Offline Upsampling

[bogi]

$2250 here in the USA

 

Uff ... more than my whole audio chain! )

I just checked, that JRiver really provides this functionality.

[Geoffrey Armstrong]

Geoffrey, I downloaded it from here: A new resampler DSP for foobar2000 - Hydrogenaudio Forums

 

Thanks Bogi. I did that; but the max sample rate it will allow me is 192khz or 4x. How are you getting up to 352.8 with it?

[Geoffrey Armstrong]

I began builtin to AuI 1-bit audio resolution on 2,8 and 5,6 MHz. Result of the convertion will *.dsd file.

 

As I understand DSD is trademark by Philips. This standard use pulse-dencity modulation. Patented decoder only. I will create own (other than Phillips's algorithm) decoder for converting multibit sequence to on-bit sequence for any sampling rate.

 

This is great new Yuri! When you say "any sample rate" do you mean both incoming PCM sample rates and outgoing DSD rates? For example, will you be able to create a DSD256 fie?

[bogi]

Thanks Bogi. I did that; but the max sample rate it will allow me is 192khz or 4x. How are you getting up to 352.8 with it?

 

The combo box contains values only up to 192k, but just write 352800 in the input box. It really functions so.

[audiventory]

Thanks, Yuri, this is very encouraging and welcome news. When you say the result will be a *.dsd file, do you mean rather than .dsf or .dff (which, as far as I know, are the file types my software players will play - .dsf preferred because it is better for metadata)?

 

Jud, *.dff file demand license fee. I plan bought it in the future due it's popular format.

 

DSF fine support standard id3 metadata. And I process input dsd files. Now I will process output dsd also. DSF already has part of programming code, so release 1-bit stream via DSF will faster.

 

As I right understand you have player for DSF?

[Geoffrey Armstrong]

Thanks Geoff, I appreciate the info. At some point I may try this (other than the part where you play the resulting file with HQPlayer ).

 

You're welcome Jud.

 

I've been brainstorming with myself this morning, which is why I came up with a crazy idea. What if we could combine Yuri's offline up-sampling with a player and have the files up-sampled, read into memory or a temporary space on disk, then played and finally trashed at the end?

 

I'm thinking that Yuri's re-sampler seems to be very efficient and still manages to sound better than any offline up-sampling I've tried in the past. For an average size album you might have to wait 5 - 10 minutes before the album could begin to play in glorious 352.8 or even DSD128/256 sound. If this could be combined with a player like HQPlayer to read those files and pay them back, we could have the best of both worlds. The benefits of offline up-sampling, which seems to sound a bit better, with the convenience (almost) of real-time up-sampling/conversion.

 

The more I think about it the less crazy it seems. After all Amarra makes you wait now, if you select the play from cache/memory option and XXHighEnd definitely makes you wait quite a while for reading into memory, shutting down all the services; etc.

 

In case you're old enough to remember playing vinyl, obviously I'm too young then I guess those old guys would spend a good 5 to 10 minutes cleaning the disk and stylus professionally before being able to play just 20 minutes of music. In today's world of instant gratification, have we all become to used to having it now?

 

In my wildest dreams I even imagine taking files from "out there" in the cloud and doing the same. e.g. with Qobuz or other internet/cloud streams.

 

My objection to offline up-sampling was the amount of space the files take up. Then, as you mentioned, you'd probably need to trash them at some point anyway, once these methods have been replaced by something better.

 

Perhaps we should have a vote: How long would people be willing to wait for substantially improved SQ from each album from their existing libraries?

[audiventory]

This is great new Yuri! When you say "any sample rate" do you mean both incoming PCM sample rates and outgoing DSD rates? For example, will you be able to create a DSD256 fie?

 

Nothing difference what input sampling rate value. It may be 44 kHz / 1 bit, but I don't think you like it

 

While I don't know how will sounding of these encoding. Before I don't make experiments with 1-bit encoding. Anyway I report what I ahieved. I assume in Settings appears some (1 or 2) parameters of 1-bit encoding which you can tune for achieve more pure sound.

