
it may caused by not using the correct precision when doing the computation.
jds element iv, using the same chip, has transparent dsp result. (see attached)

I don't know the chip they use for DSP but if it's XMOS 's shared abilities I guess jds has found a way to dedicate more horsepower to it.
To get really transparent, RME-grade results though, one needs to use a dedicated DSP chip as they do.

Of course I would argue about the audibility and the result will probably be ok, it's the technical excellency we're talking here.
Proper DSP is a whole different game, both by gear AND users.

 

[Bleib]

Topping store has just sent my black unit, they gave me an accurate shipping time, 1-2 weeks to change the color

 

[audiofun]

I don't know the chip they use for DSP but if it's XMOS 's shared abilities I guess jds has found a way to dedicate more horsepower to it.
To get really transparent, RME-grade results though, one needs to use a dedicated DSP chip as they do.

Of course I would argue about the audibility and the result will probably be ok, it's the technical excellency we're talking here.
Proper DSP is a whole different game, both by gear AND users.

the fpga and dsp chips RME uses for EQ are very, very, very old (10+ years old). too old that in the past year they were forced to swap it with newer versions of the chips, because the old model chip is no longer being manufactured.
the latest xmos chip Topping, JDS, and many other manufacturers use has dedicated dsp processors and the hardware is many generations newer thus more powerful.
proper EQ is not rocket science. A good firmware programmer can produce bug free code in an hour. There're many sample implementations on the internet. XMOS also provided reference implementation to their clients as well.

Contrary to your belief, RME's solution is not many times more powerful than current Topping/JDS offerings.
Even ESS's ADC chip has much more performant EQ capabilities than RME's solution.
I'm not saying RME products are worse. I love them very much. I'm just talking from silicon perspective.
Similar to Nintendo, they don't need to use the best chip to win their customers.
RME's advantage is they control everything, from usb audio implementation to dsp/routing, to software/firmware and drivers, so they are not depended on certain suppliers.
And they can work on some custom technologies no one else can.

audibility may suffer once user have many low frequency eq bands.

 

[Roland68]

the fpga and dsp chips RME uses for EQ are very, very, very old (10+ years old). too old that in the past year they were forced to swap it with newer versions of the chips, because the old model chip is no longer being manufactured.
the latest xmos chip Topping, JDS, and many other manufacturers use has dedicated dsp processors and the hardware is many generations newer thus more powerful.
proper EQ is not rocket science. A good firmware programmer can produce bug free code in an hour. There're many sample implementations on the internet. XMOS also provided reference implementation to their clients as well.

Contrary to your belief, RME's solution is not many times more powerful than current Topping/JDS offerings.
Even ESS's ADC chip has much more performant EQ capabilities than RME's solution.
I'm not saying RME products are worse. I love them very much. I'm just talking from silicon perspective.
Similar to Nintendo, they don't need to use the best chip to win their customers.
RME's advantage is they control everything, from usb audio implementation to dsp/routing, to software/firmware and drivers, so they are not depended on certain suppliers.
And they can work on some custom technologies no one else can.

audibility may suffer once user have many low frequency eq bands.

RME is one of the best examples that high-quality implementation and software are far more important than simply components with theoretically better performance.
But unfortunately, this is also much more expensive to develop.

 

[Sokel]

the fpga and dsp chips RME uses for EQ are very, very, very old (10+ years old). too old that in the past year they were forced to swap it with newer versions of the chips, because the old model chip is no longer being manufactured.
the latest xmos chip Topping, JDS, and many other manufacturers use has dedicated dsp processors and the hardware is many generations newer thus more powerful.
proper EQ is not rocket science. A good firmware programmer can produce bug free code in an hour. There're many sample implementations on the internet. XMOS also provided reference implementation to their clients as well.

Contrary to your belief, RME's solution is not many times more powerful than current Topping/JDS offerings.
Even ESS's ADC chip has much more performant EQ capabilities than RME's solution.
I'm not saying RME products are worse. I love them very much. I'm just talking from silicon perspective.
Similar to Nintendo, they don't need to use the best chip to win their customers.
RME's advantage is they control everything, from usb audio implementation to dsp/routing, to software/firmware and drivers, so they are not depended on certain suppliers.
And they can work on some custom technologies no one else can.

audibility may suffer once user have many low frequency eq bands.

It takes a lot more of a good firmware programmer for proper DSP implementation, as you read at the thread I linked it took miniDSP years to apply a partial fix even though its forum was full of complains all this time for example.

Combining the right elements is the key, yes, along with not allowing (or at least warning) end-users about issues like bad filter stacking, etc.
People designing the DSP implementation and interface with formal studies about it are rare and well-payed (obviously) .

There are scattered posts about it in many threads, conclusion is the same: DSP is a field that not many (of the formally educated about it) give for little or free.

 

[audiofun]

RME is one of the best examples that high-quality implementation and software are far more important than simply components with theoretically better performance.
But unfortunately, this is also much more expensive to develop.

It takes a lot more of a good firmware programmer for proper DSP implementation, as you read at the thread I linked it took miniDSP years to apply a partial fix even though its forum was full of complains all this time for example.

Combining the right elements is the key, yes, along with not allowing (or at least warning) end-users about issues like bad filter stacking, etc.
People designing the DSP implementation and interface with formal studies about it are rare and well-payed (obviously) .

There are scattered posts about it in many threads, conclusion is the same: DSP is a field that not many (of the formally educated about it) give for little or free.

You are both right.

