[marcmccalmont]

I am not quite sure what you mean by "would a 192khz signal accelerate this process with regards to how the mscaler operates here as compared to standard redbook files?".

With Blu2 I have tried some 192 MHz files which were recorded as 192 MHz files as opposed to being upsampled. I compare these through Blu2 compared to the 44.1 MHz versions of the same recording also fed through Blu2. For whatever reason, I preferred the 192 files.

I might have miss interpreted the post if the original file is 192 KHz sampling rate I prefer that too but the difference is not as much as I had thought it would be

 

[musickid]

From my understanding a 192kHz file at 24 bit contains more information about the original recording than an identical recording sampled at 44.1kHz. So i was wondering whether or not blu2 post hardware upsampling to 705/768kHz would reproduce exactly the same recording with exactly the same sound and sonic qualities regardless of what resolution the audio is that it is "fed". With interpolation at 16 bits and higher the initial sample rate should not matter if i'm right??If this is the case does that mean a blu2 owner no longer needs to invest in high-res music/downloads as redbook files result in exactly the same end result. For me the implications here seem staggering if i have understood this correctly using my simplistic analogy.

 

[rayl]

Am i right in saying with an mscaler the sample rate resolution of the incoming digital audio signal no longer matters? Whether 44.1 or 192khz everything will be 'hardware' upsampled to 705/768khz. The WTA1 stage of blu2 will interpolate to 16 bit accuracy and will substitute the "missing" parts regardless? Or would a 192khz signal accelerate this process with regards to how the mscaler operates here as compared to standard redbook files?

That is the belief and aspiration. One of the goals of the davina ADC project is to create a high res reference so side by side listening can confirm or refute the statement. Maybe we might need a longer filter- or maybe not.

 

[rayl]

From my understanding a 192kHz file at 24 bit contains more information about the original recording than an identical recording sampled at 44.1kHz. So i was wondering whether or not blu2 post hardware upsampling to 705/768kHz would reproduce exactly the same recording with exactly the same sound and sonic qualities regardless of what resolution the audio is that it is "fed". With interpolation at 16 bits and higher the initial sample rate should not matter if i'm right??If this is the case does that mean a blu2 owner no longer needs to invest in high-res music/downloads as redbook files result in exactly the same end result. For me the implications here seem staggering if i have understood this correctly using my simplistic analogy.

Exact is a tough word.

Here is what is known. If you had perfect ADC, and the input signal is truly bandwidth limited to half the sampling rate (22.05), and you had an infinite reconstruction filter, and you have ‘perfect’ math— then 705.6, or even odd ball numbers like 6666 can all be calculated from 44.1...

But we stick to exact multiples bec we want to get closer to perfect math without blowing up the require computational load... rounding errors are lessened with integer multiples.

There is also the 16 vs 24 but depth question. (Which is intertwined with perfect math ... bec the more bit depth you want on output, the more but depth you need in intermediary calculations and in any case to get to ‘perfect’, that is, enough bit depth so rounding only impacts precision beyond your desired output bit depth).

We have only 1MM taps today, not an infinite. We will get to infinite in an infinite number of days (or years if you want- same size infinity!).

My own hypothesis is that 44.1 may result in some aliasing, or we need to move the filters down more than we want to achieve 22.05 bandwidth, but I can’t see needin more than 88.2.

 

[Jawed]

A 16-bit 44.1KHz version of a 24-bit 176.4KHz recording introduces two errors:

1. decimation - reducing the sample rate, even with the same bit depth, always reduces the bandwidth which introduces timing errors (noise adds uncertainty)
   
2. reduced bit-depth - will introduce timing errors

A 44.1KHz version of a 192KHz recording introduces another error:

• non-integer ratio of sample rates - adds errors like 1 and 2 above at the same time, making them worse again

If you have to A-B to hear the difference, then it's probably not worth the bother buying the "high res original".

The real benefit of high resolution recordings is the reduced aliasing. Frequencies above Nyquist present during sampling will be vastly less than for recordings made natively at 44.1KHz. Most of the history of digital recording has ignored Nyquist during analogue to digital conversion, resulting in recordings (or transfers from analogue) that are fundamentally broken.

No matter how good Rob makes the upsampling, the aliasing errors in native 44.1KHz recordings will always be there. Indeed Rob's better designs will make the aliasing errors easier to hear. The aliasing errors in 176.4KHz or 192KHz recordings will also always be there, but they will be tiny compared with 44.1KHz-native recordings.