 

Probably it will quite slow processing. But I think first task is getting pure sound. If I can do it, then possibly work with performance.

 

Now I can't promise but try.

[4est]

What about a third computer to upsample and send to a Ram Disk or HD akin to XXHighend's playback drive?

You're welcome Jud.

 

I've been brainstorming with myself this morning, which is why I came up with a crazy idea. What if we could combine Yuri's offline up-sampling with a player and have the files up-sampled, read into memory or a temporary space on disk, then played and finally trashed at the end?

 

I'm thinking that Yuri's re-sampler seems to be very efficient and still manages to sound better than any offline up-sampling I've tried in the past. For an average size album you might have to wait 5 - 10 minutes before the album could begin to play in glorious 352.8 or even DSD128/256 sound. If this could be combined with a player like HQPlayer to read those files and pay them back, we could have the best of both worlds. The benefits of offline up-sampling, which seems to sound a bit better, with the convenience (almost) of real-time up-sampling/conversion.

 

The more I think about it the less crazy it seems. After all Amarra makes you wait now, if you select the play from cache/memory option and XXHighEnd definitely makes you wait quite a while for reading into memory, shutting down all the services; etc.

 

In case you're old enough to remember playing vinyl, obviously I'm too young then I guess those old guys would spend a good 5 to 10 minutes cleaning the disk and stylus professionally before being able to play just 20 minutes of music. In today's world of instant gratification, have we all become to used to having it now?

 

In my wildest dreams I even imagine taking files from "out there" in the cloud and doing the same. e.g. with Qobuz or other internet/cloud streams.

 

My objection to offline up-sampling was the amount of space the files take up. Then, as you mentioned, you'd probably need to trash them at some point anyway, once these methods have been replaced by something better.

 

Perhaps we should have a vote: How long would people be willing to wait for substantially improved SQ from each album from their existing libraries?

[Geoffrey Armstrong]

Nothing difference what input sampling rate value. It may be 44 kHz / 1 bit, but I don't think you like it

 

While I don't know how will sounding of these encoding. Before I don't make experiments with 1-bit encoding. Anyway I report what I ahieved. I assume in Settings appears some (1 or 2) parameters of 1-bit encoding which you can tune for achieve more pure sound.

 

Probably it will quite slow processing. But I think first task is getting pure sound. If I can do it, then possibly work with performance.

 

Now I can't promise but try.

 

…and I'm sure that's already very much appreciated here Yuri. Thank you!

[Geoffrey Armstrong]

What about a third computer to upsample and send to a Ram Disk or HD akin to XXHighend's playback drive?

 

…or perhaps one computer and an NAA as EuroDriver suggested. I like the Ram Disk idea.

[4est]

Well that is already doable and sounds great! My post was in acknowledgement of someone saying that NAA was ALMOST as good as offline upsampling. Files could be sent to the intermediate box which could even be an NAS running a version of HQPlayer Embedded or something, then parsed out the the NAA.

…or perhaps one computer and an NAA as EuroDriver suggested. I like the Ram Disk idea.

[Geoffrey Armstrong]

Well that is already doable and sounds great! My post was in acknowledgement of someone saying that NAA was ALMOST as good as offline upsampling. Files could be sent to the intermediate box which could even be an NAS running a version of HQPlayer Embedded or something, then parsed out the the NAA.

 

To get the best out of HQPlayer though you still need some serious CPU horsepower. A NAS just wouldn't be sufficient. The poly-sinc family of filters are the best sounding and are quite demanding of CPU.

 

That's why it's a good idea to perform some up-sampling first, either offline or as a pre-process, then HQPlayer has less work to do. I see an NAA as providing better immunity than a direct USB connection to the computer that's doing the work. The problem is the DAC's supported by NAA's are limited. There are no drivers as yet for the Exasound DACs to work with an NAA, for example. I haven't been able to get an Antelope Zodiac DAC or a KingRec UCD s/pdif - USB interface to work with mine either. I've managed to get the Grimm LS1 speakers and the Oppo 105D to play with the NAA.

