Hello Everyone,
This is a review and detailed measurements of the Sony CDP-X333ES CD Player and Transport.
Sony CDP-X333ES - Presentation
I got interested by this player because it uses the first 1bit DAC from Sony (CXD2552), together with their oversampling filter CXD1244, successor of the CXD1144. The newer version included a Noise shaper to go with the 1bit conversion, that is to reduce quantization errors due to the 1bit decimation. As a matter of facts, the X333ES had to bigger brothers (or sisters depending how you see it), the X555ES and the X777ES which shared the exact same conversion. They were the second iteration of Sony to host this 1bit conversion (after the CDP-X33/55/77ES).
By the way, checking the CXD2552 datasheet, I saw it was including Noise shaping processing too. Which one of the two is active in this player between the CXD1244 or the CXD2552, I can't tell.
This is a beautiful player, me thinks, and I would have not guessed it's nearly 35 years old. Yes, it was released in 1991, and is the lightest (11kg "only"), cheapest of the "ES" range. Whatever ES means, nobody is sure, says the web, but at the end of this presentation, I think you'll agree with me that "Excellence Series" suits it well.
The front face is busy with the nice pad to jump directly from track 1 to 20, I love it, so practical. The drive is faster than light, as usual with the magnetic KSS laser heads, it's addictive.
The back is more classical:
Fixed and Variable RCAs, very common at the time, and only one optical SPDIF output.
The inside shows nothing fancy, but it's still nice to look at:
We see only one power transformer, but some (early?) models have two, one for digital and analog sections.
The servo board is on top of the conversion board. The latter offers full dual mono from after the oversampling filter, which means there are two CXD2552 DACs, of course:
KSS-271A head (theoretically only found in the higher x555ES version, but hey, here you go):
The drawer is much slower to open and close than the previously reviewed Sony CDP-557ESD, but the mech itself is equality fast to skip several tracks, FFW, etc...
I've read so many (random) comments about the sound quality of the Sony 1bit DAC. The review of the X555ES (with the same conversion) that you can find on Stereophile is epic in that way, talking about the absence of rendering space, quite bright sterilized music without reverb and with neutered front to back (squashed depth), being no more than a $300 sounding machine (at best) with a phasing anomaly in the mids. Funny, I say.
I certainly had to make an opinion for myself, by measuring it, as I should find so much defects never seen before. And so, of course, I was tempted
Sony CDP-X333ES - Measurements (Analog out)
All measurements performed with a Cosmos E1AD (grade G) and the Cosmos Scaler (100kohms from unbalanced input).
I am now consistent with my specific measurements for CD Players, as I described them in the post “More than we hear”, and as I reported them for the Onkyo C-733 review. Over time, this will help comparing the devices I reviewed.
The Sony outputs 1dB more than the standard 2Vrsm form the fixed RCA outputs. The two channels were very well balanced (less than 0.02dB), which is very good. The single-ended outputs are non-inverting.
----
As usual, let's start with my standard 999.91Hz sine @0dBFS (without dither) from the Test CD (RCA out):
This is a reasonable amount of distorsion for the time. The right channel has more H3 but same THD overall. SNR is good (in presence of signal) and at the limit of the CD audio format for a full scale signal. Very good.
Next is the same view but at -6dBFS, as I usually show:
Even better as the signal goes down. The distorsion components are very limited. Note that the sine is at the exact 999.91Hz, which means we have a precise clock here, and I'll be back to that later in the review. Last note, the dashboard above shows the right channel results, left one has 10dB less distorsion.
I think you already spotted that this is a silent player, but let's have a closer look:
Sorry, the print screen says linear scale, but wrong, it's logarithmic. Anyways, nothing to see here, despite the FFT length at 512k (capture is done at 192kHz sampling rate) which lowers a lot the random noise in the view. So despite only one power transformer, this is a nice silent CD player which did not need two
----
Next in the list is the bandwidth:
You can see the two channels matching at 0.02dB, and this is very flat (-0.2dB at worst). I’d love to that precision more often.
And since this is the job of the oversampling filter, let's have a look at the rest of what it does:
You can appreciate the filter response with very limited increase of noise beyond 30kHz which is due to the noise shaping. The attenuation of the two tones 18kHz + 20kHz is more than 90dB, which was very good for the time, much better than Philips chips. It was the SOTA filter of the time. If you're not convinced, have a look at the much more recent Myryad Z210 CD Player I recently reviewed.
----
Let's have a look at the multitone test that a lot of you like very much:
This is as good as it can get, and that is 18bits+ free of distorsion. It's not so often I see that, especially for a very old CD Player.
----
Let's move on to the jitter test:
Nailed! This is again a very beautiful trace, especially for the era. The red trace is the WAV file analyzed by the computer. The blue trace is the Sony. Not lateral spikes, that means no jitter.
----
Started with the Teac VRDS-20 review, and on your request + support to get it done (more here), I'm adding now an "intersample-overs" test which intends to identify the behavior of the digital filtering and DAC when it come to process near clipping signals. Because of the oversampling, there might be interpolated data that go above 0dBFS and would saturate (clip) the DAC and therefore the output. And this effect shows through distorsion (THD+N measurement up to 96kHz):
I kept some references and will keep the same for other reviews, so you can quickly compare. The results of the Sony CDP-X333ES means that it has no headroom in the oversampling filter. In fact, it starts saturating as soon as with +0.01dB overs. It's not "hot Masters" friendly.
----
Let's continue with the good old 3DC measurement that Stereophile was often using as a proof of low noise DAC. It is from an undithered 997Hz sine at -90.31dBFS. With 16bits, the signal should appear (on a scope) as the 3DC levels of the smallest symmetrical sign magnitude digital signal:
That is a good trace, even if the Sony CDP-557ESD with R2R conversion did better. On top of that, I must mention the linearity which is very good for the time, again. For instance, this is a -100dBFS sine (with rectangular dither) that is perfectly represented:
The total absence of distorsion spikes is to be noted.
