DX340 iBasso developed Discret DAC, Easily replaceable batteries, Amp Modules! ***New firmware on 1st page. v1.04 local update. ***

[Whitigir]

I Spent some more time to figure out what this 1 bit discrete and 16X8E exactly is. I was going to write up a whole bunch of Abracadabra......But that would be too much to sink in the time of a cup of coffee. Please allow me to summarize everything.

Spoiler: Potentially Hazardous and headache inducing

What is 16:4 ? Yeah, we get 4x. This is 4 active signals polarities of L+ L- R+ R- This is your 16X

What is that 8E ? Why don't other DAP/DAC advertise it ? well....it is because they were never a Discrete to begin with. They were always an integrated process. FPGA is an integrated, DAC-IC is an integrated. While FPGA based like Chord and Hiby are only Algorithim/filter integrated, the DAC-IC is both algorithm/filters + Current or Voltage summing integrated as well. This allowing FPGA a more flexibility to design their own Current/Voltage mode by themselves. But skipping the Complexities of Analog filtering (mind you, in order to design a good Analog filters at Square Waves is not an easy feast)

Discrete 1 bit is FPGA but Algorithm Digital filters, are separated and discretely done from Analog filtering which is also a LPF. Because each Polarities of L+ R+ L- R- is carrying an 8X time delayed, and then summed into 1 final signals. This is an equivalent to 8X Analog Over Sampling. But this is done at the Digital 1 bit DSD Domain. We calling it analog, because any continuous Square wave is actually 1 bit....because it can only be 1 or 0, not both ... Up or Down only...but is continuously so.

This is why other DAC/DAP do not disclose about their own Elements. They will tell you how much Oversampling they are doing, 2-4-8X ? but never the Elements. Others like Hiby is advertising 128X PWM, that is an 4X oversampling of 4x32 bits = 128. Because those 4*32 Sampling with digital filterings were all done at Once due to the integrated DSP processing from FPGA or DAC-IC, we just do not hear about the Elements. It is just an algorithms.

1 bits is calling it 8 elements, because Each 1 of these element is an element of 1 signals (8*4=32), time cycled as 1x sample, simultaneously feeding the discrete Resistor time shifting toward the next resistors in order to filter out the noises. That is why it is 8 Elements , and it takes 16X of each elements in order to have what Ibasso have....they simply stated it out loud an 16X8E....Magically, Ibasso dont want to words salad it into (XQuanTimz) Similar to DARWIN.....What exactly is DARWIN ? so what is XQuantimz ? LOLLLLL....just 16x8E is enough. It is essentially 8X overSampling but at the DSD level of 1 bit and not at the multibit processing by PWM stages

Back in the time of D16 and my inquiries for @Paul - iBasso , I was confirmed that the overall sampling rate of the D16 was 128X. This is 16*8=128X or 128 Elements. This really means that all of the Feedbacks, and sampling processing is only Done at the Analog Domains and not multibit PWM domains.

Updated: the digital domain is also 128X rate. That is 4X oversampling as well.

At least, this is what I have been gathering, allow me to share with everybody. The elements of the DACs can also be designed differently as well, for example, an FPGA with everything integrated could have 20 Cores, with each Cores to dedicatedly be processing it own PWM+OverSampling+filters together. Then this will be looked at as a 20Elements designs. But this can not be used as a way to disclose it, because these cores processing can greatly varies with different designs. With 1 bit discrete, we can clearly see the Elements and it components on the PCB board. You can count it on D16 picture

Yes, I searched around Hiby, and it appears to me that R8II is FPGA based and not 1 bit discrete, which set it greatly different to 340 in design architecture

I just found out more about the 1Bit Discrete from Paul.

The digital algorithm domain is done with 4X oversampling into 128X rate. This is as shown in the block which I missed 48*128 =6.144

It isnt NOS but OS topology

This is DSD128 which is Twice the standard DSD64 from SACD rate. Which is 4X away from Nyquist theorem. Therefore the Oversampling factor is 4X and into 128X rates.

Some people may be confused when I mentioned Analog Oversampling. So, oversampling is when 1 cycle delay is compared to the previous signals to be summed and canceling out the Error = an oversampling, or (Feedbacks). It is just a deep dive into the technologies itself. But to understand it better, simply call this an 8E analog Filters instead

 

[Whitigir]

I just found out more about the 1Bit Discrete from Paul.