[4est]

I realize the horsepower and subsequent sonic hit when using too much cpu %. I was suggesting a device in between (NAS or even Ram Disk) to be used as a large temp buffer. If you want to split duties, one may be able to run two instances of HQPlayer (even on the smae maching). The first Desktop version > Network version. I would not be surprised if Miska or someone else could work out the code. Jack even?

To get the best out of HQPlayer though you still need some serious CPU horsepower. A NAS just wouldn't be sufficient. The poly-sinc family of filters are the best sounding and are quite demanding of CPU.

 

That's why it's a good idea to perform some up-sampling first, either offline or as a pre-process, then HQPlayer has less work to do. I see an NAA as providing better immunity than a direct USB connection to the computer that's doing the work. The problem is the DAC's supported by NAA's are limited. There are no drivers as yet for the Exasound DACs to work with an NAA, for example. I haven't been able to get an Antelope Zodiac DAC or a KingRec UCD s/pdif - USB interface to work with mine either. I've managed to get the Grimm LS1 speakers and the Oppo 105D to play with the NAA.

[Geoffrey Armstrong]

I realize the horsepower and subsequent sonic hit when using too much cpu %. I was suggesting a device in between (NAS or even Ram Disk) to be used as a large temp buffer. If you want to split duties, one may be able to run two instances of HQPlayer (even on the smae maching). The first Desktop version > Network version. I would not be surprised if Miska or someone else could work out the code. Jack even?

Apologies, I misunderstood.

[PeterSt]

What about a third computer to upsample and send to a Ram Disk or HD akin to XXHighend's playback drive?

 

I just found this thread, and if Yuri can offer an offline upsampler of any kind, I'll plug it into XXHighEnd in minutes. Ok, hours. It's all made for that and won't consume more disk space than the current playlist after upsampling.

 

(btw Geoff I load my complete (FLAC) albums into memory in a second or two, the PC running at 430 MHz, so not sure what you may be doing wrong).

 

Regards,

Peter

[Geoffrey Armstrong]

I just found this thread, and if Yuri can offer an offline upsampler of any kind, I'll plug it into XXHighEnd in minutes. Ok, hours. It's all made for that and won't consume more disk space than the current playlist after upsampling.

 

(btw Geoff I load my complete (FLAC) albums into memory in a second or two, the PC running at 430 MHz, so not sure what you may be doing wrong).

 

Regards,

Peter

 

I wasn't referring to the amount of time it takes to load the files into memory, so much as the time it takes to stop services in unattended mode, Peter. Anyway my point was, I can live with the wait for improved SQ. I was wondering what other people might consider a reasonable delay befoe playback begins, if it yields considerable improvements.

[PeterSt]

Geoffrey,

From yesturday I intensive think about including DSD into AuI. Of course, DSD encoding must appear in AuI. Now I try find DSD license owner. Also possible variant creating my own algorithm of 1-bit modulation encoding compatible with DSD (DSD is protected trademark).

 

Now I have rights of using DSF files, but algorithm encoding PCM to DSD is protected by this patent https://www.google.ru/patents/WO2006129215A2?cl=en As I know (while not exact) protected patent formula only - algorithm of converting PCM to DSD, but not sigma-delta modulation.

 

Yuri

 

Hi Yuri,

 

That patent you refer to is not about converting PCM to DSD as such. This is at least not how I read it;

It is about a method of converting PCM to DSD in lossless fashion, resulting in a "stream" that is both PCM and DSD compatible (also nice) for hardware which adapts to it. Actually the result is PCM with a a small portion of added bit code that allows the hardware to convert it to DSD if wanted.

 

So I wouldn't worry and you can do what you want.

Btw, I didn't read the whole patent, but as I see it at first glance it involves hardware, while it all can be done in software just the same (again, looked at first glance). This wouldn't break the patent ...

If you really want to know (do it this exact way) I first must read it more thoroughly.

