Consensus about upsampling to 512 DSD

[Audiophile Neuroscience]

11 minutes ago, beerandmusic said:

Hopefully we can quit talking about the word consensus and get to the subject.

 

Isn't Consensus the subject ?

 

11 minutes ago, beerandmusic said:

PS - there is a majority consensus among those that have software upsampled to DSD512 to a NOS DAC with proper resources that it is preferred vs not upsampling to a NOS DAC....

 

How do you know that?

[beerandmusic]

23 minutes ago, Audiophile Neuroscience said:

 

Isn't Consensus the subject ?

 

no, the consensus was a statement of fact (i believe it is an unspoken fact-grin)....the proposed topic of discussion was software that upsamples to DSD512....read post one.

[beerandmusic]

 

PS - there is a majority consensus among those that have software upsampled to DSD512 to a NOS DAC with proper resources that it is preferred vs not upsampling to a NOS DAC...

 

27 minutes ago, Audiophile Neuroscience said:

How do you know that?

 

many reasons, but mainly that i believe those that have tried are not stupid or deaf (wink)

[Summit]

On 6/30/2018 at 4:40 PM, firedog said:

You are free to trust you own perceptions as much as you want- no one can argue with what you hear. 

What is disputable is when you say your perceptions “prove” something to anyone other than you; or when you do a sighted comparison and don’t accept that you may have subconscious biases that affect what you hear. 

The classic example is wine tasting, where it has been shown that even experts judge wines totally differently when the tasting isn’t sighted. When it is sighted, the more expensive\more well regarded wines pretty much always win.

Some people would rather do a non sighted comparison, as it may prevent them (among other things) from spending money on an item that doesn’t really sound better than a less expensive item or an item they already own.

 

I have not stated that perceptions “prove” anything. Measurements are data that correlate with some SQ aspects, but besides transducers they often tell me very little about how something actually sound and if I would like it or not. We should strive for accuracy no doubt, but the question is which measurements is important and how was they obtained. The classic way to get good measurements is to filter out noise or to use global feedback. Both methods can be done to sound good if made right. Both can also be made so that they soften and dulling the sound. To listen to music we need an audio system and how the whole system sound in the room is what’s important. Measuring one gear won’t tell anything about synergy between different gear and under different loads etc.       

 

I’m certain bias is real and something we should try to minimize. I believe in non-sighted comparison as the best way to eliminate bias and have said it many times before. Blind test on audio gear is done by listening with our ears and not to read measurements. Measurements is one thing that can bias people to think that what they read is what they hear.

 

I believe that if I can’t hear the difference between two gear in the same audio system there is no point to change what I have, even if the other one may have 0,001 % better THD or whatever. I use a pragmatic approach using my ears then evaluating audio gear but are not denying bias. The design and measurements can make me interested in listening to some gear, but not for evaluations which I like and what to buy.    

[Summit]

If you ask people on a Muslim forum if they consider the Quran, their holy book, to be the verbatim word of God you will get much more that says yes than if the same question is asked on a Christian forum. 

 

If you ask people on a pro Trump forum if building a new wall to Mexico is a good idea you will get much more that says yes than if the same question is asked on a pro democrat forum. 

 

If you ask people on a Ferrari forum if Italian sports cars are better than American you will probably get much more that says yes than if the same question is asked on a forum about Ford Mustang. 

 

What I try to illustrate and explain is that it’s important to get a representative survey sample to assure that inferences and conclusions can reasonably extend from the sample to the population/audiophiles as a whole.

 

One of the things that (I believe) have made CA to such a big audio forum is the tweaking with upsampling on software like HQplayer and XXHighEnd. Because it’s not only the size of the responders that is of importance, but also to get an adequate sample of the population for the question at hand. Upsampling and the ability to choose different software is one of the biggest pros of using a computer compared to spinning discs or are streaming from a network player.

 

If doing a survey on a forum like CA, with many members that is in favour of using a computer to play music, one should calculate that they probably are much more inclined to upsample the signal before sending it to the DAC than other audiophiles that are spinning discs or are streaming from a network player.

