Testing DVD/Blu-ray players as CD transports using a null test

[Blew]

After reading many of @NTTY's excellent CD transport reviews, I was curious to find out whether either of the two DVD/BD players that I currently use as CD transports were actually outputting CD audio correctly. Both players output CD audio at 44.1KHz, which I verified by connecting a Topping DAC via optical. The two DVD drives are a Pioneer DV-686A-s and a Samsung BD-H6500. More details of the systems they're used in are in my signature below.

I performed a crude null test on the players by performing the following:

1. Fed the players digital out via Toslink optical to a Hifime UR23 optical to USB converter plugged into a Raspberry Pi 4.
2. Played the last track of a CD and recorded it with arecord to a wav file in 44.1KHz 16 bit 2 channel LE.
3. Copied the wav to my laptop and loaded it into Audacity.
4. Loaded my own FLAC rip of the CD track into Audacity, which I'd verified as accurate against the AccurateRip database.
5. Trimmed the silence from both files.
6. Aligned the audio samples precisely.
7. Inverted the recorded wav file.
8. Mixed them together.

The mix from the Pioneer recording came out as perfectly silent, suggesting that the Pioneer plays CD audio correctly. However, the mix from the Samsung recording came out with low level noise. I played the noisy mix back by increasing the gain by 30dB and I could hear the track's music playing as if it was on a quiet radio in the next room. So, it wasn't just random noise, it was a heavily attenuated version of the same track!

Can anyone explain what the Samsung is doing to cause this? Is it resampling audio back and forth between the original 44.1KHz CD audio and 48KHz of DVD audio? Is my testing process flawed in any way?

I have a couple of other Sony Blu-ray players that I could potentially use as CD transports which I would like to similarly test. They are the BDP-S390 and BDP-S5500, but they only have coaxial S/PDIF digital output and I need Toslink optical. As jitter doesn't matter too much, since I feed the CD audio into a MiniDSP 2x4HD which uses ASRC to resample all audio anyway, will a simple coax/toslink converter such as the one below be sufficient for this?

https://www.amazon.com.au/gp/product/B08B89D3T2/ref=ewc_pr_img_2?smid=A1SI5LOR0XI016&th=1

 

[RandomEar]

This could be a simple alignment problem. Or it could be more. I would recommend DeltaWave for a deeper comparison. Also, the Samsung manual mentions it's got a "Digital Output" menu, which affects how DVD and BluRay audio are processed and streamed via the HDMI and optical output. The settings supposedly include PCM or bitstream output and resampling. It doesn't mention CDs but they could be affected, too.

 

[Redacted]

You can also bit-compare a pair of FLAC files in Foobar2000....
Sometimes it shows different offsets or other padding to be present, but the actual audio content may be the same.

 

[staticV3]

use Deltawave for performing useful Null tests.

it will correct alignment, clock drift, and amplitude automatically to show the residuals that you actually care about, while still displaying the correction values should those be of interest.

 

[Blew]

This could be a simple alignment problem. Or it could be more. I would recommend DeltaWave for a deeper comparison.

Thanks. When I compared the tracks in Audacity I aligned the track time to the exact sample by zooming right in to the samples.

I ran a comparison in DeltaWave of both the Pioneer and the Samsung recordings with the CD rip. DeltaWave reported that the Pioneer recording was 100% accurate, but the Samsung recording was not. So that's in line with what I saw when comparing in Audacity. I've pasted the Samsung comparison below.

Spoiler: Samsung comparison in DeltaWave

DeltaWave v2.0.25, 2026-07-08T20:41:08.1594964+10:00
Reference: 18 Baby Don't Cry.wav[L] 11198492 samples 44100Hz 16bits, stereo, MD5=00
Comparison: samsung-babydontcry.wav[L] 25475573 samples 44100Hz 16bits, stereo, MD5=00
Settings:
Gain:True, Remove DC:True
Non-linear Gain EQ:True Non-linear Phase EQ: True
EQ FFT Size:65536, EQ Frequency Cut: 0Hz - 0Hz, EQ Threshold: -500dB
Correct Non-linearity: False
Correct Drift:True, Precision:30, Subsample Align:True
Non-Linear drift Correction:False
Upsample:False, Window:Kaiser
Spectrum Window:Kaiser, Spectrum Size:32768
Spectrogram Window:Hann, Spectrogram Size:4096, Spectrogram Steps:2048
Filter Type:IIR, window:Kaiser, taps:262144, minimum phase=False
Dither:False bits=0
Trim Silence:True
Enable Simple Waveform Measurement: False
Correct for MP DC Filter = False, DC Filter frequency=0.5 Hz