 

Regards,

Peter

[PeterSt]

I wasn't referring to the amount of time it takes to load the files into memory, so much as the time it takes to stop services in unattended mode, Peter. Anyway my point was, I can live with the wait for improved SQ.

 

I understood the latter Geoff, and it is good reasoning.

But just in case let me rephrase : It takes a second or two before all plays, including the shutting off of all. BUT do it in Minimized OS Mode ...

 

Best regards,

Peter

 

PS: Keep up the good work !

[4est]

Peter: My apologies for not keeping up, but can XXHighend plays DSD?

I just found this thread, and if Yuri can offer an offline upsampler of any kind, I'll plug it into XXHighEnd in minutes. Ok, hours. It's all made for that and won't consume more disk space than the current playlist after upsampling.

 

(btw Geoff I load my complete (FLAC) albums into memory in a second or two, the PC running at 430 MHz, so not sure what you may be doing wrong).

 

Regards,

Peter

[Jud]

Jud, *.dff file demand license fee. I plan bought it in the future due it's popular format.

 

DSF fine support standard id3 metadata. And I process input dsd files. Now I will process output dsd also. DSF already has part of programming code, so release 1-bit stream via DSF will faster.

 

As I right understand you have player for DSF?

 

Yes, I know Audirvana+ will play .dff, .dsf or .iso (the latter is all the tracks on a DSD album in one compressed .iso file). HQPlayer will I believe play at least .dff and .dsf, don't know about .iso. And I haven't used JRiver, so I can't comment - but there are plenty of other people in the thread who will know. I don't know if any of them will play a DSD file with a .dsd file name extension. My preference is .dsf for various reasons (metadata support, I like dealing with individual tracks better than a whole album as an .iso file).

[Jud]

In case you're old enough to remember playing vinyl, obviously I'm too young then I guess those old guys would spend a good 5 to 10 minutes cleaning the disk and stylus professionally before being able to play just 20 minutes of music. In today's world of instant gratification, have we all become to used to having it now?

 

 

I was *much* faster than that at cleaning the disk and stylus! (You'd have to be crazy to do the full bore cleaning every time. I wasn't that insane, just did the full cleaning the first time before playing the record, then used a carbon fiber brush the rest of the time.)

[PeterSt]

Peter: My apologies for not keeping up, but can XXHighend plays DSD?

 

Hey Forrest,

 

Yes and no. No because I never made it public, yes because ... well ...

 

But let's say that I had other priorities forever. Still have, but if now offline upsampling filtering is for the better to you guys I'd like to grab some opportunity just because XXHighEnd can do these things right out of the box without as said - the inconveniences of explicitly dealing with it; just play the file, give it to the upsampling offline algorithm by means of a parameter setting and notice that it has done it under the hood.

All IOW, it keeps the processing out of the playback loop and this should be for the better indeed (OK, this is what I myself am telling all the time).

I can do it with Arc Prediction filtering just the same but this is a bit moot because of how all is organized. But if wanted (and you want to wait for it) ...

(but notice there's a 2GB limit for .WAV which (.WAV) is always the result for PCM).

 

 

Peter

[Jud]

 

Still have, but if now offline upsampling filtering is for the better to you guys I'd like to grab some opportunity just because XXHighEnd can do these things right out of the box

I can do it with Arc Prediction filtering just the same but this is a bit moot because of how all is organized. But if wanted (and you want to wait for it) ...

(but notice there's a 2GB limit for .WAV which (.WAV) is always the result for PCM).

 

 

Peter

 

You know I'd like to see this.

[PeterSt]

Jud, yes. But I didn't know this thread existed. Edit :

And you know that I know that you know that ...

 

So OK.

The when is a bit tight. So I will see.

 

Anyway thanks,

Peter

[joelha]

In reference to Jud's post, I too would like to see XXHighEnd offer off-line processing.

 

I like the idea of being able to use the player I want and get the quality of the file I want as well, even if it means I have to have more storage in the process.

 

I think off-line processing would be a great option for any software package that offers particulary good upsampling or filtering.

 

Joel