[GRUMPYOLDGUY]

Didn't Rob state about one year ago that getting more FPGA performance in a Mojo size/design is impossible? So would this be a micro-USB to USB-C and other cosmetic refresh update only?

Aren't they using a Xilinx Artix part? That's spun on a 28nm process. Kintex and Virtex UltraScale architectures are on a 20nm process and USP families on a 16nm process. There is without question higher density options on faster fabric.

I'm sure Altera has some good offerings as well, but there would be some porting needed since the DSP block architectures are a little different.

Higher cost? Without question. Better performance? Without question.

 

[Kentajalli]

Aren't they using a Xilinx Artix part? That's spun on a 28nm process. Kintex and Virtex UltraScale architectures are on a 20nm process and USP families on a 16nm process. There is without question higher density options on faster fabric.

I'm sure Altera has some good offerings as well, but there would be some porting needed since the DSP block architectures are a little different.

Higher cost? Without question. Better performance? Without question.

Actually, I do remember Mr. Watts saying (was it in an interview??) that he could get more out of Mojo, if it wasn't because of his limited miliwatt budget (power usage) and heat generation within Mojo.
I think if Mojo starts to be a bit better, the gap between Mojo and Hugo2 might get narrower, so I wonder if politics has played any part with delayed Mojo upgrade launch.
I wonder ...

 

[Kentajalli]

A side experiment:
- My equipment at the moment is Mojo, USB connected to Huawei mate 20 pro. Sennheiser IE 400 pro (slightly modified).
Interconnect is a DIY 750mm long, very thin cable with a small ferrite bush near DAC end. I use Neutron media player on the phone as player.
I was playing around with options within Neutron, specifically the oversample function. (settings>audio hardware>oversampling)
Neutron can also do a simple resample, meaning outputting any bitrate from any bitrate, which means nothing! it is only there for compatibility with various DACs.
The oversampling, on the other hand is a smart upsampling, it works in 2X , 4X, 8X ...
It is CPU intensive, and at 16X can get unstable, lowering latency under audio hardware helps sometimes.
So here is the issue:
I can swear that 4X and 8X oversampling within the media player improves the sound, while everything tells me it shouldn't!
I do know that Mojo upsamples, and it does a wonderful job of it, and the fact that I am oversampling before Mojo gets its hands on the signal is not wise.
But after a week or so, there is no denying, it improves the sound.
The difference is subtle but very noticeable. the tone is the tone, it does not change that.
What does happen is more 3D stereo imaging. the sound is no longer between the ears, but I get an out of head imaging.
Instruments seem to be better focused in a 3D environment - the sense depth gets more pronounced.
For example, the solo piano at the begining of Al Stewart: Year of the cat , was usually a little congested piano panned to far left.
With oversampling at 8X, it becomes an actual piano in a room with echoes of that room about couple meters away!
Otherwise resolution and tonal colour does not get altered.
I posted something on Neuron thread here in Head-fi but got no replies.
I wonder if anyone has had similar experience here, they'd be willing to discuss.

 

[Johnfg465vd]

A side experiment:
- My equipment at the moment is Mojo, USB connected to Huawei mate 20 pro. Sennheiser IE 400 pro (slightly modified).
Interconnect is a DIY 750mm long, very thin cable with a small ferrite bush near DAC end. I use Neutron media player on the phone as player.
I was playing around with options within Neutron, specifically the oversample function. (settings>audio hardware>oversampling)
Neutron can also do a simple resample, meaning outputting any bitrate from any bitrate, which means nothing! it is only there for compatibility with various DACs.
The oversampling, on the other hand is a smart upsampling, it works in 2X , 4X, 8X ...
It is CPU intensive, and at 16X can get unstable, lowering latency under audio hardware helps sometimes.
So here is the issue:
I can swear that 4X and 8X oversampling within the media player improves the sound, while everything tells me it shouldn't!
I do know that Mojo upsamples, and it does a wonderful job of it, and the fact that I am oversampling before Mojo gets its hands on the signal is not wise.
But after a week or so, there is no denying, it improves the sound.
The difference is subtle but very noticeable. the tone is the tone, it does not change that.
What does happen is more 3D stereo imaging. the sound is no longer between the ears, but I get an out of head imaging.
Instruments seem to be better focused in a 3D environment - the sense depth gets more pronounced.
For example, the solo piano at the begining of Al Stewart: Year of the cat , was usually a little congested piano panned to far left.
With oversampling at 8X, it becomes an actual piano in a room with echoes of that room about couple meters away!
Otherwise resolution and tonal colour does not get altered.
I posted something on Neuron thread here in Head-fi but got no replies.
I wonder if anyone has had similar experience here, they'd be willing to discuss.