The digital algorithm domain is done with 4X oversampling into 128X rate. This is as shown in the block which I missed 48*128 =6.144

It isnt NOS but OS topology

This is DSD128 which is Twice the standard DSD64 from SACD rate. Which is 4X away from Nyquist theorem. Therefore the Oversampling factor is 4X and into 128X rates.

Some people may be confused when I mentioned Analog Oversampling. So, oversampling is when 1 cycle delay is compared to the previous signals to be summed and canceling out the Error = an oversampling, or (Feedbacks). It is just a deep dive into the technologies itself. But to understand it better, simply call this an 8E analog Filters instead

The way I understand this in simpler term. The 8x of 1 bit and 4X of multibit is a total rate of 1024X. That is , if we just want to giggle about it....DSD1024X

However, the way the system is doing it is advanced, split and sharing the loads under discrete ways. We can not technically call it DSD1024 but DSD128X.

That is why the 8E discrete and 1 bit design is such an innovative design. Especially when shrinking it down and squeezing it into a Portable device

All resulting in a performances that you can witness and observe! The 340 is such an amazing design and implementation

 

[Whitigir]

Some more to Dive in. Since someone also debated with me that Ibasso isnt innovative, but perfecting the design as N7 cayin was the first to put 1 bit discrete in a DAP.

Well, the 1 bit discrete design was from DIY communities, then MolaMola then other higher end designers. So, we can keep tracing this down. But if we simply look at it this way, each designer have their own implementations and so is Ibasso

N7 is an 4X of 8E elements 32Elements

D16 and 340 is an 16X8E which is 4X of the N7 and or 128E , that is why I called out about Ibasso being innovative in their products.

Pictures as shown by Cayin and Ibasso. You can count the elements. These are dual sided

So there, you have your answer. The Ibasso 340 is 4X the elements of Cayin N7

A simpler term ? N7 is single DAC design and Ibasso is Dual

 

[voja]

Some more to Dive in. Since someone also debated with me that Ibasso isnt innovative, but perfecting the design as N7 cayin was the first to put 1 bit discrete in a DAP.

Well, the 1 bit discrete design was from DIY communities, then MolaMola then other higher end designers. So, we can keep tracing this down. But if we simply look at it this way, each designer have their own implementations and so is Ibasso

N7 is an 4X of 8E elements 32Elements

D16 and 340 is an 16X8E which is 4X of the N7 and or 128E , that is why I called out about Ibasso being innovative in their products.

Pictures as shown by Cayin and Ibasso. You can count the elements. These are dual sided

So there, you have your answer. The Ibasso 340 is 4X the elements of Cayin N7

A simpler term ? N7 is single DAC design and Ibasso is Dual

I cannot imagine anyone saying that iBasso isn't innovative. It's even more insane to imagine anyone to expect iBasso to invent a completely new DAC architecture/technology (however it's called). Cayin is a respected manufacturer, their DAPs seem to be complex compared to what others are doing. It always appeared to me that it's about the taste preference between iBasso and Cayin. As a whole, I think iBasso featured more innovation and "firsts" in its DAPs, but if we are only talking sound, I think it's a matter of preference.

I don't think there's need for a debate about this. I'm always rooting for innovation, but with some of these manufacturers, it seems like it's all about being the FIRST to use it, and not much about integrating it correctly. It's as if it's a race who gets to claim the right to say that they were place #1. Imo, iBasso always prioritized the implementation of its innovations and firsts. I mean, we are talking about a company that delayed the whole release of a product (SR2) because the adapter didn't turn out how they wanted it to... I'll leave it at that.

 

[Senka]

Good day everyone!

Its been between 30-40 hours of burn in on the DX340 and half of that was ear time if not more since it came in, I've really been putting it through all its paces.

let me get some of the bugs I found out of the way before we get to the good stuff.

If you activate always on display and choose no lock as your lock screen option the DX340 will go to Always on display then insta open into home page without waiting for tap to wake. Oh yeah... Tap to wake does not work if any of those other two options are active.

I just left double tap to wake on and left the other two off for now, Maybe included in one of the updates, added to the list.

Mango would bug and restart tracks after I took my transducers out and put them back in, Don't know if this is normal or not.