Discarding Reference: Start=0s, End=0s
Discarding Comparison: Start=0s, End=0s

Initial peak values Reference: -0.311dB Comparison: -0.312dB
Initial RMS values Reference: -12.361dB Comparison: -15.931dB

Null Depth=9.956dB
X-Correlation offset: -197899 samples
Drift computation quality, #1: Excellent (0μs)


Trimmed 0 samples ( 0.00ms) front, 0 samples ( 0.00ms end)


Final peak values Reference: -0.311dB Comparison: -0.311dB
Final RMS values Reference: -12.313dB Comparison: -12.313dB

GS: -4487.505669 ms, -0.001 dB (L), -101.0 dBFS (L), [-102.4dBFSA (L)]

Gain= -0.001dB (0.9999x) DC=0.00002 Phase offset=-4487.505669ms (-197899 samples)
Difference (rms) = -101.01dB [-102.36dBA]
Correlated Null Depth=117.07dB [105.93dBA]
Clock drift: 0 ppm


Files are NOT a bit-perfect match (match=47.13%) at 16 bits
Files are NOT a bit-perfect match (match=0.19%) at 24 bits
Files match @ 49.9131% when reduced to 15.99 bits


---- Phase difference (full bandwidth): 0.0196723667686429°
0-10kHz: 0.01°
0-20kHz: 0.02°
0-24kHz: 0.02°
Phase EQ not computed
Timing error (rms jitter): 1.3ns
PK Metric (step=400ms, overlap=50%):
RMS=-106.2dBFS
Median=-110.0
Max=-91.2

99%: -92.5
75%: -109.25
50%: -109.97
25%: -110.65
1%: -112.27

gn=1.00012001512478, dc=1.52109580692905E-05, dr=0, of=-197899

DONE!

Signature: 13f2b1621758b9e3ba69c19f47c0c004

RMS of the difference of spectra: NaNdB
DF Metric (step=400ms, overlap=0%):
Median=-91.9dB
Max=-73.5dB Min=-94.5dB

1% > -94.14dB
10% > -93.33dB
25% > -92.63dB
50% > -91.86dB
75% > -90.96dB
90% > -89.73dB
99% > -75.75dB

Linearity 25.7bits @ 0.5dB error

Also, the Samsung manual mentions it's got a "Digital Output" menu, which affects how DVD and BluRay audio are processed and streamed via the HDMI and optical output. The settings supposedly include PCM or bitstream output and resampling. It doesn't mention CDs but they could be affected, too.

I have it set to PCM, which I assumed was the same as bitstream when playing CD audio while also decoding DVD and Blu-ray audio. Perhaps PCM routes all audio through a resampler on its way out? I'll try bitstream and see if there is any difference.

I'm still curious about how the Samsung's CD/PCM processing causes the very faint track audio to appear after a null comparison in Audacity!

 

[Blew]

I changed the digital output to bitstream. Also set the audio delay to 0, just in case that affects anything. Unfortunately, it made no difference. The Samsung BD-H6500 appears to always manipulate CD audio on the digital output regardless of settings.

The DeltaWave comparison is below:

Spoiler: Samsung second comparison

DeltaWave v2.0.25, 2026-07-09T00:01:59.7409313+10:00
Reference: 18 Baby Don't Cry.wav[L] 11198492 samples 44100Hz 16bits, stereo, MD5=00
Comparison: samsung-bitstream-babydontcry-trimmed.wav[L] 11166892 samples 44100Hz 16bits, stereo, MD5=00
Settings:
Gain:True, Remove DC:True
Non-linear Gain EQ:True Non-linear Phase EQ: True
EQ FFT Size:65536, EQ Frequency Cut: 0Hz - 0Hz, EQ Threshold: -500dB
Correct Non-linearity: False
Correct Drift:True, Precision:30, Subsample Align:True
Non-Linear drift Correction:False
Upsample:False, Window:Kaiser
Spectrum Window:Kaiser, Spectrum Size:32768
Spectrogram Window:Hann, Spectrogram Size:4096, Spectrogram Steps:2048
Filter Type:IIR, window:Kaiser, taps:262144, minimum phase=False
Dither:False bits=0
Trim Silence:True
Enable Simple Waveform Measurement: False
Correct for MP DC Filter = False, DC Filter frequency=0.5 Hz