 

BAM If upsampling to DSD 512 sound good to you and you have the gear to play it on, go for it. Most companys have so that you can try their software before buying.

[mansr]

7 hours ago, tailspn said:

Actually no, the structure of DSD bit and framing organization has nothing to do with which software is used to operate on it. That organization is in the original DSD Scarletbook, and later expanded upon with faster DSD bitrates.

 

Clicks CAN be minimized by selecting the point in time of zero crossing. I used to do that years ago at the cost of several hours per album. The problem is then rendering out the sliced tracks. That rendering process discards the bits within a incomplete frame where the slice was made. If you can suggest (prove) an application that allows arbitrary slicing at bit points within a frame, which then separates out the resulting tracks without discarding bits back to the frame foot, I'd be most interested 

 

I can only speak from my experience using Pyramix, which is the professional DAW of at least 80% plus of production studios world wide producing DSD content.

OK, I don't know the SACD requirements, and I don't have experience with Pyramix. I was talking about the file formats used for downloads. DSF is by far the most popular, and it supports any number of samples exactly. DSDIFF does indeed require a multiple of 8 samples (sorry, I misremembered). As you correctly point out, DSD data itself has no frame structure, so there is no inherent reason to restrict the cut points. SACD, being based on the DVD physical medium, may well require sector alignment. I know DVD does. What size "frames" does Pyramix use?

[mansr]

43 minutes ago, Summit said:

What I try to illustrate and explain is that it’s important to get a representative survey sample to assure that inferences and conclusions can reasonably extend from the sample to the population/audiophiles as a whole.

This is especially true if the responders are self-selected.

[tailspn]

5 hours ago, mansr said:

...DSD data itself has no frame structure, so there is no inherent reason to restrict the cut points. SACD, being based on the DVD physical medium, may well require sector alignment. I know DVD does. What size "frames" does Pyramix use?

Yes, I believe a raw channel modulator output has no frame structure. But immediately thereafter it's packaged (formatted) into the DSDIFF format for storage, processing, and playing out. It's not a Pyramix invention or proprietary  mechanism, but inherent to the DSDIFF (DSD) format.

 

This DSDIFF 1.5 specification pdf may help to explain:

 

http://www.sonicstudio.com/pdf/dsd/DSDIFF_1.5_Spec.pdf

 

Detailed in it are DSD Frames made of data Chunks, which in turn are various support chunks including sectorized raw DSD data bits. To the best of my knowledge, frames are time based, so at different DSD bitrates, there will be a different number of bits contained in each frame/chunk.

 

The point remains however, whether dff, or dsf at any available bitrate, DSD data is organized into frames. Any DSD editor not separating a full album file into tracks (slicing) along those frame boundaries will corrupt that frame, and it in turn will be discarded. The effect then when playing will be a click at the track transition bridging two different DC offset levels in one bit time.

 

The original question (Ted, Mr Wizard, boggie?), does sacd_extract contain the action of automatically searching for the nearest frame boundary to the PQ Marker to perform the track slice, or just do the slice within a frame time corresponding to the PQ Marker time?

 

Hope this helps,

 

Tom

 

 

[mansr]

1 minute ago, tailspn said:

Yes, I believe a raw channel modulator output has no frame structure. But immediately thereafter it's packaged (formatted) into the DSDIFF format for storage, processing, and playing out. It's not a Pyramix invention or proprietary  mechanism, but inherent to the DSDIFF (DSD) format.

DSDIFF is just one of several formats. DSF, which is far more common for downloads, has no such limitation.

 

1 minute ago, tailspn said:

This DSDIFF 1.5 specification pdf may help to explain:

 

http://www.sonicstudio.com/pdf/dsd/DSDIFF_1.5_Spec.pdf

 

Detailed in it are DSD Frames made of data Chunks, which in turn are various support chunks including sectorized raw DSD data bits. To the best of my knowledge, frames/chunks are time based, so at different DSD bitrates, there will be a different number of bits contained in each frame/chunk.