I don't have neutron player on android phone (using UAPP) but I've got JRiver on my PC. I remember there being an option to upsample, gonna try and see if I notice any changes.

This reminds me of my XD-05 days. Back then, I set up Foobar2000 to do a real time upsample to DSD cause to my ears it removed the digital sharpness from the sound and was more pleasing to listen to.

 

[Johnfg465vd]

I've done close to an hour of listening with 8x resample in Jriver using Sox resampler. Used mostly Sundara and sometimes Meze 99 Classic (Modded). Between the two (no resample vs 8x) there definately is a difference but it was very small and took multiple A B comparisions to notice. Still, could not say one was better than the other though. Maybe the way Neutron does signal processing effects the end result? I remember trying a bunch of android music players a few months ago, neutron (trial), UAPP, PowerAMP, Onkyo, Hiby.... Etc and while all other apps sounded the same, Neutron & PowerAMP had a slightly different sound. Take that info with a grain of salt, cause it's been a while since I did that test.

Out of curioucity, Since I've got slightly better gear than the ones I had before, I tried converting PCM to DSD on Jriver and used XD-05 again. There was a very noticeable but small difference. The digital glare/sharpness was softened a bit, so there definately is a change when resampling, but differences are mostly minor.

 

[Kentajalli]

I've done close to an hour of listening with 8x resample in Jriver using Sox resampler. Used mostly Sundara and sometimes Meze 99 Classic (Modded). Between the two (no resample vs 8x) there definately is a difference but it was very small and took multiple A B comparisions to notice. Still, could not say one was better than the other though. Maybe the way Neutron does signal processing effects the end result? I remember trying a bunch of android music players a few months ago, neutron (trial), UAPP, PowerAMP, Onkyo, Hiby.... Etc and while all other apps sounded the same, Neutron & PowerAMP had a slightly different sound. Take that info with a grain of salt, cause it's been a while since I did that test.

Out of curioucity, Since I've got slightly better gear than the ones I had before, I tried converting PCM to DSD on Jriver and used XD-05 again. There was a very noticeable but small difference. The digital glare/sharpness was softened a bit, so there definately is a change when resampling, but differences are mostly minor.

Thank you for your effort.
So did I.
Jriver, does a simple resample - it is not an interpolation upsample, what gives the game away, is the amount of CPU usage while resampling, hardly anything.
As such, Jriver resampling is only good if you happen to have an older DAC that say does not support anything above (say) 96kHz, so Jriver can stay compatible with it.
This resampling does not alter (improve) the sound, indeed it can damage it.
I believe the oversampling feature of Neutron is an interpolation upsample, but I can not find any info regarding algorithm used, noise shapers etc.
HQPlayer (PC Mac) can do a similar job (even better), with similar effects on sound quality.(CPU intensive depending on filter and noise shaper used)
So if you feel like it, download a demo version and try HQPlayer.
If your PC is beefy enough, HQPlayer can output in DSD, and improvements are very noticable (Imaging) , but it can gloss over small details on a track.

 

[GreenBow]

I think mojo is great, even it is a decade back gatget. It outperforms mobile phones, by far... It has a full body holographic presentation ( although a little narrowed soundstage, but not to much) that not many dacs can achieve. It has plenty of power and near to zero hiss for sensitive iems. Compared to ifi micro idsd , drangonfly cobalt and others it has the more natural sound. Perhaps there are better dacs nowadays, but it still holds its own way...

I never found the Mojo soundstage to be narrow. It had the same width as another DAC I was using at the time. However with the Mojo instruments at the side of the soundstage had just as much coherernce and prominence as the centre instruments. Whereas with my other (albeit budget) DAC, I found that I was focused only on the middle (maybe 50%) area of the soundstage.

It meant I heard and noticed instruments at the sides of soundstage with the Mojo, that could go unoticed without Mojo. Or rather maybe I registered the sounds on other DACs, but didn't take them in, so to speak.

Secondly, something that the Mojo has over other DACs, that is eveident with all Chord DACs, is depth. There is way more depth. Anything from twice as much and much more. That has the effect of making the soundstage sound narrower than it is.

Other DACs I compared with were Meridian Explorer, and sound from a DAB tuner, that could also run through Chord.


