Blind test: Vinyl rip: Straight PCM vs. PCM down sampled from DSD

Hello

First of all; this thread is NOT about a comparison between DSD and PCM.

I've been testing my Tascam DA-3000 and I'd like to have your opinion on these two samples:

Sample A
Sample B

One of them is a straight PCM 96/24 file and the other one is a PCM 96/24 that has been down sampled from DSDIFF 5,6MHz. For the convertion I used AudioGate with TPDF dither. Which one sounds better to your ears? Thanks!

I also want to share with you my latest video; it serves me to show you the rest of the equipment that has been used in the making of this test. I hope you enjoy it:



Regards!

I'm not sure how well my hearing is today since I'm just getting over a really bad cold, but from what I hear through my Audio-Gd 12.1 DAC and Sennheiser HD 800's, I prefer the sound of sample B over sample A. I'm not sure what it is, but sample B seems to be a smidgen more transparent than sample A; but that could be my cold interfering.

One question - these are two separate recordings, one in DSD and the other straight 96k/24bit, correct? Have you tried the Saracon format converter by Weiss?

Download links:
http://peeplink.in/748cb3f51103

Archive password:  g7C~PJa(u,-e,5*AudioZ

I find the Weiss converter to be slightly better than anything else I've tried, including Korg's Audiogate - especially when down converting DSD to PCM. The interface is a bit clunky and dated, but the results are well worth it.

PS: Awesome video as always! 

Downloaded the files to an USB flash drive, then plugged it to my Sony BDP 173 Blu-ray player, which is plugged to my Hi-Fi setup (Nuova Elettronica preamp and power amp, Mission 701 loudspeakers).

Sample A has got a bit more bass and sounds a bit more lively. The bass gives the impression of higher volume, but in fact the volume is the same for both samples.
Sample A is definitely more faithful to the original recording.

SPOILER:
Sample A is the straight PCM, sample B is the downsampled one.

Thank you both for participating

...I prefer the sound of sample B over sample A. I'm not sure what it is, but sample B seems to be a smidgen more transparent than sample A; but that could be my cold interfering.

I posted the same test on another forum and sample B won hands down...

One question - these are two separate recordings, one in DSD and the other straight 96k/24bit, correct?

Yep, that's correct Both samples were recorded on the same day at the same time (one after the other LOL!)

Have you tried the Saracon format converter by Weiss?

Download links:
http://peeplink.in/748cb3f51103

Archive password:  g7C~PJa(u,-e,5*AudioZ

I find the Weiss converter to be slightly better than anything else I've tried, including Korg's Audiogate - especially when down converting DSD to PCM. The interface is a bit clunky and dated, but the results are well worth it.

Thanks!! Downloading as I type I'll keep you posted about my impressions. Many thanks!

Sample A has got a bit more bass and sounds a bit more lively. The bass gives the impression of higer volume, but in fact the volume is the same for both samples.
Sample A is definitely more faithful to the original recording.

Although I prefer sample B, I agree with you to some extent: simple A seems to have more bass...

SPOILER:
Sample A is the straight PCM, sample B is the downsampled one.

You revealed the "secret", how dare you! You naughty boy...LOL!   

I prefer the sound of sample B over sample A. I'm not sure what it is, but sample B seems to be a smidgen more transparent than sample A; but that could be my cold interfering.

I posted the same test on another forum and sample B won hands down...

Although I prefer sample B, I agree with you to some extent: sample A seems to have more bass...

Honestly I don't understand that.

I mean, tastes are personal and debating on them is totally useless, but the fact the people likes better the sample with less bass (that is, the right amount of bass, because digital won't add anything: it can only lose details) makes me think that their setups can't properly handle bass, thus making the sound less transparent and focused (more bass = more information to reproduce = more complexity, which need a perfect reproduction). So, to their ears the sample with less bass sounds "better", but actually they are being misled.

IMHO.

Although I prefer sample B, I agree with you to some extent: sample A seems to have more bass...

Honestly I don't understand that.

Well, I dunno the reason why the others prefer B over A, but in my case I did not say that I preferred sample B "because" it had more bass. To me, it feels as if B had more 3D imaging and air between instruments, more "sparkle" and dynamism...but yes, OTOH I can hear more bass in sample A. Of course, they're very subtle diffenrences. I can feel the same differences (to a greater or lesser degree) listening with my AKG K701 (connected direct to the Tascam or using my headphone preamp), through speakers, pc... Anyway, you have better hearing than me and maybe I'm just being misled, I don't rule out anything

BTW, I tried Saracon converter but...:



 

Well, I dunno the reason why the others prefer B over A, but in my case I did not say that I preferred sample B "because" it had more bass. To me, it feels as if B had more 3D imaging and air between instruments, more "sparkle" and dynamism...

I got that. But I'm saying that I hear in sample A what you're hearing in sample B!
3D imaging and air between instruments, and dynamism, I hear them better in sample A. Bass is of course an ingredient in that recipe. I just suspect that in a setup that can't proprly handle bass, bass actually kills all of that, while in a well tuned setup, bass actually emphasizes all that.

