Upsampling by dac, can't hear an improvement.

[unbalanced output]

11 hours ago, jabbr said:

1) There are outstanding examples of software upsampling to both R2R based and SDM based DACs.

2) upsampling PCM to a SDM DAC might not have a benefit. 

3) The DAC in question is not typical and benefits of offboard upsampling depends on the DAC being able to accept.

4) Only rarely does something work for everything.

 

I think your DAC does internal PCM->SDM conversion as so if accepts DSD in, that might be better 

 

Came here for that. When the DAC has its own internal filter, the upsampled data is "shaped" by the filter and should look very similar to the original filtered data. 

 

With the R2R on the Lampizator, upsampling does have a perceptible difference. Perhaps not to the point that one could easily spot but is noticeable. But I can certainly spot the difference between playing nos and upsampled to DSD (using the 2nd engine). Just the other day I thought the sound was somewhat changed from one day to the other (!) and indeed I had disabled upsampling - I don't think I would be able to tell the difference if I was comparing PCM streams. 

[Lebouwsky]

21 minutes ago, unbalanced output said:

 

Came here for that. When the DAC has its own internal filter, the upsampled data is "shaped" by the filter and should look very similar to the original filtered data. 

 

With the R2R on the Lampizator, upsampling does have a perceptible difference. Perhaps not to the point that one could easily spot but is noticeable. But I can certainly spot the difference between playing nos and upsampled to DSD (using the 2nd engine). Just the other day I thought the sound was somewhat changed from one day to the other (!) and indeed I had disabled upsampling - I don't think I would be able to tell the difference if I was comparing PCM streams. 

I think we're on the same boat when it comes to upsampling PCM. But when it comes to upsampling PCM to DSD I have no experience.  I recently invested in good PCM dac. When you discribe the PCM tot DSD upsampling as "somewhat changed" I don't feel the need to go DSD real soon. What's your opinion about PCM to DSD?

[firedog]

13 hours ago, plissken said:

Hey Alex, when are you going to do that Double Blind test that you ran away from like a scalded pig @ WBF forum?

 

10 hours ago, plissken said:

I didn't name call anyone FD. I simply stated a fact. 

Maybe technically not name calling as you used the word "like", but certainly comes off as name calling. The "scalded pig" phrase could have been omitted and your comment would have made he same point.

[firedog]

24 minutes ago, Lebouwsky said:

I think we're on the same boat when it comes to upsampling PCM. But when it comes to upsampling PCM to DSD I have no experience.  I recently invested in good PCM dac. When you discribe the PCM tot DSD upsampling as "somewhat changed" I don't feel the need to go DSD real soon. What's your opinion about PCM to DSD?

 

I wouldn't get audiophillia nervosa over comments about upsampling you read on this site. I had a ESS based Dac that sounded better with music upsampled to DSD, and now I have a DAC that upsamples everything to PCM and sounds better that way.

If you like the sound of your DAC, I wouldn't worry about it too much.

[unbalanced output]

28 minutes ago, Lebouwsky said:

I think we're on the same boat when it comes to upsampling PCM. But when it comes to upsampling PCM to DSD I have no experience.  I recently invested in good PCM dac. When you discribe the PCM tot DSD upsampling as "somewhat changed" I don't feel the need to go DSD real soon. What's your opinion about PCM to DSD?

 

I haven't heard other DSD DACs that are capable of playing native streams without any kind of conversion. I think the Lampi is quite unique when it comes to DSD implementation and theyre certainly one level above the mainstream (t+a is another example of a dac praised for is Dsd). I have heard one or two other DACs where DSD was not all that great.

 

Mind you that the Lampizator actually has two separate circuits which share the same interface and clock, therefore  the difference may not be only due to the digital stream but due to the implementation as well. As fellow ? ? just wrote, another DAC may have a better R2R circuit (I'm sure there are better face in this respect than the Lampi) but not a so good DSD engine and the result may be the opposite. Just stick to the one which sounds best on your equipment...

[jabbr]

3 hours ago, Lebouwsky said:

I think we're on the same boat when it comes to upsampling PCM. But when it comes to upsampling PCM to DSD I have no experience.  I recently invested in good PCM dac. When you discribe the PCM tot DSD upsampling as "somewhat changed" I don't feel the need to go DSD real soon. What's your opinion about PCM to DSD?

 

The standard format for CDs is PCM. The internal format for most ADCs and DACs is SDM/DSD. Many DAC chips are very complicated and first upsample/convert PCM to DSD/SDM internally before converting to analogue. In these cases there can be a benefit for outboard conversion to DSD as this step is math intensive.

 

Other DACs which are truly native PCM are called R2R based on the resistor ladder. These DACs have no benefit to outboard DSD conversion. 

 

I use  @Miska'S HQPlayer for  outboard DSD conversion (DSD512) and either HQPlayer or @PeterSt XXHighEnd for PCM upsampling before feeding into the DAC (up to 700 kHz range). I find this makes a terrific improvement for me with CD source in particular.