Discarding Reference: Start=0s, End=0s
Discarding Comparison: Start=0s, End=0s

Initial peak values Reference: -0.311dB Comparison: -0.312dB
Initial RMS values Reference: -12.361dB Comparison: -12.35dB

Null Depth=10.134dB
X-Correlation offset: 9550 samples
Drift computation quality, #1: Excellent (0μs)


Trimmed 0 samples ( 0.00ms) front, 0 samples ( 0.00ms end)


Final peak values Reference: -0.311dB Comparison: -0.311dB
Final RMS values Reference: -12.338dB Comparison: -12.338dB

GS: 216.553288 ms, -0.0011 dB (L), -101.0 dBFS (L), [-102.4dBFSA (L)]

Gain= -0.0011dB (0.9999x) DC=0.00002 Phase offset=216.553288ms (9550 samples)
Difference (rms) = -101.02dB [-102.37dBA]
Correlated Null Depth=106.87dB [105.72dBA]
Clock drift: 0 ppm


Files are NOT a bit-perfect match (match=47.23%) at 16 bits
Files are NOT a bit-perfect match (match=0.19%) at 24 bits
Files match @ 49.9623% when reduced to 15.99 bits


---- Phase difference (full bandwidth): 0.0208348733871072°
0-10kHz: 0.01°
0-20kHz: 0.02°
0-24kHz: 0.02°
Phase EQ not computed
Timing error (rms jitter): 1.5ns
PK Metric (step=400ms, overlap=50%):
RMS=-106.5dBFS
Median=-110.0
Max=-91.7

99%: -92.58
75%: -109.27
50%: -109.95
25%: -110.62
1%: -112.13

gn=1.00012228921263, dc=1.52545190429251E-05, dr=0, of=9550

DONE!

Signature: c737e38dc1e0a995d4631dce9e6e6546

RMS of the difference of spectra: NaNdB
DF Metric (step=400ms, overlap=0%):
Median=-91.8dB
Max=-73.9dB Min=-94.5dB

1% > -94.15dB
10% > -93.33dB
25% > -92.61dB
50% > -91.82dB
75% > -90.88dB
90% > -89.64dB
99% > -75.47dB

Linearity 25.7bits @ 0.5dB error

 

[Blew]

I purchased the PORTTA Toslink/Coaxial adapter mentioned above. It arrived today, so I was able to test the two Sony Blu-ray players (BDP-S390 and BDP-S5500) using DeltaWave. Both were perfectly bit accurate, despite being modern SoC based players with HDMI. It's good to know that not only are the Sony players bit-accurate but that the Toslink/Coax adapter has no issues either.

I also used DeltaWave to test a JVC XV-Y360A DVD player I have lying around. It actually has a front display with track information and physical play/pause/skip buttons, which is a rarity these days. Sadly, it was not bit accurate. I performed the same null test comparison in Audacity that I did with the Samsung BD-H6500 above, where I could see low level noise left that was not completely cancelled out. However, when I played it back at +36dB I heard faint crackling noise, which is different to the faint music that was left over from the Samsung.

DeltaWave output is below, if anyone is interested.

Spoiler: Sony BDP-S390 DeltaWave comparison

DeltaWave v2.0.25, 2026-07-09T21:51:10.1896928+10:00
Reference: 18 Baby Don't Cry.wav[L] 11198492 samples 44100Hz 16bits, stereo, MD5=00
Comparison: sony-s390-babydontcry.wav[L] 11814359 samples 44100Hz 16bits, stereo, MD5=00
Settings:
Gain:True, Remove DC:True
Non-linear Gain EQ:True Non-linear Phase EQ: True
EQ FFT Size:65536, EQ Frequency Cut: 0Hz - 0Hz, EQ Threshold: -500dB
Correct Non-linearity: False
Correct Drift:True, Precision:30, Subsample Align:True
Non-Linear drift Correction:False
Upsample:False, Window:Kaiser
Spectrum Window:Kaiser, Spectrum Size:32768
Spectrogram Window:Hann, Spectrogram Size:4096, Spectrogram Steps:2048
Filter Type:IIR, window:Kaiser, taps:262144, minimum phase=False
Dither:False bits=0
Trim Silence:True
Enable Simple Waveform Measurement: False
Correct for MP DC Filter = False, DC Filter frequency=0.5 Hz

Discarding Reference: Start=0s, End=0s
Discarding Comparison: Start=0s, End=0s

Initial peak values Reference: -0.311dB Comparison: -0.311dB
Initial RMS values Reference: -12.361dB Comparison: -12.593dB