This is not to say that my setup is better than other people's ones, but I just noticed this fact a long time ago. To me this is the same reason that made digital take over analog 30 years ago: digital clearly lacks bass compared to analog, so in many not so great Hi-Fi setups digital sounded crisper and better due to its lack of bass. But in a properly performing setup, analog is definitely better.

Again, in my humble opinion.

Although I prefer sample B, I agree with you to some extent: sample A seems to have more bass...

Honestly I don't understand that.

Well, I dunno the reason why the others prefer B over A, but in my case I did not say that I preferred sample B "because" it had more bass. To me, it feels as if B had more 3D imaging and air between instruments, more "sparkle" and dynamism...but yes, OTOH I can hear more bass in sample A. Of course, they're very subtle diffenrences. I can feel the same differences (to a greater or lesser degree) listening with my AKG K701 (connected direct to the Tascam or using my headphone preamp), through speakers, pc... Anyway, you have better hearing than me and maybe I'm just being misled, I don't rule out anything

BTW, I tried Saracon converter but...:



 

Hmm... Interesting! 

So, how did the Saracon conversion sound? I use the Saracon software mainly for down-converting 96k/48k 24 bit files to Redbook 44.1k/16 for CD mastering. When I use it for DSD to PCM, it usually sounds better than Audiogate - but not always.

To start off, I'd like to try and explain some of my own thoughts, opinions and experiences on the subject of digital audio and formats.

Again, this is just my humble opinion.

First, PCM vs DSD audio. Yes, I am aware of the technical specifications, strengths and limitations of both sampling methods (S/N vs frequency, noise floor, etc.). And yes, there is compelling research and lengthy articles written on both sides of the PCM vs DSD debate. But to me; to my own ears, DSD sounds more like analog than PCM. Just to put things into perspective, here's a list of gear that I use for all of my critical listening:

Sources:
   • Audio GD NFB-12.1 DAC; Wolson WM8741 chipset, DAC output via RCA's
   • Sony DVP-S9000ES CD/DVD/SACD player; analog output (for SACD) and digital coax output
   • Korg MR-2000S DSD/PCM master recorder; 5.6Mhz DSD ~ 192k/24bit PCM, Cirrus Logic 4398 DAC
   • Revox A77 Mark II Dolby reel-to-reel tape deck; 15ips, Maxell UD/XLII tape
   • Focusrite Liquid 56 firewire interface; Cirrus Logic CS4272 DAC chip, up to 192k/24bit, analog and digital outputs
   • Creative Soundblaster X-Fi Fatality Pro sound card; coax digital & analog output up to 96k/24bit (Wolfson WM8775 chipset)
   • Parasound Halo C2 preamp; AKM 4393 96K/24bit DAC, analog outputs

Amps:
   • Parasound HCA-2205A; 5 channel, 220W high current amp
   • Burson Audio HA-160 headphone amp
   • Audio GD NFB-12.1 (headphone amp)

Monitors:
   • Sennheiser HD 800 headphones with Cardas cable upgrade
   • Sennheiser HD 650 headphones
   • Paradigm Studio 100 v2 full range speakers with a Paradigm Servo 15 subwoofer
   • Paradigm Studio 20 v2 bookshelf speakers
   • Mackie MR5 studio monitors with a Paradigm PDR10 subwoofer

When I evaluate and listen to audio samples, I mainly use my PC; which is: my X-Fi sound card -> Wireworld gold coax cable-> Audio GD NFB-12.1-> Wireworld Eclipse RCA cables-> Burson Audio HDA-160 -> Sennhesier HD 800's. With this setup, the differences between files is very easy to hear since the Sennheiser HD800's tend to magnify the fine details in the music. This setup is limited to 96k/24bit audio, so when I listen to .dsf and other DSD audio formats, I use the Foobar2000 audio player with the DSD plugin, which converts DSD to PCM 96k/24bit. Or, I may copy the .dsf file to my Korg MR-2000S and run the analog outputs to my Burson Audio HA-160 headphone amp along with my HD 800 headphones. In both cases, the DSD audio files will almost always sound better than straight PCM. There are times when I can't tell them apart, but most of the time, it's quite obvious.

It's kind of like the difference between my Revox A77 and any hi-res PCM audio, the A77 always wins hands down. Technically, using specifications alone - high resolution digital audio should sound superior since PCM has better S/N and theoretically better dynamic range, but even with a bit of tape hiss and a slight high frequency roll-off (compared to high resolution PCM), the A77 sounds more real and 'present' than the digital reproduction. In fact, just passing a digital source through the A77 improves the perceived audio quality, IMHO. And when it comes to capturing the sound qualities of my A77, DSD is the only format that comes close to reproducing the realness and presence I hear when I listen to the A77 live.

