Upsampling - Server vs DAC

[Gieseler Audio]

I have been playing with upsampling lately & on my DAC's the general consensus is it definnately sounds better, especially on the AKM Klein. Using my new Tecktronix scope it looks a lot better too. I use very gentle post DAC filter so higher Rez material works much better with this type of filter.

I would like to get some feedback from you guys as to which sounds better, upsampling in software/sever or hardware /DAC? There is a nice new AKM upsampler chip that I was thinking of having a play with.

Any experience/feedback appreciated.

Cheers

Clay

[acg]

Definitely software is my vote Clay.

[ljmac]

Software is definitely the most flexible, but also the most complex for the end user. It should also be cheaper in theory, but I strongly suspect firmware has greater SQ potential, in that it can be more effectively integrated with the DAC for lower jitter etc.

 

On my MDAC I definitely prefer the SABRE DAC's built-in filters over the MDAC's DSP filters, and in my experience with software so far I prefer to run it straight and let the DAC do the work. However, with something like the Phasure where the DAC doesn't do any of its own filtering, I suspect software may work very well.

 

I guess the best advice I can offer based on my experience is do one or the other, but not both.

[bhobba]

Hi Guys

 

I personally prefer it in hardware because recloking reduces jiter.

 

Thanks

Bill

[ljmac]

On further consideration, given the increasing popularity of dedicated hardware music servers and the theoretical SQ advantages, I think you'd better do everything via hardware.

[Primare Knob]

It differs per DAC design. It is trying to find out at which data stream the DAC operates in optimal performance.

http://www.stereo.net.au/forums/index.php?/topic/78727-high-res-audio-and-what-is-left-of-it/

[davewantsmoore]

I would like to get some feedback from you guys as to which sounds better, upsampling in software/sever or hardware /DAC? 

 

IMHO it's less about where we do the upsampling ... and more about whether it is done in "realtime" or not.     Doing the oversampling seperately from the playback infers a big advantage that we can use infinitely complex filters, becuase it doesn't matter how long they take  (or whether the time taken per sample is even or not) ..... it imparts much more flexibility.

 

 

Cons of up-sampling in realtime means:

• Practical limitations on the filters which can be employed  (because the computation must happen within a certain timeframe)
• Electrical noise, or electrical complexity of the devices directly involved in the playback/DA process may increase

 

Cons of up-sampling before hand (non-realtime)

• More data needs to be transferred from the source to the DAC
• Other parts of the DAC (or playback chain, in addition to just after the oversampler) may need to operate faster  (ie.  the transport, and the transport -> DAC interface)

[davewantsmoore]

A DAC like the ESS Sabre has a sophisticated/fast internal ASRC oversampler/reclocker ..... it is interesting to read what it's lead designer and other high end designers say (it works very well) about turn off the DAC chips internal oversampling.... and instead feeding the chip with a fast external clock, and high sample rate material    (eg.  oversampled before input, eg.   44.1 -> 352.8kss-1)

[Gieseler Audio]

Thanks guys for all the info. I tend to favour software up sampling in the server as it is simple, Also i just noticed that I can Upsample in Daphile too which is great as I use it at work.

On the the other hand there is this brand new Premium Audio AKM SRC coming out shortly which looks amazing.

http://www.akm.com/akm/en/product/detail/0056/

It also has a total bypass mode so I'm guessing you could switch it in or out, great for comparisons. Thinking of prototyping up a DAC using this SRC up front just for fun. probably combine it with the new AKM Verita 4490EQ DAC which I have just designed a PCB for.

What do you think?

Edited February 27, 2015 by Gieseler Audio

[matth]

Thinking of prototyping up a DAC using this SRC up front just for fun. probably combine it with the new AKM Verita 4490EQ DAC which I have just designed a PCB for.

What do you think?

 

Hi Clay,

the idea to combine the AK4137 SRC with the AK4490 in an upcoming Klein DSD DAC is in my view most appealing. The AK4137 is able to convert all PCM and DSD content to DSD256 and the AK4490 in your DSD DAC will do an excellent job to convert DSD256 to analogue. To keep costs at a minimum I would like to propose to get rid of both SPDIF inputs and design a pure USB DAC. That would be the perfect DAC for me.

 

KR Matt

Edited August 1, 2015 by matth

[Gieseler Audio]

Hi Matt,

Actually I looked at the new AKM SRC a while ago & thought, wow that is an amazing device! Good idea as well to incorporate it into a DAC with just USB input, no I2S switching so the USB board I2S lines would be more direct. You have got me thinking now. I might have a go at prototyping up a PCB & give it a go. I only ever use USB myself myself so a DAC like that would definately take less time & parts to build & probably sound better. Maybe include galvanic isolation on the USB I2S outputs & some nice reclocking as well. OK leave it with me & see what I can come up with, that is when I get time in between filling Klein DAC orders.

[Gieseler Audio]

Update,

No stock at Digikey yet on the AK4137, looks like Q2 2015 before production starts so must be close now.

I will get on to AKM direct & see if I can get some samples to get the prototype happening.

Edited August 1, 2015 by Gieseler Audio

[Martykt]

Hi Clay,

I can't speak for software upsampling as having no experience with it (though do have experience with 64 bit DAW software but that's a story for another day) but I do have experience with hardware upsampling specifically through an Esoteric P03 Transport while I don't know for sure whether Esoteric uses a AKM upsampling chip to do their upsampling I do know Esoteric uses AKM chips in their dacs so is a reasonable assumption that they may use AKM upsampling chips as well.

I can say confidently the hardware upsampling in my transport works extremely well and makes a big difference.

It's great to know there's guys out there like you trying different things, keep up the great work

[Addicted to music]

Just remember that not all of us use USB.....

[matth]

Hi Matt,

Actually I looked at the new AKM SRC a while ago & thought, wow that is an amazing device! Good idea as well to incorporate it into a DAC with just USB input, no I2S switching so the USB board I2S lines would be more direct. You have got me thinking now. I might have a go at prototyping up a PCB & give it a go. I only ever use USB myself myself so a DAC like that would definately take less time & parts to build & probably sound better. Maybe include galvanic isolation on the USB I2S outputs & some nice reclocking as well. OK leave it with me & see what I can come up with, that is when I get time in between filling Klein DAC orders.

Thank you Clay for thinking about,

IMO there are three benefits:

You get rid of both the input switch, which somehow degrades the sound quality and the WM8804 with all the SPDIF processing. Further, you are not depending on a software player for converting the files. If an external software converter is better than the AK4137 it is possible to bypass the AK4137 anyway.

KR Matt

Edited August 1, 2015 by matth

[matth]

Update,

No stock at Digikey yet on the AK4137, looks like Q2 2015 before production starts so must be close now.

I will get on to AKM direct & see if I can get some samples to get the prototype happening.

Clay,

good luck and all the best, the chance is great that it will be the worlds first DAC with both AK4137 and AK4490 and most certainly without any competition at its price point and above.

KR Matt

Edited August 2, 2015 by matth