Null Depth=10.084dB
X-Correlation offset: -100256 samples
Drift computation quality, #1: Excellent (0μs)


Trimmed 0 samples ( 0.00ms) front, 0 samples ( 0.00ms end)


Final peak values Reference: -0.311dB Comparison: -0.311dB
Final RMS values Reference: -12.337dB Comparison: -12.337dB

GS: -2273.378685 ms, 0 dB (L), -224.3 dBFS (L), [-226.4dBFSA (L)]

Gain= 0dB (1x) DC=0 Phase offset=-2273.378685ms (-100256 samples)
Difference (rms) = -224.25dB [-226.38dBA]
Correlated Null Depth=227.87dB [227.46dBA]
Clock drift: 0 ppm


Files are a bit-perfect match at 16 bits
Files are NOT a bit-perfect match (match=99.99%) at 24 bits


---- Phase difference (full bandwidth): 1.94617580519077E-09°
0-10kHz: 0.00°
0-20kHz: 0.00°
0-24kHz: 0.00°
Phase EQ not computed
Timing error (rms jitter): 0ps
PK Metric (step=400ms, overlap=50%):
RMS=-225.0dBFS
Median=-225.5
Max=-218.1

99%: -220.58
75%: -223.73
50%: -225.48
25%: -227.22
1%: -241.22

gn=1.00000000000009, dc=0, dr=0, of=-100256

DONE!

Signature: 646b7ef2e7b76b791f101fc3d05f9b24

RMS of the difference of spectra: NaNdB
DF Metric (step=400ms, overlap=0%):
Median=-300dB
Max=-138.2dB Min=-300dB

1% > -300.0dB
10% > -300.0dB
25% > -300.0dB
50% > -300.0dB
75% > -145.47dB
90% > -142.66dB
99% > -139.71dB

Linearity 32.0bits @ 0.5dB error

Spoiler: Sony BDP-S5500 DeltaWave comparison

DeltaWave v2.0.25, 2026-07-09T22:27:58.3391543+10:00
Reference: 18 Baby Don't Cry.wav[L] 11198492 samples 44100Hz 16bits, stereo, MD5=00
Comparison: sony-s5500-babydontcry.wav[L] 13969942 samples 44100Hz 16bits, stereo, MD5=00
Settings:
Gain:True, Remove DC:True
Non-linear Gain EQ:True Non-linear Phase EQ: True
EQ FFT Size:65536, EQ Frequency Cut: 0Hz - 0Hz, EQ Threshold: -500dB
Correct Non-linearity: False
Correct Drift:True, Precision:30, Subsample Align:True
Non-Linear drift Correction:False
Upsample:False, Window:Kaiser
Spectrum Window:Kaiser, Spectrum Size:32768
Spectrogram Window:Hann, Spectrogram Size:4096, Spectrogram Steps:2048
Filter Type:IIR, window:Kaiser, taps:262144, minimum phase=False
Dither:False bits=0
Trim Silence:True
Enable Simple Waveform Measurement: False
Correct for MP DC Filter = False, DC Filter frequency=0.5 Hz

Discarding Reference: Start=0s, End=0s
Discarding Comparison: Start=0s, End=0s

Initial peak values Reference: -0.311dB Comparison: -0.311dB
Initial RMS values Reference: -12.361dB Comparison: -13.321dB

Null Depth=10.064dB
X-Correlation offset: -116713 samples
Drift computation quality, #1: Excellent (0μs)


Trimmed 0 samples ( 0.00ms) front, 0 samples ( 0.00ms end)


Final peak values Reference: -0.311dB Comparison: -0.311dB
Final RMS values Reference: -12.337dB Comparison: -12.337dB

GS: -2646.553288 ms, 0 dB (L), -224.3 dBFS (L), [-226.4dBFSA (L)]

Gain= 0dB (1x) DC=0 Phase offset=-2646.553288ms (-116713 samples)
Difference (rms) = -224.25dB [-226.38dBA]
Correlated Null Depth=227.87dB [227.46dBA]
Clock drift: 0 ppm


Files are a bit-perfect match at 16 bits
Files are NOT a bit-perfect match (match=99.99%) at 24 bits


---- Phase difference (full bandwidth): 1.94617580519077E-09°
0-10kHz: 0.00°
0-20kHz: 0.00°
0-24kHz: 0.00°
Phase EQ not computed
Timing error (rms jitter): 0ps
PK Metric (step=400ms, overlap=50%):
RMS=-225.0dBFS
Median=-225.5
Max=-218.1

99%: -220.58
75%: -223.73
50%: -225.48
25%: -227.22
1%: -241.22

gn=1.00000000000009, dc=0, dr=0, of=-116713

DONE!