 

The point remains however, whether dff, or dsf at any available bitrate, DSD data is organized into frames. Any DSD editor not separating a full album file into tracks (slicing) along those frame boundaries will corrupt that frame, and it in turn will be discarded. The effect then when playing will be a click at the track transition bridging two different DC offset levels in onee bit time.

DSDIFF packs 1-bit samples into 8-bit bytes. The resulting per-channel byte streams are then interleaved and written to disk. If the number of samples to write isn't a multiple of 8, the file has to be padded or the odd samples at the end discarded. DSF uses a different block structure, but more importantly, it records the exact number of samples in the file header. If the final block is incomplete, reading software will know this and can stop processing at the actual end. I know this because I have done it.

[tailspn]

36 minutes ago, mansr said:

DSDIFF is just one of several formats. DSF, which is far more common for downloads, has no such limitation.

 

DSDIFF packs 1-bit samples into 8-bit bytes. The resulting per-channel byte streams are then interleaved and written to disk. If the number of samples to write isn't a multiple of 8, the file has to be padded or the odd samples at the end discarded. DSF uses a different block structure, but more importantly, it records the exact number of samples in the file header. If the final block is incomplete, reading software will know this and can stop processing at the actual end. I know this because I have done it.

That could well be, for Pyramix only operates on dff files, and converts any dsf file back to dff. .dsf is a delivery format, not a recording or processing format. It's the end of the line from my perspective, and of little interest.

 

But that doesn't answer my question; does sacd_extract routinely produce track transition click/ticks from slicing ISO DSD files? (which BTW are .dff). If so, does that mean its slicing at the PQ Marker time point within a dff frame, then later converting to dsf?

 

Also, does any software exist that automatically slices ISO files into tracks from the embedded PQ Markers that does not click at the track transitions?

 

Thanks!

[Kal Rubinson]

46 minutes ago, tailspn said:

But that doesn't answer my question; does sacd_extract routinely produce track transition click/ticks from slicing ISO DSD files? (which BTW are .dff). If so, does that mean its slicing at the PQ Marker time point within a dff frame, then later converting to dsf?

There is a method which combines the work of several people but is listed on CA as Bogi's method at 

Since I have been using it, I have not heard any clicks although, I must admit that I rarely heard them before this.   I believe that DACs vary in their susceptibility to making the track transition errors audible.

[Darryl R]

1 hour ago, tailspn said:

That could well be, for Pyramix only operates on dff files, and converts any dsf file back to dff. .dsf is a delivery format, not a recording or processing format. It's the end of the line from my perspective, and of little interest.

 

But that doesn't answer my question; does sacd_extract routinely produce track transition click/ticks from slicing ISO DSD files? (which BTW are .dff). If so, does that mean its slicing at the PQ Marker time point within a dff frame, then later converting to dsf?

 

Also, does any software exist that automatically slices ISO files into tracks from the embedded PQ Markers that does not click at the track transitions?

 

Thanks!

 

Mr. Tailspn,

 

Your patience with us is impressive (you never allow it to put you into a tailspin ).

 

I sent a private message to Mr. Wicked asking if he considered frames when slicing, though it looks like he hasn't visited the site since 2014.  Since concatenating the files (with a free editor such as the one from TEAC) eliminates the pops on my system, it does bring this into question.

[beerandmusic]

9 hours ago, Em2016 said:

 

Audirvana too.

 

There's no consensus about the best software either.

 

But just to share my own observations I've used Roon, Audirvana and HQPlayer Embedded.

 

In terms of SQ, I prefer HQPlayer. Jussi, it's developer, has been working on his filters for over 20 years he says. For DACs that support DSD512 over USB (or I2S) I'm yet to see Jussi say there's a DAC that measures better NOT up-sampling to DSD512 (i.e. I believe all DACs he's measured perform best at DSD512, if they support it).

 

Roon is relatively new to DSD up-sampling (just over 12 months actually) so I expect their filters and modulators to improve over time but it sounds quite nice.