It's hard to say which of these errors is the most important, judging by listening tests, because you need the right recordings (files) to start with. Also you need the right DAC. And BluDAVE may not be good enough to show the errors completely.

 

[spotforscott]

Does anybody use their Blu2 with DVD concerts through TV? Is there any way of dealing with lip sync issue?

 

[AndrewOld]

Am i right in saying with an mscaler the sample rate resolution of the incoming digital audio signal no longer matters? Whether 44.1 or 192khz everything will be 'hardware' upsampled to 705/768khz. The WTA1 stage of blu2 will interpolate to 16 bit accuracy and will substitute the "missing" parts regardless? Or would a 192khz signal accelerate this process with regards to how the mscaler operates here as compared to standard redbook files?

I haven’t got an Mscaler, but I would be surprised and even concerned if a genuine 192kHz recording sounded no better than the same recording downsampled to 44.1kHz both played through the M Scaler. There is more information in the original recording. You should hear more, or certainly no less. But what do I know.

 

[musickid]

I wasn't thinking downsampled to 44.1 more the original recording mastered to 192 and 44.1.

 

[dmance]

Maybe off-topic, but at the Montreal Audio Show this weekend I listened to the BluDave for a long stretch and had them switch the mid/high M-Scaler upsampling switch on Blu2 and found I preferred the mid (500k taps). Less detail, yes, but also less screechy ...at least to my ears although other people just nodded in approval when 'told' which was supposed to sound better. Just my guess, but this was probably an issue with the increased RF on the dual coax caused by the FPGA working harder for the more taps. Do any bludave owners notice this?
Thanks

 

[musickid]

When i listened to daftpunk through bludave through oppo pm1 headphones it was crystal clear with hyper detail. Dave on its own was much warmer IMHO. Dave also sounded quieter at the same volume which has been explained before. I think bludave cannot be appreciated or understood in a quick a/b.

 

[Triode User]

When i listened to daftpunk through bludave through oppo pm1 headphones it was crystal clear with hyper detail. Dave on its own was much warmer IMHO. Dave also sounded quieter at the same volume which has been explained before. I think bludave cannot be appreciated or understood in a quick a/b.

My experience is the complete opposite. I find BluDave warmer and a fuller sound with bass more defined and obvious compared to solo Dave. BluDave volume is 3dB less than Dave on its own.

 

[AndrewOld]

I wasn't thinking downsampled to 44.1 more the original recording mastered to 192 and 44.1.

How else do you get an original 192 recording down to 44.1 without downsampling? That is what downsampling is!

 

[musickid]

an original 44.1kHz recording unless i'm mistaken.

 

[marcmccalmont]

Maybe off-topic, but at the Montreal Audio Show this weekend I listened to the BluDave for a long stretch and had them switch the mid/high M-Scaler upsampling switch on Blu2 and found I preferred the mid (500k taps). Less detail, yes, but also less screechy ...at least to my ears although other people just nodded in approval when 'told' which was supposed to sound better. Just my guess, but this was probably an issue with the increased RF on the dual coax caused by the FPGA working harder for the more taps. Do any bludave owners notice this?
Thanks

No but I have 32 ferrites on my coax!

 

[Triode User]

Maybe off-topic, but at the Montreal Audio Show this weekend I listened to the BluDave for a long stretch and had them switch the mid/high M-Scaler upsampling switch on Blu2 and found I preferred the mid (500k taps). Less detail, yes, but also less screechy ...at least to my ears although other people just nodded in approval when 'told' which was supposed to sound better. Just my guess, but this was probably an issue with the increased RF on the dual coax caused by the FPGA working harder for the more taps. Do any bludave owners notice this?
Thanks

Basically, no.

Something was not right elsewhere I’m guessing. BluDave is the opposite of screechy even with cheap bnc and no ferrites (which are, after all only fine tuning). When I first got my Blu2 I switched to the mid position on the upscaling and a friend in the room said that was the best demonstration of just how good the full MScaling is.

Don’t get this RF thing on the BNC out of proportion. Rob Watts himself has said it is a side show compared to the huge improvement due to the MScaler and is more a matter of very fine tuning when taken in the context of the overall sound of Blu2.

To prefer the mid position is bizarre and to my mind means something odd was going on.