Signature: d253c2bfcfbe7ffd1dbd3cc0238931be

RMS of the difference of spectra: NaNdB
DF Metric (step=400ms, overlap=0%):
Median=-300dB
Max=-138.2dB Min=-300dB

1% > -300.0dB
10% > -300.0dB
25% > -300.0dB
50% > -300.0dB
75% > -145.47dB
90% > -142.66dB
99% > -139.71dB

Linearity 32.0bits @ 0.5dB error

Spoiler: JVC XV-Y360A DeltaWave comparison

DeltaWave v2.0.25, 2026-07-09T22:45:59.9679678+10:00
Reference: 18 Baby Don't Cry.wav[L] 11198492 samples 44100Hz 16bits, stereo, MD5=00
Comparison: jvc-babydontcry.wav[L] 17641600 samples 44100Hz 16bits, stereo, MD5=00
Settings:
Gain:True, Remove DC:True
Non-linear Gain EQ:True Non-linear Phase EQ: True
EQ FFT Size:65536, EQ Frequency Cut: 0Hz - 0Hz, EQ Threshold: -500dB
Correct Non-linearity: False
Correct Drift:True, Precision:30, Subsample Align:True
Non-Linear drift Correction:False
Upsample:False, Window:Kaiser
Spectrum Window:Kaiser, Spectrum Size:32768
Spectrogram Window:Hann, Spectrogram Size:4096, Spectrogram Steps:2048
Filter Type:IIR, window:Kaiser, taps:262144, minimum phase=False
Dither:False bits=0
Trim Silence:True
Enable Simple Waveform Measurement: False
Correct for MP DC Filter = False, DC Filter frequency=0.5 Hz

Discarding Reference: Start=0s, End=0s
Discarding Comparison: Start=0s, End=0s

Initial peak values Reference: -0.311dB Comparison: -0.311dB
Initial RMS values Reference: -12.361dB Comparison: -14.334dB

Null Depth=9.917dB
X-Correlation offset: -192629 samples
Drift computation quality, #1: Excellent (0μs)


Trimmed 0 samples ( 0.00ms) front, 0 samples ( 0.00ms end)


Final peak values Reference: -0.311dB Comparison: -0.311dB
Final RMS values Reference: -12.328dB Comparison: -12.328dB

GS: -4368.004535 ms, 0 dB (L), -107.7 dBFS (L), [-108.5dBFSA (L)]

Gain= 0dB (1x) DC=0 Phase offset=-4368.004535ms (-192629 samples)
Difference (rms) = -107.73dB [-108.49dBA]
Correlated Null Depth=120.06dB [116.6dBA]
Clock drift: 0 ppm


Files are NOT a bit-perfect match (match=87.23%) at 16 bits
Files are NOT a bit-perfect match (match=1.38%) at 24 bits
Files match @ 87.2259% when reduced to 16 bits


---- Phase difference (full bandwidth): 0.00499586396279604°
0-10kHz: 0.00°
0-20kHz: 0.01°
0-24kHz: 0.00°
Phase EQ not computed
Timing error (rms jitter): 315.6ps
PK Metric (step=400ms, overlap=50%):
RMS=-109.8dBFS
Median=-110.6
Max=-102.8

99%: -104.39
75%: -109.44
50%: -110.65
25%: -111.83
1%: -120.77

gn=1.00000446702057, dc=6.07501820304306E-07, dr=0, of=-192629

DONE!

Signature: ecae0c9525e7cc9b260b43cb5a8bcd41

RMS of the difference of spectra: NaNdB
DF Metric (step=400ms, overlap=0%):
Median=-98.7dB
Max=-96.9dB Min=-121.4dB

1% > -120.47dB
10% > -102.43dB
25% > -100.08dB
50% > -98.67dB
75% > -97.91dB
90% > -97.49dB
99% > -97.12dB

Linearity 26.8bits @ 0.5dB error

In summary, the bit-accurate DVD/Blu-ray players were:
Pioneer DV-686A-s
Sony BDP-S390
Sony BDP-S5500

The DVD/Blu-ray players not bit accurate were:
Samsung BD-H6500
JVC XV-Y360A