 

DACs that won't bypass their own internal filters and preprocessing units would not be expected to show the same improvements with outboard upsampling. Does your DAC have an R2R ladder? Does it accept DSD input? If it has an R2R ladder then you don't want to convert to DSD. If it doesn't accept DSD in then you can't. Otherwise worth trying.

[plissken]

4 hours ago, firedog said:

 

Maybe technically not name calling as you used the word "like", but certainly comes off as name calling. The "scalded pig" phrase could have been omitted and your comment would have made he same point.

 

Oh, I can't stop laughing. Sorry about the micro-aggression. I'll prepare your safespace now. 

 

And there's no technically about it. Look up the definition some time.

[unbalanced output]

1 hour ago, jabbr said:

 

The standard format for CDs is PCM. The internal format for most ADCs and DACs is SDM/DSD. Many DAC chips are very complicated and first upsample/convert PCM to DSD/SDM internally before converting to analogue. In these cases there can be a benefit for outboard conversion to DSD as this step is math intensive.

 

Other DACs which are truly native PCM are called R2R based on the resistor ladder. These DACs have no benefit to outboard DSD conversion. 

 

I use  @Miska'S HQPlayer for  outboard DSD conversion (DSD512) and either HQPlayer or @PeterSt XXHighEnd for PCM upsampling before feeding into the DAC (up to 700 kHz range). I find this makes a terrific improvement for me with CD source in particular.

 

DACs that won't bypass their own internal filters and preprocessing units would not be expected to show the same improvements with outboard upsampling. Does your DAC have an R2R ladder? Does it accept DSD input? If it has an R2R ladder then you don't want to convert to DSD. If it doesn't accept DSD in then you can't. Otherwise worth trying.

 

Correct, many DACs which accept both DSD and PCM will first convert to their native input (mostly PCM) which is only then converted to 1 bit before the SDM. Converting a PCM stream to DSD in this case is kinda pointless. 

 

Without a conversion chip, a R2R cannot accept a DSD stream - what the Polish guys do is to divert the DSD streams from the ladder to the DSD circuit.

[Ralf11]

math intensive ===> better GPUs

 

 

it will be fun to rip the AI system out of your self-driving Tesla 500 and re-purpose it...

[firedog]

8 hours ago, plissken said:

 

Oh, I can't stop laughing. Sorry about the micro-aggression. I'll prepare your safespace now. 

 

And there's no technically about it. Look up the definition some time.

Looked it up. Still don't understand why it isn't an insult. I know nothing about micro aggressions. I just don't know why basic civility can't be maintained.

[jabbr]

6 hours ago, unbalanced output said:

Correct, many DACs which accept both DSD and PCM will first convert to their native input (mostly PCM) which is only then converted to 1 bit before the SDM. Converting a PCM stream to DSD in this case is kinda pointless. 

 

The conversion from PCM to DSD is math intensive and it may be more cost effective to use a commodity CPU+GPU to do the conversion in which case DSD512 is easy. This makes things much easier for the following analogue filter and thus more easily avoids artifacts from creeping back down into the audible range e.g. the oft described CD "brickwall" filter

[plissken]

42 minutes ago, firedog said:

Looked it up. Still don't understand why it isn't an insult. I know nothing about micro aggressions. I just don't know why basic civility can't be maintained.

 

So you are of the opinion that what Alex said isn't hypocritical?

 

Oh, I agree with you that basic civility can't be maintained. You are talking to the wrong person about this. Alex made it personal in this thread. It was fine up to that point. 

[4est]

11 minutes ago, plissken said:

 

So you are of the opinion that what Alex said isn't hypocritical?

 

Oh, I agree with you that basic civility can't be maintained. You are talking to the wrong person about this. Alex made it personal in this thread. It was fine up to that point. 

Really now? In rereading, Alex never mentioned you and yet you likened him to a scalded pig straightaway. Methinks you have got it wrong here.

[Ralf11]

can we all just yell "poot on yoot" into the computer screen and end the OT fluff shouting, so the thread can continue on upsampling?

[Lebouwsky]

If you'd imagine a track like a digital picture with a certain resulution, lets say 800x600 dots of red blue yellow and the dot in the upper right is red. If the picture would be upsampled to a higher resulution by a factor 4 (1600x1200), all the 4 upper right square dots should be red. The picture remains the same, there's no more info, just more dots or bits. To me, with limited knowledge of digital music files, that's what upsampling in pcm does.

 

On the other hand, imagine this same picture, but contrast is made higher, or tones are softened or whatever filter is being used. The picture can be changed to a result, more apealing to the viewers preferences. I can imagine that these filters have more freedom of movement when the picture is brought to a higher resulution. I can also imagine these filters have more freedom if a 4th colour is added to the pallet (cyan, magenta, yellow key). 

 

Feel free to shoot.

 

[jabbr]

15 hours ago, Lebouwsky said:

Feel free to shoot.