Also, IMHO, not all DAC chips are created equal; and not all implementations and applications of DAC chips are ideal. For example, the very best PCM digital-to-analog conversion I've heard (from the gear that I own) is not my Audio GD NFB-12.1 DAC, but my Parasound Halo C2. Even though the AKM 4393 DAC chip used in the C2 may not be 'state of the art'; Parasound's implementation of the DAC chip is simply outstanding.

Even though I spent quite a lot of money over the years on audio equipment, most of what I own is still considered 'mid-fi' by audiophile snobs - and I understand and accept that. I freely admit that I don't own the 'absolute best' equipment there is - far from it, but over the years I have heard real high-end gear - stuff that costs as much as a mansion, which actually helped me understand just how high of sound quality you can achieve if you don't care about the price. It also gave me enough perspective to be able to fairly judge everything below that line. But just to be clear, I'm not claiming to be an expert, not at all.

After finding out which the source of the samples, I loaded each sample into Reaper and found something interesting when I compared them under a 'microscope', if you will. If you look at the attached photo, you can see sample A tends to round off some of the peaks at certain places while sample B has thinner, more refined peaks with more 'spikes'. Maybe what I'm hearing is the difference between the smoother waveform of sample A and the jagged waveform of sample B:

[attach=1]

The straight PCM file seems to 'smooth over' the peaks, making everything less detailed and less airy, at least to my ears anyway. The fact that the DSD to PCM file retains the definition of the peaks seems to reinforce the notion that the converter had a higher resolution source file to begin with (the DSD original recording) in which to use as a reference when creating the output file; resolution and detail that seems to be missing from the straight PCM recording. The only way to confirm this would to examine the waveform of the straight DSD recording, which at the moment, is all but impossible since no readily available software (that I'm aware of) that can can edit/view native DSD files.

All of this is just my humble opinion and I don't proclaim to be an expert; I can only describe what I hear. It would be interesting to hear the original DSD (.dsf) file on my own equipment for comparison.

Maybe if I get some free time I'll post some of my own PCM vs DSD recordings so we can all compare. In the meantime, if you want to play back DSD recordings on your own computer, I highly recommend Foobar2000 and the DSD plugin.

In the meantime, here's an interesting comparison of PCM and DSD by Ayre where you can download high resolution PCM and DSD samples and listen to the differences yourself. You'll need playback software like Foobar2000 and the DSD plugin to listen to the DSD files. To me ears, the difference is quite obvious, but I'll leave it up to you to pick which one you prefer.

First of all, thank you very much for your detailed reply 

All of this is just my humble opinion and I don't proclaim to be an expert; I can only describe what I hear

From my point of view you are an expert (valid for Luca too!). It's funny, you know...many people call me expert, but I don't think I have experienced half of what you have, so go figure I think it's a matter of different points of view...I understand that a 20 year old beginner could consider me an expert so I guess you'll understand if I call you expert 

And when it comes to capturing the sound qualities of my A77, DSD is the only format that comes close to reproducing the realness and presence I hear when I listen to the A77 live.

You nailed it, that is the feeling I have when it comes to transferring the "goodness of vinyl" to digital with my Tascam: PCM sounds great but DSD gets closer to the real thing

OK, here we go, I've just recorded these files. It would be great if you could analyze them "under microscope" and let us know what you think:

1-Straight-dsdiff.dff
2-Converted-to-96-24-PCM-AudioGate-TPDF-Dith.wav
3-Converted-to-96-24-PCM-Saracon-TPDF-Dith.wav
4-Straight-PCM.wav

This is fun and very very interesting 

PD: The cart I used for these samples is the Ortofon 2M with an OM40 stylus (yeah, I weird looking combination, LOL!). The record is the MFSL edition of Aja, yes BTW, I feel the right channel is a bit dominant, I dunno why...it may be the record 

Very interesting. 

One of the first things I noticed about the PCM files is how much louder they are (and a higher noise floor) compared to the DSD source. Are you normalizing or adding gain while converting? I know Saracon's default for DSD conversion is to add 6dB to the output file, but I always use '0' gain when converting DSD to PCM. Have a listen to what that sounds like:

DSD to PCM 96/24 - Saracon - TPDF - 0 Gain:
Straight-dsdiff-d2p.wav

Adding gain or normalizing the output usually adds a digital 'glare' and harshness to the audio, so I never use it. 

I'm still auditioning and examining the files - I'll write up a more thorough reply when I'm done.

Just for comparison, here's a 96k/24bit sample recording I did of the same song (MFSL Aja) with my Apogee Rosetta 96K AD converter (a year or so ago), using my Rega P25 turntable, Zyx Omega G cartridge and Audio Research PH5 preamp.

BC-PCM_96k24b-Rega-Zyx-Omega.wav

Let me know what you think.

As soon as the new loudspeakers I got will have a proper break-in period (and after replacing caps in the crossover and the speaker cables), I will listen to the various samples with my Blu-ray player, which should be able to play DSD. Then, I'll tell you my opinion.

Brian, you got PM.