Soundstage depth is nothing gimmicky either. Hearing thedepth of a sound, is just as real as hearing either side of a sound. Off the shelf DAC chips, in cheaper implementations, don't have anything like the resolving power as Chord. Hence the image on those DACs comes across with less depth.

I adore the depth from Chord's DACs.

 

[Johnfg465vd]

@Kentajalli Hey, sorry I could not reply sooner, was busy with some work. I tried HQPlayer 4 and after setting it up, I get around 20% CPU (Intel i5, 44,100 Hz to 705,600 Hz) usage when playing back music (attached settings and filter used). I was afraid I'd need a more powerful CPU, but everything worked fine (maybe cause I checked CUDA Offload?). There is a definite sound quality improvement, albeit very very small. I did not notice any increase in soundstage or depth, but instruments moving around the soundstage were a bit more easy to follow and had more air. Also, the vocals seem nicer, I can't put my finger on what it is though, resolution, detail levels and tonality is the same but something seems different (transients change?). Needs more testing. Again. the improvements were very small, for me personally, they are not worth the horrible interface of HQPlayer or possible increased battery consumption on android with neutron's oversampling. Depending on your gear, your mileage may vary I guess.

I installed Neutron (Eval) from playstore and was able to try the App for ~25 Mins without problems, after that I started getting Evaluation period over message every time I tried to play something. With not enough time and in a noisy environment, I didn't notice any positive or negative changes in sound using neutron with or without 8x oversample.

 

[alekc]

@Kentajalli Hey, sorry I could not reply sooner, was busy with some work. I tried HQPlayer 4 and after setting it up, I get around 20% CPU (Intel i5, 44,100 Hz to 705,600 Hz) usage when playing back music (attached settings and filter used). I was afraid I'd need a more powerful CPU, but everything worked fine (maybe cause I checked CUDA Offload?). There is a definite sound quality improvement, albeit very very small. I did not notice any increase in soundstage or depth, but instruments moving around the soundstage were a bit more easy to follow and had more air. Also, the vocals seem nicer, I can't put my finger on what it is though, resolution, detail levels and tonality is the same but something seems different (transients change?). Needs more testing. Again. the improvements were very small, for me personally, they are not worth the horrible interface of HQPlayer or possible increased battery consumption on android with neutron's oversampling. Depending on your gear, your mileage may vary I guess.

I installed Neutron (Eval) from playstore and was able to try the App for ~25 Mins without problems, after that I started getting Evaluation period over message every time I tried to play something. With not enough time and in a noisy environment, I didn't notice any positive or negative changes in sound using neutron with or without 8x oversample.

From my experience mobile devices can't handle properly upsampling due to limited CPU power and quick battery drain so any player like Neutron makes little sense from my perspective. On the other hand when I'm in move I prefer dedicated transport like xDuoo X10T II which also limits additional software installation. Still considering tiny changes vs usability I prefer usability and longer enjoyment of a music on single charging.

Indeed HQPlayer has terrible interface, and it was one of the reasons why I've moved to Audirvana and cd transport with upsampling. Audirvana plays nicely with Mojo and while provides a fraction of oversampling option comparing to HQP it is less CPU hungry (at least on x64).

 

[Kentajalli]

@Kentajalli Hey, sorry I could not reply sooner, was busy with some work. I tried HQPlayer 4 and after setting it up, I get around 20% CPU (Intel i5, 44,100 Hz to 705,600 Hz) usage when playing back music (attached settings and filter used). I was afraid I'd need a more powerful CPU, but everything worked fine (maybe cause I checked CUDA Offload?). There is a definite sound quality improvement, albeit very very small. I did not notice any increase in soundstage or depth, but instruments moving around the soundstage were a bit more easy to follow and had more air. Also, the vocals seem nicer, I can't put my finger on what it is though, resolution, detail levels and tonality is the same but something seems different (transients change?). Needs more testing. Again. the improvements were very small, for me personally, they are not worth the horrible interface of HQPlayer or possible increased battery consumption on android with neutron's oversampling. Depending on your gear, your mileage may vary I guess.

I installed Neutron (Eval) from playstore and was able to try the App for ~25 Mins without problems, after that I started getting Evaluation period over message every time I tried to play something. With not enough time and in a noisy environment, I didn't notice any positive or negative changes in sound using neutron with or without 8x oversample.