 

Audrivana was quite nice in my experience.
 

The differences for me are subtle, rather than night and day.

 

All of my above experiences were with an iDSD Black Label and Pro-Ject S2 DAC.

 

I've heard the T+A DAC8 DSD can be even more revealing of differences due to better DAC performance overall (discrete design and better power supply and analogue sections too). It's a deal breaker for me until it supports DSD512 over Linux. At the moment DSD512 is only supported on Windows but Linux support is slowly being worked on.

 

Thanks for this information, especially about Audirvana also supporting DSD upsampling.  I understand they now make a windows product and it may be worth trying.  I own jriver and hqplayer already, but i don't care for the interface of either one....and i have no interest in buying a "subscription based product" (roon), besides it seems hqplayer is better than roon for sq, which i already own.  This is the kind of information i was seeking.

[Jud]

15 hours ago, Em2016 said:

 

Some good objective reasons to explain this discussed by @Miska and @mansr in this short (3 pages) thread:

 

https://www.computeraudiophile.com/forums/topic/43621-ifi-audio-ships-pro-idsd-dacstreamer/

 

 

I have seen these differences discussed, and it’s quite useful I have one of the DACs measured (micro-iDSD).

 

But I’m not sure how audible raising the noise floor from really damn low to damn low ought to be. ? 

 

Also, I’m not sure how it would relate to the subjective effects I’m thinking I hear, which include greater stability of instrument and vocal positions, an overall better defined and slightly larger soundstage, and increased prominence and ease of following both lead and backing vocals.

[4est]

22 minutes ago, beerandmusic said:

 

Thanks for this information, especially about Audirvana also supporting DSD upsampling.  I understand they now make a windows product and it may be worth trying.  I own jriver and hqplayer already, but i don't care for the interface of either one....and i have no interest in buying a "subscription based product" (roon), besides it seems hqplayer is better than roon for sq, which i already own.  This is the kind of information i was seeking.

MusicBee can be a library front end for HQPlayer. It is a great inexpensive product for general use, and works great with HQPlayer. It doesn't have Tidal though.

 

Drag and drop works well too.

[beerandmusic]

13 minutes ago, 4est said:

MusicBee can be a library front end for HQPlayer. It is a great inexpensive product for general use, and works great with HQPlayer. It doesn't have Tidal though.

 

Drag and drop works well too.

now this sounds interesting...i have seen musicbee before and i liked it's simplistic interface....can you provide step-by-step how to configure it to work with hqplayer?

[4est]

29 minutes ago, beerandmusic said:

now this sounds interesting...i have seen musicbee before and i liked it's simplistic interface....can you provide step-by-step how to configure it to work with hqplayer?

No it has been awhile, read the manual. It is not hard if I can do it. It is cheap and works great as I recall. If no other option becomes available (that's a hint bibo), I intend to go back to it when my Roon subscription is up. I'm not going to sugar coat it, I hate the way Roon works, and even more so their pushing of MQA. I am just going to give up streaming PCM>DSD.

[beerandmusic]

13 minutes ago, 4est said:

No it has been awhile, read the manual. It is not hard if I can do it. It is cheap and works great as I recall. If no other option becomes available (that's a hint bibo),

 

he has had many hints over the years, and i don't suspect he will ever do anything about the interface.  It probably takes a different type of programmer where he doesn't have expertise and his profits must not be good enough to warrant contracting with a gui developer.  If it is simple to configure any gui using hqplayer, without having nuances, like how to start, configure, playback, etc...then perhaps using musicbee is something he should push and provide a guide for.  I am sure there are many that are discouraged with his interface where they just gave up on hqplayer.

[Miska]

2 hours ago, mansr said:

DSDIFF is just one of several formats. DSF, which is far more common for downloads, has no such limitation.

 

I think most of DSD content is mastered in DSDIFF (the Philips format) and only converted to DSF (the Sony format) for delivery. Not sure if the Sonoma workstation uses DSF as file format or if it also uses DSDIFF.