For some reason Neutron player would crash on me every time I tried to turn on 64 bit, OS seemed to work fine

Roon ARC still resamples everything to 44.1 with android... still... Bummer!

This was me yesterday for a huge portion of the day. A/B to Mojo 2+Poly pulling out the BX2 PRO as the amp (its gold plated copper case should be here next week... big hype) cutting board and some stands offs with a thermal pad it is for now.

Oh I also noticed that the refresh rate of the screen was making my camera freak out as well photo and video, I think it was the dimming dropping down the refresh rate.

For those that use Roon make sure to go on the iBasso website and download the ASIO drivers for windows, Mac is golden.

this next part is for windows users.

The WASAPI drivers will auto install off the DX340 as soon as you plug it in.... But... You must enable USB dac mode for the computer to see it as a sound card and to output through windows. If you only use the WASAPI drivers with Roon you will be hard locked to DoP and not able to play native DSD.
Grab the ASIO drivers and make sure to engage USB DAC mode when it asks you to plug it in, Just having the USB in only sees the DX340 as a hard drive to drop files into.

Why you talking about Roon this is a DAP my guy... I get it but this DAC has potential and is doing something special.

Okay some sounds impressions from the above listen.

At first things that popped out at me was how it stages... I am super sensitive to spatial cues and this is doing some magic. The DX340 is wider taller and deeper than the Mojo 2+Poly. Where it really takes the edge is how well separated the images are and the space between them. The DX340 has a more defined image edge than the Mojo 2 Poly... This also has to do with how close you are to everything with the Mojo 2+Poly, The M2P stage is more intimate than the DX340, The 340 takes that step back from the stage. Drums and how kits ware placed in the stage were the easiest to pinpoint out in the AB on how wide and tall the stage truly was. Yes that is something inherent to the transducer and good amps enhance on top of this.

M2P - Warmer, Thicker note weight, more diffuse image, more intimate a few feet closer to the stage. Class A timbre
DX340 - Accurately musical is the term I would like to use, Less bass thickness for better more articulate bass texture and response.

Tonality had me @#$%ed for the first 20 hours... It took some time for my brain to burn in but also the unit to start opening up and thicken up a little. I am far off the 2-300 hours this thing will need but damn... If these are the changes now, should be getting sweeter as the days go on. Just like everything else I have on the desk, They all sound better after some warm up.

I stuck with both high gains and H2 on for all my testing, I did notice there is a smaller jump from low to medium than there is medium to high gain for the AMP15, Adding in the 12v fills everything out that little bit more... Looking forward to finding a LPSU and a battery for portable(need to find one for CanJam)

There was either grounding noise or electrical noise every once and a while when i would use the 12V and or have the USB C in for DAC mode. I am planning on getting a grounded IC and a grounding adapter to test out, especially for when the 12v and USB are in question. It would go away as soon as I touched the case...

Yo at 550 grams this thing is Thicc, When I add it to the Pro with its case... that thig is 680 grams + 550 grams... I did not take the weight into much consideration but hey... we in it now.

Being able to find something to replace or be aside to my Mojo2+Poly has been a huge surprise, Not that I did not expect this thing to be good... Did not expect to like it this much.

Need to put some more time on it and keep er burnin in.

This thing sounds amazing being feed some DSD 256... Roon had changed my whole listening way, needing a solution that's Roonless but also can work as a desktop is next level for my particular needs. The Mojo2+Poly had this on lock but the use case outside a Roon server for android is... trash.

The one thing I am most excited for is getting back to using Tidals native discovery... So good!

As the burn in continues i'll update with any changes.

I hope everyone has a wonderful Friday

Jam On

Pozdrav

 

[FooFighter]

Besides A&K and Fiio I am not seeing DAPs getting Roon support.
There was an attempt on Cayin side but haven't heard from it for a while now.
Wondering if and when the once discussed Ibasso desktop project will see some light.
That could have the potential maybe for desktop features including Roon support, who knows.
Actually I think it's more about the certification process than a real technical limit

 

[Senka]

Besides A&K and Fiio I am not seeing DAPs getting Roon support.
There was an attempt on Cayin side but haven't heard from it for a while now.
Wondering if and when the once discussed Ibasso desktop project will see some light.
That could have the potential maybe for desktop features including Roon support, who knows.
Actually I think it's more about the certification process than a real technical limit

100% Its not just the hardware limitations for the protocols but its the endpoint propriety tech behind it.