Thank you.
The point of this exersise was not to praise various software upscalers, but to investigate what effect, improvements (if any) they produce when connected to a Chord Mojo.
After all, Mojo also upscales, so how could a humble $10 Android software improve the sound?
To me it certainly does, specially at 16X (found a way to keep it stable) - this means upscalling a mere 44.1kHz all the way to 705.6kHz.
I was wondering if it can be replicated by others, and could they notice what I get.
At least with HQPlayer you could replicate some of my findings. BTW , to my ears, poly-sinc-mqa-mp with no dither or LSN15 , does the trick.
Upscaling to DSD slightly improves the sound but requires more CPU.
Neutron does not use that much extra battery processing to 16X. Last night I listened to about 3 hours, my battery was down 15% as a result - perhaps 5% extra.
The more I try to find a way to convince myself, that it isn't so - the more I hear that it is so, while the book says it shouldn't!
If nothing else, at least I found a way to improve the sound of the mighty Mojo, for free.

Regarding the very short eval period on Neutron before nagging starts, I recommend leaving a feedback for them, perhaps they listen.
25 mins is a JOKE - a week is reasonable.

 

[Rob Watts]

My strong recommendation is never upsample- the WTA filter built into Mojo will do a much better job of recovering the timing of transients of the original analogue than any other (including ultra long tap length filters) and you will be rewarded with better placement accuracy, timbre variation, instrument separation and focus. Much more importantly, better musicality (defined as getting emotional with the music). Of course, if you like soft bass, flat soundstage with artificial width, and the loudest instrument to dominate with a hardend edge, then go for software upsampling...

 

[Kentajalli]

My strong recommendation is never upsample- the WTA filter built into Mojo will do a much better job of recovering the timing of transients of the original analogue than any other (including ultra long tap length filters) and you will be rewarded with better placement accuracy, timbre variation, instrument separation and focus. Much more importantly, better musicality (defined as getting emotional with the music). Of course, if you like soft bass, flat soundstage with artificial width, and the loudest instrument to dominate with a hardend edge, then go for software upsampling...

Thank you.
This is the book I was referring to !
I am aware of your assertions and I do believe you and indeed I have recommended to others on these very pages, the same.
But yet somehow ....

 

[alekc]

My strong recommendation is never upsample- the WTA filter built into Mojo will do a much better job of recovering the timing of transients of the original analogue than any other (including ultra long tap length filters) and you will be rewarded with better placement accuracy, timbre variation, instrument separation and focus. Much more importantly, better musicality (defined as getting emotional with the music). Of course, if you like soft bass, flat soundstage with artificial width, and the loudest instrument to dominate with a hardend edge, then go for software upsampling...

Dear @Rob Watts thank you for explanation. Now I know that my subjective feelings are in fact objective and comes from Mojo design.

 

[Johnfg465vd]

Regarding the very short eval period on Neutron before nagging starts, I recommend leaving a feedback for them, perhaps they listen.
25 mins is a JOKE - a week is reasonable.

If I remember right, the Eval period is 30 days, My trial period ended way back when I tested all the different Music Players. It just took ~25 mins for the App to realize my trial period was over.

The more I try to find a way to convince myself, that it isn't so - the more I hear that it is so, while the book says it shouldn't!

At the end of the day, if you like the changes that extra processing, or mods... etc bring and are able to enjoy music, that's all that matters.


@alekc Hey, could you share your opinions on the X10T II as a transport for Mojo. For a while now, I've been mulling over getting the X10T II. While I would love to get a Ploy, it costs as much as a new Mojo, maybe some day I'll get it but for now xDuoo is cheaper and available locally (no shipping + customs). I remember someone mentioned that sound quality improved when using xDuoo with Mojo, something to do with less noise or interference... Have you noticed any changes in sound quality between Mojo paired with X10T II and your phone? Have you tried the Bluetooth to USB/SPDIF option, I'm thinking it's a very convenient way to get wireless connectivity when using YouTube or watching movies etc. and not seriously listening to music via Hiby Link or local storage.

 

[Johnfg465vd]

My strong recommendation is never upsample- the WTA filter built into Mojo will do a much better job of recovering the timing of transients of the original analogue than any other (including ultra long tap length filters) and you will be rewarded with better placement accuracy, timbre variation, instrument separation and focus. Much more importantly, better musicality (defined as getting emotional with the music). Of course, if you like soft bass, flat soundstage with artificial width, and the loudest instrument to dominate with a hardend edge, then go for software upsampling...

Thanks for your explanation, makes sense that Mojo would do a better Job of reconstructing the original signal. I'm curious about something, and sorry if this has been discussed before. Would using EQ effect the transients timing or how mojo reconstructs the source signal? I never had good results when using EQ and I'm curious if it's because of my lack of experience using an Equalizer or If there is more to it.