It doesn't work that way. Not exactly like a photo but upsampling wouldn't just replicate pixels. Consider an antialiasing filter. If this were at a specific corner frequency, upsampling, without increasing information content, would reduce the blur caused by the antialiasing filter. 

[Rexp]

I've tried both hardware and software upsampling and it didn't make typical poor quality digital files sound significantly better. I am intrigued enough to give the new Chord upsampler a go though as this review states it transforms redbook:

https://audiobacon.net/2017/08/22/chord-blu-mk-2-m-scaler-cd-transport-review-digital-dressed-in-analog/

 

[pkane2001]

8 hours ago, jabbr said:

It doesn't work that way. Not exactly like a photo but upsampling wouldn't just replicate pixels. Consider an antialiasing filter. If this were at a specific corner frequency, upsampling, without increasing information content, would reduce the blur caused by the antialiasing filter. 

 

Photo/image resampling is very similar to audio resampling. With filters, interpolation, and ringing. Except it's in two dimensions. Nobody does upsampling by duplicating pixels, that's the worst possible way to do it. I assume nobody does audio upsampling by duplicating audio samples, either. Upsampling operation acts as an averaging filter and introduces a blur. A proper filter helps recover most (but not all) of the original information that was in the sample, but it does not add to or improve on existing information. The best it can do is not to destroy too much.

 

[Ralf11]

I think the real problem here is a lack of confirmation bias.  I am crowd funding a device which will automatically add a big but controllable dosage to your system.  Send money!

 

or send lawyers guns and money...

[rich-40]

Besides the fact that you try to make a sound of x bits, turn to y bits...... ,which data do you add??in the freq domain that is....nothing,it's nice it adds some sound you are missing, and then a daw can  help you or a dsp   or just a 200euro soundenchancer and gear like that , but have you got the best clock available, otherwise i don't see the point of upsampling.

[Lebouwsky]

This night for the first time I used HQplayer, installed on a W10 laptop, sms200 in HQplayer NAA mode. The sms200 usb output is connected to an Uptone Regen, which gets straight in a Schiit Eitr usb to spdif converter by a hard usb adapter. From there my AMR dp777 is fed by a 75 ohm coax. The Eitr is quite special by the way.

 

Anyway, HQplayer upsamples everything to 192kHz, with the poly-sync shrt filter and the NS4 dither. I don't have a clue what they mean or what they do, but the sound is superb. I just can't stop listening, it's so emotional engaging. OMG!!! My system has never sounded so good.

[wushuliu]

Wait blind testing that can withstand serious scrutiny and meet criteria for peer-review can be had just by flying someone out to your house? OMG I had no idea. Are ABX kits available on Amazon?

 

I didn't realize scientifically accurate blind testing could just be done by posting on the internet! Awesome! Here I was thinking you'd have to do all this work of establishing criteria, equipment, personnel... pfft. 

 

Thank you Internet!

[Lebouwsky]

1 hour ago, wushuliu said:

Wait blind testing that can withstand serious scrutiny and meet criteria for peer-review can be had just by flying someone out to your house? OMG I had no idea. Are ABX kits available on Amazon?

 

I didn't realize scientifically accurate blind testing could just be done by posting on the internet! Awesome! Here I was thinking you'd have to do all this work of establishing criteria, equipment, personnel... pfft. 

 

Thank you Internet!

I see you speak fluent Sarcasm.

[OldBigEars]

@ lebouwsky I played around with software and hardware upsampling, including DSD, for quite some time.  My listening acuity must clearly be impaired as I struggled to hear any substantial improvements with anything beyond 24/96.  The hours I wasted twiddling with Izotope settings may be amongst the least productive of my life.  There are so many things - in my experience - that matter more.  People are obsessed by the numerical size of music files in much the same way that 15 years ago many people valued digital cameras based primarily on their maximum megapixel size. I'm convinced it's a passing fad.

[Lebouwsky]

30 minutes ago, OldBigEars said:

@ lebouwsky I played around with software and hardware upsampling, including DSD, for quite some time.  My listening acuity must clearly be impaired as I struggled to hear any substantial improvements with anything beyond 24/96.  The hours I wasted twiddling with Izotope settings may be amongst the least productive of my life.  There are so many things - in my experience - that matter more.  People are obsessed by the numerical size of music files in much the same way that 15 years ago many people valued digital cameras based primarily on their maximum megapixel size. I'm convinced it's a passing fad.

 

It's true that some people are obsessed be numbers, that was the initial thought when I started this thread. But then I was adviced by a fellow member in this thread to at least try upsampling in the software by HQplayer instead of hardware.

 

The improvement really was there, even my wife, who doesn't care about hifi, noticed, and I even didn't tell her something changed. And I'm not biased (no investment was done, free trial) rather sceptical. Maybe it has something to do with the fact that the sms200 now acts as a NAA and the heavy lifting is don by the laptop? I don't know, but this is a keeper.