Heck the Fiio JM21 has all to DSD iirc and they confirmed it will be getting Roon endpoint at some future update.

Right now I am okay with it not having endpoint, It works great with USB for when I'm at the desk. Mainly got the DX340 to de-tether from Roon. Big plus it can be plopped down for some desktop use. Coming from a Cyan 2 tube amp desktop chain, this is gonna be a change and for the better for my use case.

Side not i did get it to thermal throttle and lower the screen brightness last night for maybe 20-30 sec. Had it charging with the 12v.

One thing I did also notice was... When I had it in USB C DAC mode, it does drain the battery but it just stayed at 0% still running, I was expecting it to shut off, It did when I unplugged it.

iirc Paul stated if you charge the battery to 100% and keep it plugged in it has an intelligent bypass mode that shuts the battery off while pulling from the USB.

 

[pawdiophile]

Bringing this beast on the road.

 

[Loftprojection]

I got so many questions as soon as I mentioned reviewing DX340, so I decided not to wait until I get to my full review (which is still coming) and put together this detailed First Look write-up, focusing on the new design and sound analysis, along with a handful of comparisons and DAC/amp pair-ups all within iBasso family. I hope you enjoy the read.

Very nice review Twister, thanks for sharing. One thing that would be really nice to know, if you have time with your full review, is does the DAC in DX340 sounds the same or some differences to the DAC in D16. So connecting "DX340 LO to PB5" vs "DX340 to D16 to PB5".

Thanks also for mentioning coming amp cards, new AMP16 with JAN tubes has the potential to be incredible.

 

[Joe Bloggs]

I Spent some more time to figure out what this 1 bit discrete and 16X8E exactly is. I was going to write up a whole bunch of Abracadabra......But that would be too much to sink in the time of a cup of coffee. Please allow me to summarize everything.

Spoiler: Potentially Hazardous and headache inducing

What is 16:4 ? Yeah, we get 4x. This is 4 active signals polarities of L+ L- R+ R- This is your 16X

What is that 8E ? Why don't other DAP/DAC advertise it ? well....it is because they were never a Discrete to begin with. They were always an integrated process. FPGA is an integrated, DAC-IC is an integrated. While FPGA based like Chord and Hiby are only Algorithim/filter integrated, the DAC-IC is both algorithm/filters + Current or Voltage summing integrated as well. This allowing FPGA a more flexibility to design their own Current/Voltage mode by themselves. But skipping the Complexities of Analog filtering (mind you, in order to design a good Analog filters at Square Waves is not an easy feast)

Discrete 1 bit is FPGA but Algorithm Digital filters, are separated and discretely done from Analog filtering which is also a LPF. Because each Polarities of L+ R+ L- R- is carrying an 8X time delayed, and then summed into 1 final signals. This is an equivalent to 8X Analog Over Sampling. But this is done at the Digital 1 bit DSD Domain. We calling it analog, because any continuous Square wave is actually 1 bit....because it can only be 1 or 0, not both ... Up or Down only...but is continuously so.

This is why other DAC/DAP do not disclose about their own Elements. They will tell you how much Oversampling they are doing, 2-4-8X ? but never the Elements. Others like Hiby is advertising 128X PWM, that is an 4X oversampling of 4x32 bits = 128. Because those 4*32 Sampling with digital filterings were all done at Once due to the integrated DSP processing from FPGA or DAC-IC, we just do not hear about the Elements. It is just an algorithms.

1 bits is calling it 8 elements, because Each 1 of these element is an element of 1 signals (8*4=32), time cycled as 1x sample, simultaneously feeding the discrete Resistor time shifting toward the next resistors in order to filter out the noises. That is why it is 8 Elements , and it takes 16X of each elements in order to have what Ibasso have....they simply stated it out loud an 16X8E....Magically, Ibasso dont want to words salad it into (XQuanTimz) Similar to DARWIN.....What exactly is DARWIN ? so what is XQuantimz ? LOLLLLL....just 16x8E is enough. It is essentially 8X overSampling but at the DSD level of 1 bit and not at the multibit processing by PWM stages

Back in the time of D16 and my inquiries for @Paul - iBasso , I was confirmed that the overall sampling rate of the D16 was 128X. This is 16*8=128X or 128 Elements. This really means that all of the Feedbacks, and sampling processing is only Done at the Analog Domains and not multibit PWM domains.

