Pioneer DV-717 (CD/DVD) Measurements

[ReDFoX]

Hello! This is a quick review/measurements set of a Pioneer DV-717 CD/DVD player, tested in CD mode.

I was loaned this unit just for a few days, so I took an opportunity to measure such a high-end (for early 2000s) device.

Well, at least it looks cool. Released in 1998 it had a price of ~1550$ in 1999 (adjusted to inflation). Unfortunately, looking at the back of the unit:

we see a disappointingly small variety of outputs: no YPbPr and HDMI (only CVBS and Y/C), no multichannel output and no SACD support (released in 1999). Sooo, what do we have?

Multichannel DVD-A? NO
SACD? NO
24/192? NO
Balanced out? NO
Breakfast in bed? NO (at least mine unit doesn't)

Quite a disappointment, though I'm more interested in CD mode quality, since no one will use DVDs with an analog output in 2025. At least it has dual mono dedicated audio path, though SMPS () can possibly degrade warm analog sound of your spinning disks. This was the moment I felt scammed with this unit for the first time: most of the weight is concluded in the base, not in PCBs/drive unit(

All signals were generated with REW.
DV-717 was fed into line input of my E-MU1616 at 192kHz (E-MU has ultrasonic noise, so I wouldn't show anything above 48k).
Analysis settings, unless said otherwise, are:

• 32k FFT
• 8 averages
• Dolph-Chebyshev 200 window
• 0% overlap

Starting with single tone THD measurement at 1kHz:

Uh-oh... My initial reaction was:

WHAT THE HELL?? I mean, I was so impressed (in a bad way), that I loopbacked my 1616 just to check that my interface works correctly. Look at that mess! That's not a compact-cassette deck, what's up with those sidebands that are almost as high as an initial harmonic distortion signal? I can't even smell 16 bits of resolution, but, yeah, 24/96 DAC!
Btw, note that DV-717 was outputting less then standard -10dBu.

I tried capturing PSU noise, but I'm not shure whether it's a Pioneer problem, or non-existent ground in my flat...

Next up is multitone (FFT length - 512K):

Not terrible, not great either( Note the high noise floor (compared to my Marantz CD-67) and strange bell-shaped noise at ~3kHz (we'll see some anomalies there later).

What about frequency response, the first and most important (unless broken) aspect of any digital device?

It's so bad I had to change vertical graph scale! I think I found the reason people praise DV-717 for its sound.

100% this is audible, since FR drops almost 4dB from 4 to 20kHz. At this point it's an e-waste (at least by my standards), but lets dig deeper!

I combined CCIF IMD test with filter response (255 averages):

Just look at this thing! It's already NOS-like and just needs a few tubes on the output to make it 1% THD for extreme warmth.

Almost no attenuation, thus terrible image problems. With every new measurement I'm uncovering a new low .

I'm glad that Pioneer cares about foxes and bats and allows them to listen to hidden details in your Patricia Barber CDs

I'm gonna speedrun the rest of the measurements, since the aren't as entertaining...

SMPTE IMD test at 256k FFT

shows same strange behaviour around 3kHz, though not awful in general.

J-Test (256k FFT)

is messy, with a lot of PSU noise and sidebands...

Finally, a THD vs frequency measurement! I specifically enabled 2nd&3rd harmonics on the graph

Just as expected, but now we got a 1.7kHz peak for some weird reason.

Manufacturer claims 0.002% THD, but I measured 0.008!

Needless to say, this is a disaster in almost every aspect, from IO to performance... At leas it won an EISA award!

I think a need my own Panther-like ratings... As a canine and fox enjoyer, I'll rate this luxurious garbage can a surprised fox. This fox perfectly recreates my face during entire measurement process.

Not a traumatized fox (yet), because that's not my unit

 

[AnalogSteph]

Hey, at least it has RGB output (on the SCART) if you want that..

Starting with single tone THD measurement at 1kHz:

I smell some clipped ASRC shenanigans, especially since the multitone looks fairly clean. Maybe that's what the "Hi-Bit" chip (PD0236AM) does? Should be bypassable fairly easily if need be.

Odd construction in general. They're using the left channel of one PCM1716 and the right one of another. Was crosstalk that much of an issues with these or what? Or just for bragging rights? ("Look, I'm all dual mono!")

Looks like the unit should be moddable to "IEC Class II with ground" fairly easily (connect PE to C2/C3/C4 junction) - not like it would do much good for you and your ungrounded outlets. (BTW, in theory that means that you shouldn't be operating any IEC Class I devices at your place at all. Their safety relies on a protective earth connection with resistance low enough to blow a fuse or trip a breaker if the live ever were to come into contact with chassis.)

 

[Scytales]

Thanks for this set of measurements, RedFox !