 

So the DSDIFF properties apply to mastering stages. And as a side effect, you probably don't find much DSF content that wouldn't have multiple-of-8 number of bits.

 

But given the bitrate, worrying about this 8-bit boundary is not much of an issue, even less at DSD128 or DSD256 rates. It has more to do about the waveform position where the tracks are cut (you don't want to cut at the peak). This isn't much problem when you play SACD because you are not going to have "gapless" transitions between unrelated tracks. This is more relevant when you take that content off SACD disc and put it into files and then make a playlist that mixes tracks from multiple SACD discs in random order. So in this respect, mastering an SACD disc is different from mastering a DSD download... This is why I do pop suppression when playing DSD tracks from a playlist and don't do it when playing DSD tracks from album (yes, HQPlayer makes difference between "album" and "playlist" transports).

 

[Miska]

12 minutes ago, beerandmusic said:

he has had many hints over the years, and i don't suspect he will ever do anything about the interface.

 

I just made a facelift to the alternative GUI (touch/fullscreen mode) built into HQPlayer, but so far only one people has commented about it...

 

But also when I look at various other players, I also completely fail to understand what good someone sees in many of the interfaces.

 

13 minutes ago, beerandmusic said:

It probably takes a different type of programmer where he doesn't have expertise and his profits must not be good enough to warrant contracting with a gui developer.

 

Maybe, I'm not a graphics artist kind of guy... And yes, so far I haven't got money to hire anybody.

 

[beerandmusic]

30 minutes ago, Miska said:

 

I just made a facelift to the alternative GUI (touch/fullscreen mode) built into HQPlayer, but so far only one people has commented about it...

 

But also when I look at various other players, I also completely fail to understand what good someone sees in many of the interfaces.

 

 

Maybe, I'm not a graphics artist kind of guy... And yes, so far I haven't got money to hire anybody.

 

 

I respect what you have brought to the community!

I think you should charge double and hire someone though (wink)..i would be a buyer....maybe if you ask for investment ahead of time, you would find a lot of support (kind of like fundme) amongst your current clients.

How many buyers do you have?  If half of them were to give you $50 today, would that be enough to have someone develop a simplistic interface like musicbee?

 

 

[jabbr]

1 hour ago, Miska said:

But also when I look at various other players, I also completely fail to understand what good someone sees in many of the interfaces.

 

Let's see ... Roon has an IOS app that enables search, selection, volume control etc from my iPhone. The graphics allow browsing an artists works etc. I can also switch between different devices, even direct music to my iPhone... now only if Roon could allow me to select among different NAAs ...

[tailspn]

2 hours ago, Miska said:

I think most of DSD content is mastered in DSDIFF (the Philips format) and only converted to DSF (the Sony format) for delivery. Not sure if the Sonoma workstation uses DSF as file format or if it also uses DSDIFF.

 

Sonoma only works with DSDIFF, since it is DSD64 only, and is strictly a DAW. The Sony spec for DSF came out quite a few years after the Sonoma development and heyday as a way to promote DSD delivery outside of SACD, and after most people using Sonoma as a DAW had already shelved it. There are a few exceptions, including myself, most notably 5/4 Productions, Bruce Brown at Puget Sound, and a few others. Sonoma was the first DSD DAW available, and while very capable, was soon eclipsed by the Philips technology Merging Technologies Pyramix. It, and its SADiE competitor however were pure DSD post production editors without PCM conversion, which capability may soon return.

[Jud]

1 hour ago, jabbr said:

 

now only if Roon could allow me to select among different NAAs ...

 

There’s RealVNC. ? (I know that’s not the convenient all-in-one interface you’re after.)

[jabbr]

1 minute ago, Jud said:

 

There’s RealVNC. ? (I know that’s not the convenient all-in-one interface you’re after.)

 

I use VNC currently. I'm tempted to write an IOS app just so I can switch HQP remotely using the XML interface ... but the interface doesn't allow iteration over the NAAs. I could use SSH on my iPhone too ...