Updated: the digital domain is also 128X rate. That is 4X oversampling as well.

At least, this is what I have been gathering, allow me to share with everybody. The elements of the DACs can also be designed differently as well, for example, an FPGA with everything integrated could have 20 Cores, with each Cores to dedicatedly be processing it own PWM+OverSampling+filters together. Then this will be looked at as a 20Elements designs. But this can not be used as a way to disclose it, because these cores processing can greatly varies with different designs. With 1 bit discrete, we can clearly see the Elements and it components on the PCB board. You can count it on D16 picture

Yes, I searched around Hiby, and it appears to me that R8II is FPGA based and not 1 bit discrete, which set it greatly different to 340 in design architecture

I just found out more about the 1Bit Discrete from Paul.

The digital algorithm domain is done with 4X oversampling into 128X rate. This is as shown in the block which I missed 48*128 =6.144

It isnt NOS but OS topology

This is DSD128 which is Twice the standard DSD64 from SACD rate. Which is 4X away from Nyquist theorem. Therefore the Oversampling factor is 4X and into 128X rates.

Some people may be confused when I mentioned Analog Oversampling. So, oversampling is when 1 cycle delay is compared to the previous signals to be summed and canceling out the Error = an oversampling, or (Feedbacks). It is just a deep dive into the technologies itself. But to understand it better, simply call this an 8E analog Filters instead

From what I'm reading, from the headline spec to these extra notes from Whitigir, the R8II's DAC section is more similar than different to the DX340. Both do oversampling in stages into very high rates, e.g. DSD128, before being converted to analog via multiple 1-bit DACs with different delays (that's what multiphase PWM in our advertising stands for) to be combined into a smooth signal. In our case the AA filter after the first 4x oversampling is customizable, that's what you choose between (or roll your own and save onto the player) in the Darwin AA filter choice.

I would like to clarify that oversampling is talking about increase in sampling rate, where 128x means the final sampling rate is 128x 44.1kHz or 48kHz. The bit depth remains constant at 32 bits.

Moreover, as seen in this photo, the FPGA of the R8II is only responsible for the digital oversampling process, whereas the downstream processes (including the 16 DACs) are discrete analog circuitry

There's also a harmonics generation thing on the DX340, for which I haven't read how it is done, but I did read from twister's review that there are five modes available--whereas on the RS6 or R8II, there is one type of harmonics generation, for which there is a 256 position slider.

I am posting simply to clarify technical points raised regarding our own product.

 

[Whitigir]

From what I'm reading, from the headline spec to these extra notes from Whitigir, the R8II's DAC section is more similar than different to the DX340. Both do oversampling in stages into very high rates, e.g. DSD128, before being converted to analog via multiple 1-bit DACs with different delays (that's what multiphase PWM in our advertising stands for) to be combined into a smooth signal. In our case the AA filter after the first 4x oversampling is customizable, that's what you choose between (or roll your own and save onto the player) in the Darwin AA filter choice.

I would like to clarify that oversampling is talking about increase in sampling rate, where 128x means the final sampling rate is 128x 44.1kHz or 48kHz. The bit depth remains constant at 32 bits.

Moreover, as seen in this photo, the FPGA of the R8II is only responsible for the digital oversampling process, whereas the downstream processes (including the 16 DACs) are discrete analog circuitry


There's also a harmonics generation thing on the DX340, for which I haven't read how it is done, but I did read from twister's review that there are five modes available--whereas on the RS6 or R8II, there is one type of harmonics generation, for which there is a 256 position slider.

I am posting simply to clarify technical points raised regarding our own product.

Thank you Joe, glad to see another active Rep! I was corrected on your Darwin design. So R8ii with DARWIN is indeed a 1 bit discrete DAC and not just like Chord FPGA.

 

[ginandbacon]

So, my DX320 died 2 weeks ago. Just unplugged the USB C cable while it was charging and it must have shorted something because the screen is gibberish and unresponsive to touch. I thought about connecting to it via ADB but I doubt that's going to tell me anything useful..