The behavior you discovered has a very simple explanation, I think: this unit must uses a custom interpolation filter Pioneer called "Legato Link". On more recent Pioneer players, this feature can be disable, but on this DV-717, apparently not : https://fc.darty.com/notices/DOCUMENTATION/PIONEER/0751880/0751880_NOTCOMP.pdf

 

[GXAlan]

Yes, that’s Legato Link. As much as we criticize these CD players for not being truthful to source, with leaky filters, DSD ultrasonic noise (not this product but in general), Sony DSEE HX, and Pioneer Legato Link, maybe this is one type of seasoning that was preferred by sufficient customers?

You can read more here:

Pioneer's LegatoLink, Wadia Spline - the secrets about Up- and Oversampling

The ONE place on the net with the most objective sighted reviews.

marlene-d.blogspot.com

 

[Scytales]

Paul Miller has written about the first iteration of the Legato Link Conversion in the issue #112 (page 78 of the pdf) and issue #113 (page 17 of the pdf) of Hifi Choice.

 

[VintageFlanker]

Thanks @ReDFoX !

I had this one in the early days of my audiophile journey. (Paid 100€ or so to get it the time).

 

[ReDFoX]

@AnalogSteph
@GXAlan
@Scytales
Thank you all for Legato Link related explanations!
I was almost 100% sure that it's Legato Link's fault, but for me, as a consumer, it doesn't really matter. My point is, you can get something cheaper and better
Personally, I can't see any reasons to even mod such thing, cause it has 0 unique characteristics that will outweigh inconveniences and time spent during such process... I have Pioneer DV-565A measurements on hands (will publish later) and it's better!
Btw, I'm cooking a 4 HI-FI VHS VCR comparison, should be quite interesting

 

[NTTY]

Analysis settings, unless said otherwise, are:

• 32k FFT
• 8 averages
• Dolph-Chebyshev 200 window
• 0% overlap

Starting with single tone THD measurement at 1kHz:

View attachment 442940

32k FFT with 192kHz input is way too small. I’d recommend 256k minimum, I personally go 512k.

Also, I see the span is up to 91kHz in that view so that does not give a valid reading of THD+N, nor ENOB, of course.

You said you created all test files with REW. Did you use dither with the 1kHz test tone (16bits)? If yes, THD+N will max at -90dB, if I remember. With rectangular dither of Audacity, only 3dB resolution are lost. With 999.91Hz test tone, no need for dither, and that allows to see the max resolution of the player.

Last: the pattern of distortion looks like the issue some of us had when burning a CD with WMP and with the « equalized levels » ON. Maybe you want to double-check that too. But since I don’t see issues on your other measurements, I guess it might be the conversion of that player.

Next up is multitone (FFT length - 512K):

View attachment 442945

This one looks a little noisy (?) but the dBFS scale does not help, dBr would be better.

I combined CCIF IMD test with filter response (255 averages):

View attachment 442950

Not NOS but a slow type filter indeed.

.

J-Test (256k FFT)

View attachment 442953

This one requires linear frequency scale for better analysis. I think it suffers from the REW issue when creating 16bits jtest file. I think the side bands around the fundamental are mainly because of REW issue. You can use the beta version of REW to re-create the test file, and which fixes this issue.

Cheers

 

[ReDFoX]

32k FFT with 192kHz input is way too small. I’d recommend 256k minimum, I personally go 512k.

Also, I see the span is up to 91kHz in that view so that does not give a valid reading of THD+N, nor ENOB, of course.

You said you created all test files with REW. Did you use dither with the 1kHz test tone (16bits)? If yes, THD+N will max at -90dB, if I remember. With rectangular dither of Audacity, only 3dB resolution are lost. With 999.91Hz test tone, no need for dither, and that allows to see the max resolution of the player.

Last: the pattern of distortion looks like the issue some of us had when burning a CD with WMP and with the « equalized levels » ON. Maybe you want to double-check that too. But since I don’t see issues on your other measurements, I guess it might be the conversion of that player.




This one looks a little noisy (?) but the dBFS scale does not help, dBr would be better.


Not NOS but a slow type filter indeed.



This one requires linear frequency scale for better analysis. I think it suffers from the REW issue when creating 16bits jtest file. I think the side bands around the fundamental are mainly because of REW issue. You can use the beta version of REW to re-create the test file, and which fixes this issue.

Cheers

Oh, why on Earth my THD measurement was set to 91kHz? I can't remember setting up such a strange number...
Imma just use your test CD next time

 

[NTTY]

Oh, why on Earth my THD measurement was set to 91kHz? I can't remember setting up such a strange number...
Imma just use your test CD next time

High pass (20Hz) and low pass (20kHz) filters were probably unselected in the distortion settings of RTA

 

[ReDFoX]

High pass (20Hz) and low pass (20kHz) filters were probably unselected in the distortion settings of RTA

Yes, but I can't imagine why... I haven't touched distortion analyzer settings)
Anyway, that's a good catch, thank you! Strange that no one else pointed that out