Some off-topic posts have been removed.To be clear: This thread was split to keep discussion of subjective testing separate.
Suggested thread for such discussion - If it's purely an engineering challenge why bother designing yet another DAC?
Infractions have been given regarding this
Last edited:
Hi Terry,
Thanks for that.
Oh for sure I would expect the Mola Mola to do everything to the utmost degree. For the kind of money it commands, 12k USD, it should. I have no idea what the Mola Mola would cost in Australia. Marantz is pricey enough. The Ruby listed at 8450 AUD. There was an authorized sale earlier this year dropping the price to 5k AUD. I haven't spent that kind of money since 2000 and well before I got in to DIY.
The new 30n player lists at 2500 USD and 6k AUD, crazy difference.
I started following this thread, and other dac development threads, well before I gave any serious thought to the Ruby player. And then it all just kind of happened real fast. Still like following the threads though.
What i believe is more relevant is that i have listened to Mola Mola amps in various systems and also, on more that a few occasions, to a complete Grimm Audio setup. While i would not compare the experience to a dental appointment, the overall result is far removed from my idea of musical enjoyment. If music listening was my daily job these pieces would work just fine, but i wouldn't use them if not getting paid for it.
Sorry, I have to ask those complaining about off topic comments in this thread: and the last couple of pages pontification on various expensive boutique equipment, to the AK4499 DAC topic is...? Perhaps it would be appropriate to open a new thread on high end DAC equipment?
A little birdie told me that when AKM developed AK4499, they expected it would be received by the high end Japanese audio companies such as Esoteric in the same way that the new Rohm dac chip has been received by Luxman. But it didn't happen that way. Most of those companies decided to stop making chip dacs and pivot to aiming for FPGA dacs instead.
Could AK4499 be used in a high end design? I think so. It most likely would not be something in the flavor of Mola Mola though. AKM gave a small hint they were thinking in that direction that can be seen in the design of AK4499 eval board. I think I even wrote in some forum posts that it could sound very good in some cases, but distortion was higher if using it. Nobody paid much attention. Some assumed AKM was just being stupid.
Could AK4499 be used in a high end design? I think so. It most likely would not be something in the flavor of Mola Mola though. AKM gave a small hint they were thinking in that direction that can be seen in the design of AK4499 eval board. I think I even wrote in some forum posts that it could sound very good in some cases, but distortion was higher if using it. Nobody paid much attention. Some assumed AKM was just being stupid.
Last edited:
Does the Mola Mola have a flavour, I thought it was supposed to be transparent? Does that mean "best ever" and "high end" signify an effects box?
Are you genuinely surprised people have different interpretations of high-end, Scott?
What is a "FPGA dac" and why would it be bad (or not good enough)? Can you give some examples of commercial "FPGA dacs"?
There are a number Field Programmable Gate Array (FPGA) products, the Chord Company's products as an example.
The FPGA is comprised of millions or billions of programmable 1-bit gates that can be programmed to do many different types of things. In this case, each of these 1-bit gates is being programmed to decode a specific type of modulation (pulse-width modulation), outputting a voltage.
There's a lot more to it then that, but you are essentially using a very simple computer process to generate an output as a voltage source.
Cheers,
Greg
The FPGA is comprised of millions or billions of programmable 1-bit gates that can be programmed to do many different types of things. In this case, each of these 1-bit gates is being programmed to decode a specific type of modulation (pulse-width modulation), outputting a voltage.
There's a lot more to it then that, but you are essentially using a very simple computer process to generate an output as a voltage source.
Cheers,
Greg
I would say that the implementation is far more important than whether it's a purpose-designed device (in this particular discussion) or generic device.
I have a sabre dac that sounds very, very smooth for a reason: the various regulated supplies that I've used make a really pronounced difference. Measurable, not really.
This discussion is always hitting the far left or right, such as "Delta Sigma dac chips don't sound as natural as R2R or FPGA." Well, that has not been my experience at all.
I built two different dacs: one based on a TDA1541 and the other on an ES9018. They are both very smooth and very enjoyable. A huge effort was put into the implementation of each of these chips and the results are very similar, yet there are some differences.
Again, I just cringe at these discussions when there is much, much more to how an audio device "sounds" or audio circuit "sounds" without looking at the entire build / product.
Cheers,
Greg
I have a sabre dac that sounds very, very smooth for a reason: the various regulated supplies that I've used make a really pronounced difference. Measurable, not really.
This discussion is always hitting the far left or right, such as "Delta Sigma dac chips don't sound as natural as R2R or FPGA." Well, that has not been my experience at all.
I built two different dacs: one based on a TDA1541 and the other on an ES9018. They are both very smooth and very enjoyable. A huge effort was put into the implementation of each of these chips and the results are very similar, yet there are some differences.
Again, I just cringe at these discussions when there is much, much more to how an audio device "sounds" or audio circuit "sounds" without looking at the entire build / product.
Cheers,
Greg
Greg: An FPGA just means they didn't have the R&D budget for custom silicon. Many commerical DACs are likely to have been breadboarded on FPGA during development. In the case of Chord they like unfeasibly large FIR filters so they were forced down a custom route. The upsampling and filtering could be implemented in many technologies.
Do filters make more difference to perception that the end coversion technology? Possibly. Do people make a big fuss about a minimal or non-existant audible difference? Certainly, but that is high end audio.
Do filters make more difference to perception that the end coversion technology? Possibly. Do people make a big fuss about a minimal or non-existant audible difference? Certainly, but that is high end audio.
Trust me, to go from a general purpose FPGA to ASIC can be eye wateringly expensive. That's why FPGAs sell for hundreds of dollars. If you make 200 DACs a year no point. If you make 2 million a year then the cost of conversion becomes attractive. The Spartan 6 in the Chord Dave is midrange at about $100 in one offs. That's a lot of BOM cost in a audio product.
That’s an insurmountable inevitability. People like sharing their feelings. It’s not something that can be contained, even when being discussed on the lofty shores of a diy audio community.
For example AKM demoed a new noise cancelling chip to me a few years ago AS an FPGA emulation. It allowed us to tune the ANC of a headphone in about 15 minutes from scratch. Unfortunately the chip doesn't yet exist. probably delayed again with the fire. Also, no way you can use an FPGA for a power sensitive or space sensitive application like a headphone. Still the mask sets cost alone mean sell 1 million chips or bust.
I can't imagine what Chord is doing in the FPGA that could not be done in a Shark or Blackfin processor.