APL 3910 owners, has anyone tried APL upsampler?

Kaaos, Your right, you will be AMAZED at the improvement over the AA. I am shipping mine back to Alex next week, uuggghhh, I hate to part with it, but I will let you know after I get it back. I have found out over time, I just do what Alex suggests, as far as digital goes. Now, if the guy just got his amps, and speakers going, maybe I could stop searching....remember, the APL really comes into it's own after 500 uugghhh hours.

Steve

I thought i remembered reading comments from Alex saying something to the effect that "upsampling does nothing" and is "strictly marketing", etc... Am i confusing him with someone else??? Sean
>

Based on a search of the archives, Alex is not a fan of non-linear conversion, often refered to as 'upsampling' ie 44.1 khz to 96 khz etc. Referring to APLs website comments he is technically ony performing linear coversion, often referred to as 'oversampling' ie 44.1 to 88.2 etc. The difference between the two methods is that non linear conversion in a binary based system like digital must by definition involve truncation of bits at some point in the conversion. Linear conversion does not involve any truncation of the original signal.

This makes sense to my brain and confirms what my ears have known for much longer. I have never liked any implementation of non-linear conversion. Players and digital remasters that upsample the redbook standard to 24/96 have rarely sounded right to my ears.

In general terms Alex's oversampling implementation doubles the sampling rate and the takes the bit rate to 24 bits.

Hope that helps

Kerry

Sean, you're probably confusing him with me. My only points on upsampling is that the name "upsampling" is strictly a marketing term, not a technical one. From a technical standpoint it is simply "oversampling". Secondly, oversampling/upsampling by non-integer values to 96kHz or 192kHz results in a slight mathematical imprecision compared to oversampling/upsampling to 88.2kHz or 176.4kHz or other integer multiples of 44.1kHz.

BTW, Alex has chosen a 88.2kHz "upsampling"/oversampling filter for sonic reasons (2x44.1kHz). With this chip he is able to use 88.2/96/176.4/192/211kHz oversampling, but feels 88.2kHz sounds best. I'm not sure I'd be able to tell the difference...I'll trust Alex on this one though.

I guess my biggest gripe with the "upsampling" marketing that has overwhelmed the digital field for the last 4 years is that it was marketed as something new and implied (by virtue of 96/192kHz) that it in some way was related to, or equivalent to, the new DVD-A standard, but for redbook CD's. Maybe it was necessary to give people a reason to accept the superiority of a better designed digital filter? But dCS's "upsampling" chip is not the same as the cheap ones that followed. Unfortunately, way too many of us "buy in" without listening at times. Not all of us certainly, but quite a few. I couldn't even count how many threads I've read that contained sentiment like, "yeah, but it's not an upsampling player, so I'm going with brand x". For others, their ears, or the ears of trusted others, have convinced them that a particular player is to their liking, and if it's an "upsampling" player, this may give them a certain confidence that there are good technical reasons for its superiority. No problem with that, but it's not really accurate.

"Upsampling"/oversampling is not the source of the sonic superiority of some chips. This is simply digital multiplication and is about as basic as it gets. The real magic/technical superiority lies in what happens after the additional bogus sampling data is added. This is where the math and technical complexity is expressed in the chip/digital filter. The merits of any given digital filter are in how well it implements noise shifting, anti-aliasing, etc and ultimately how well it is able to facilitate the reconstruction of the original analog waveform. This is where high mathematics comes into play and IMHO what differentiates the various chips.

I'm not "against" oversampling/upsampling to 96kHz or 192kHz...just think those numbers are chosen for marketing reasons, even though there is a small loss of mathematical precision in doing so. Whether this is audible, I have no idea? I owned and loved the Audio Aero Capitole and it oversamples/upsamples to 192kHz. I think the Swiss Anagram chip that Audio Aero uses is a fantastic and technically sophisticated chip!! I also think their analog section is done well. Unfortunately, the user interface/remote was ridiculously bad, IMO...heck $50 DVD players did this much better. Damn French Company...did so many things so right and then dropped the ball on the easiest part! LOL

Does my "stance" make sense? Preaching to the Choir? Babbling like a Baffoon?

BTW, I just received my APL 3910 from Alex today. It does not have the "Upsampling" chip. First impressions right out of the box without even a warmup were positive, but certainly not earth-shattering. I still have some system setup issues to address and certainly some burn-in time should help. I'm optimistic though.

So far I have just two people who listened to the Upsampler in A-B test with the original 44.1/16 version while playing CD. For all of those who decided to go with the Upsampler but are not interested in the Digital Input receiver, I will be able to use the second PCM port of my DAC so you can alternate between the origianl and upsampled version and hear for your self. Until now, most companies were using Analog Devices AD1895/96. This chip is the one built into the DAC modules in Audio Aero too. The AD chip has a control port. The designer has just two choices - 96/24 or 192/24. The chip I am using is the latest deveolpment in the industry and NOT used by any other company on the market today. At least not for now. It is a product of many years research and is absolutely uniquie. It can perform any type of conversion up to 211/24. I am not going to tell you why, but the sonic improvement with this chip is not only because it upsamples. I find that 88.2/24 sounds best with it. Please keep in mind that the CS4398 DAC chips as used in my DAC board, unlike most 8x oversampling DACs on the market, have non-oversamplng digital filter for PCM. Then the PCM is converted to dual bitstream and then is overampled 32, 64 or 128 (user selectable). In the case with 88.2/24 oversampled PCM you get final dual bitstream rate equal to 128 oversampled DSD inside the DAC.

Regards,
Alex