That's why I prefer open source solutions --- because the capable people can improve them in free time.
Unfortunately due to the competitive nature, no manufacturer in the hifi industry did that.

This may change though. In the past it was very hard to hire good firmware/software programmers outside of the tech industry, because those people preferred to seek for much higher paying jobs in the silicon valley.
Working for small, unknown companies like Topping was regarded as career suicide.
Now the mass tech layoff changes everything. It's much cheaper to hire high quality applicants.

With Gen-AI boom we may see more and more high paid jobs gone forever, and we can enjoy cheaper and better products. Until no one has job to afford those products

 

[magor34]

Dumb question perhaps, but is there any reason, besides volume, and if you have the amp capability, not too much of a concern, why you'd want boost in PEQ over cutting the frequencies that you don't want boosted. Wouldn't that net out to the same result, and if so, would that bypass the noise issues?

In a professional setting it's most common to do the subtractive approach with your equalizer. In general though there's no bad way to EQ really. If it sounds good, all good.

One could do a wide cut to all frequencies excluding sub and/or mid bass ones however if you already have adjusted other bands to your taste they would likely need additional adjustments too cause their relative decrease in the EQ is lessened.
Unless you already got a flat EQ, a bass boost is easier if you know which frequencies you'd like some more off. Just my two cents though.

Only thing you need to be cautious of when boosting any given frequency is to make sure you don't run into any clipping issues.
If for example you're using the Peace plugin with Equalizer APO, the graph window will tell you how much you should lower the volume in the app.

 

[Tano]

Well, I just bought one in white for 224€. So I guess I'll find myself if the bug really happens at normal listening levels when using my EQs.

 

[REVolution]

@Tano where did you bought it?

Are there already any Pictures of the PCB available?

 

[Tano]

@Tano where did you bought it?

Are there already any Pictures of the PCB available?

Bought it on AliExpress from a store called Brightaudiostore. No PCB pics unfortunatelly, just the regular ones of the 3 colours.

 

[Roland68]

@Tano where did you bought it?

Are there already any Pictures of the PCB available?

Bought it on AliExpress from a store called Brightaudiostore. No PCB pics unfortunatelly, just the regular ones of the 3 colours.

This was released in April in connection with the FCC certification, presumably still a pre-production model.

 

[profligate]

I just realised this is usb-a rather than C, damn.

I know a is sturdier and preferred by some, some I was planning to connect a tiny usb-c switch to switch between two sources, A is going to be a more cumbersome to change.

 

[Roland68]

I just realised this is usb-a rather than C, damn.

I know a is sturdier and preferred by some, some I was planning to connect a tiny usb-c switch to switch between two sources, A is going to be a more cumbersome to change.

USB-C is just a connector, not a standard. The DAC has USB 2.0, and you can simply use a USB 2.0 cable with a USB-B and USB-C connector for the switch.
If the device had a USB-C port, absolutely nothing would change.

 

[pieterv1]

I just realised this is usb-a rather than C, damn.

I know a is sturdier and preferred by some, some I was planning to connect a tiny usb-c switch to switch between two sources, A is going to be a more cumbersome to change.

Why not just buy the appropriate USB C to B cable?

I would never trust any device advertised as a USB "switch" though - this is far from standard and you never know what kind of issues or limitations you might run into. I would probably stick with USB from one source, coax or toslink from the other, allowing you to easily switch sources with the input button at the front panel.

 

[pieterv1]

Well, I just bought one in white for 224€. So I guess I'll find myself if the bug really happens at normal listening levels when using my EQs.

Wow how did you manage to get that price?

 

[Tano]

Wow how did you manage to get that price?

AliExpress coupons.

 

[Berwhale]

I would never trust any device advertised as a USB "switch" though - this is far from standard and you never know what kind of issues or limitations you might run into. I would probably stick with USB from one source, coax or toslink from the other, allowing you to easily switch sources with the input button at the front panel.

Far from what standard? I been using a UGreen USB switch for 5 years to share peripherals between my home and work PCs. I've never had any issues with it.

Amazon.co.uk

Almost all KVM switches include USB switching as USB has been the standard for keyboards and mice for many years. Are you suggesting these should be avoided as well? (I'll admit that some of the early USB KVMs could be a little flaky, but that was a couple of decades ago!)

*edit* 5 years sounds a bit vague, it's actually 5 years and 1 day

 

[somebodyelse]

Far from what standard? I been using a UGreen USB switch for 5 years to share peripherals between my home and work PCs. I've never had any issues with it.

https://www.amazon.co.uk/dp/B01N6GD9JO

That one fails to comply to USB standards because it uses Type A sockets where it should be using Type B on the computer side connection. Apart from that it may be fine - I've never had any problems with the USB switch I use either. That said, mine does properly switch so the computer sees the devices as having been connected or disconnected each time you switch. For the KVM mouse/kbd connections they may emulate the devices on the computer side and switch which one they pass the events to rather than just switching the device, so the computer sees a connected mouse or keyboard all the time. They may also snoop on the keyboard events so they can use a hotkey combination for switching so you don't need separate switching button(s) to control the KVM.

 

[ziggurcat]

I just realised this is usb-a rather than C, damn.

USB-B, actually.

 

[Berwhale]

That one fails to comply to USB standards because it uses Type A sockets where it should be using Type B on the computer side connection.

I'm not so sure... The switch does not communicate with the host, so it is not a 'USB device'. So why would the USB device connection standards apply to it?

I wonder what the standards think about this type of thing...