How does the DX260 compare to the DX340? Seems like the DX260 is lighter and easier to carry around. While the DX340 is certainly smaller than the DX320 it sounds like it's pretty heavy. Obviously that's not the important part though.

I really liked the ROHM DAC's in the DX320 but iBasso is probably the most innovative of DAP makers IMO. They are REALLY proud of their FPGA Master 2.0 and now 3.0 with the DX260 claiming pecoseond control of the DAC's, and that's with 2.0. The fact that they went with their own DAC's, or whoever they outsourced it to, for their new reference DAP speaks volumes in their confidence in it. Obviously for years it's been the same 23 vendors, with ROHM being in very few products.

Right now I bought a FIIO KA-17 USB dongle DAC which I must say, is excellent for any IEM and most over the ear/classes back headphones. With that said, I always liked having a DAP to escape the constant smartphone announces . Yes, you can just use Mango or put your phone in airplane mode but sometimes its nice to not have those options if you know what I mean. I also wouldn't rule out the DX170 but obviously with the roughly 300 difference between the DX170 and DX260, and just over 500 difference between the DX260 and DX340.

Unfortunately audio is the one thing that's close to impossible to really judge until you hear it for yourself. Reviews and user feedback help but personal preferences and IEM's headphones used can also sometimes make a difference in the end result.

 

[DAPpower]

So, my DX320 died 2 weeks ago. Just unplugged the USB C cable while it was charging and it must have shorted something because the screen is gibberish and unresponsive to touch. I thought about connecting to it via ADB but I doubt that's going to tell me anything useful..

How does the DX260 compare to the DX340? Seems like the DX260 is lighter and easier to carry around. While the DX340 is certainly smaller than the DX320 it sounds like it's pretty heavy. Obviously that's not the important part though.

I really liked the ROHM DAC's in the DX320 but iBasso is probably the most innovative of DAP makers IMO. They are REALLY proud of their FPGA Master 2.0 and now 3.0 with the DX260 claiming pecoseond control of the DAC's, and that's with 2.0. The fact that they went with their own DAC's, or whoever they outsourced it to, for their new reference DAP speaks volumes in their confidence in it. Obviously for years it's been the same 23 vendors, with ROHM being in very few products.

Right now I bought a FIIO KA-17 USB dongle DAC which I must say, is excellent for any IEM and most over the ear/classes back headphones. With that said, I always liked having a DAP to escape the constant smartphone announces . Yes, you can just use Mango or put your phone in airplane mode but sometimes its nice to not have those options if you know what I mean. I also wouldn't rule out the DX170 but obviously with the roughly 300 difference between the DX170 and DX260, and just over 500 difference between the DX260 and DX340.

Unfortunately audio is the one thing that's close to impossible to really judge until you hear it for yourself. Reviews and user feedback help but personal preferences and IEM's headphones used can also sometimes make a difference in the end result.

The new discrete 1 bit DAC sounds quite different than the ROHM implementation in the DX320. Where the DX320 sounds lean and technical, the DX340 sounds robust and organic.

The weight of the DX340 is quite manageable, if you factor in the width and height of the player the weight will feel rather balanced compared to something like the DX300MAX/DX320MAX.

 

[Aetherhole]

The new discrete 1 bit DAC sounds quite different than the ROHM implementation in the DX320. Where the DX320 sounds lean and technical, the DX340 sounds robust and organic.

The weight of the DX340 is quite manageable, if you factor in the width and height of the player the weight will feel rather balanced compared to something like the DX300MAX/DX320MAX.

Coming from the DX320Max, the DX340 is a featherweight and puny… really no complaints from me.

 

[twister6]

Very nice review Twister, thanks for sharing. One thing that would be really nice to know, if you have time with your full review, is does the DAC in DX340 sounds the same or some differences to the DAC in D16. So connecting "DX340 LO to PB5" vs "DX340 to D16 to PB5".

Thanks also for mentioning coming amp cards, new AMP16 with JAN tubes has the potential to be incredible.

Yes, I will keep that in mind, of course! One interesting thing, while their 1bit DAC design should be the same, the D16 max LO spec is 4Vrms (4.4mm) and 2Vrms (3.5mm) while the DX340 max LO spec is 2Vrms (4.4mm) and 1Vrms (3.5mm). Don't want to speculate, but it does make me wonder if the LO voltage was lowered for compatibility with amp modules to avoid saturating the input.