----
Other measurements (not shown):
The crosstalk is very good because the DAC not receiving any signal shuts down. I need to update my test file to prevent that. This is also how Sony could justify an SNR of -118dB, which is indeed precisely what I measured when no signal goes into the DAC. But that is a trick, of course. All that said, it means that 35 years later, it performs as out of the factory...
The pitch error test shows a very precise clock with only 2.5ppm deviation, exceptional for the time and the age, and many times better than what our ears need.
----
You might know that it is now possible to go below the theoretical resolution of 16bits of the CD Audio, using a technique that is called noise shaping. This a mathematical process to reduce the quantization errors due to the format (only 16bits) and spread these errors in a smart manner so the actual resolution can be increased where it matters for our ears. The below is an FFT of the same 999.91Hz Sine test tone at full scale (0dBFS), with shape dither. To process and calculate what the player is capable in terms of resolution, I reduce the span to 20Hz - 6000Hz, instead of the standard 20Hz - 20kHz, to prevent adding the noise generated at the higher frequencies. I added the result with the Denon DCD-900NE as a reference:
Reducing the span allows the software to compute a theoretical ENOB, which is very real to our ears. With shape dither, the Denon can achieve 18.2bits of virtual resolution, while we see the true limits of the Sony CDP-X333ES at 17.1bits. The higher noise floor of the Sony is probably due to the noise shaper which might add some digital noise because its oversampling filter is an 18bits one, but also maybe because it lacks power and/or the noise shaper is less agressive than the one of the Denon. That I don't know. But this is a very good result, also because the trace is very neat
----
Last and not least, I like to run a THD vs Frequency sweep at -12dBFS as it shows how the conversion has evolved over time. I am currently using the beta version of REW and I discovered that this sweep gives better and more reliable results than before. So I decided to measure a number of CD players I still have on the bench and overlay their results. The CD players are:
As you can see, the Sony X333ES is in par with the much more recent Denon DCD-900NE and its ESS conversion. It shows the advantage of the 1bit and delta sigma conversion over the best R2R converters of the past (the Denon DCD-3560 with its dual BurrBrown PCM58K in pseudo co-linear mode exhibits a crazy good result here, for an R2R conversion). The worst trace is from the first Philips 14bits conversion (in the Marantz CD-73), hence the highest distorsion.
Sony CDP-X333ES - Measurements (Optical Out)
I've seen several of you reviewing CD players using their digital outputs, in case the results could be improved from an external DAC.
This Sony suffered an unusual issue, and I think the cause is the emitter. Have a look at the 999.91Hz @-0dBFS:
Yes, full of noise. And too bad because the clock is precise and so it's easy to have a good trace. My testing interface is not decoding/correcting anything, just taking the digital input as it comes. And it shows an issue. But when connecting to a standard DAC (an SMSL PS200 in my case), the receiver of the DAC can manage to correct the errors, as you can see with the same sine:
The result is nearly perfect, very close to the original WAV file.
I thought about a way to test the data stream input, without making it a nightmare. So I simply reused my standard 3DC levels test (997Hz @-90.31dBFS) to verify if a 1 would turn out to be a 0 sometimes (or the other way around). Theoretically, and in practice, I should see this:
At this level we must see a scare (lowest symmetrical signal in 16bits). Ringing is due to the Gibbs Phenomenon.
But with the Sony, sometimes, I see that:
Haha, a 1 turned out to be a 0 in that case (0000000000000000 instead of 0000000000000001). So I did not need to deep dive into this, there are indeed some errors in the transmission, probably some aging components (I'll check later, maybe).
But, at the end of the day, it was nice to see that the external DAC had no issue dealing with these errors. When I think some say they can hear a difference between two optical cables
Sony CDP-X333ES - Testing the drive
What would be good measurements if the drive would not properly read a slightly scratched CD, or one that was created at the limits of the norm? The below tests reply to these questions.
The drive was able to consistently continue playing, without generating typical digital clicks, with dropouts of up to 0.75mm. The interpolation effect remained hidden to my ears when it kicked-off but failed to maintain a constant flow beyond 0.75mm as it stopped playing with 1mm dropouts. The Sony had no issue with variable linear velocity and/or track pitch, as well as with HF detection. These results are a little lower than what I measure usually. I guess we have to pay the price for the speed of that drive.
Conclusion
Let's forget the issue with the digital output (which is a not an issue at the output of a DAC), and recognize the quality of this player. Besides to non-resistance to intersample overs, the results are extremely good. There is absolutely nothing bad to mention here. I'm really happy to have had the opportunity to test this first 1bit conversion offering from Sony.
This was SOTA at the time, and is still a very good for a CD player more than 30 years later. I must say I did not expect so good results. There's only the noise floor that limits the real resolution of that player to a still very good 17+bits, which is probably due to the ancient noise shaping (I'll add more information about that later).
This 1bit tech was a preview of the DSD and SACD to come. Sony folks knew what they were doing, no surprise.
Oh yes, it sounds really good to my ears, like all the others
I hope you enjoyed this review.
Have a nice week!
This is a review and detailed measurements of the Sony CDP-X333ES CD Player and Transport.
Sony CDP-X333ES - Presentation
I got interested by this player because it uses the first 1bit DAC from Sony (CXD2552), together with their oversampling filter CXD1244, successor of the CXD1144. The newer version included a Noise shaper to go with the 1bit conversion, that is to reduce quantization errors due to the 1bit decimation. As a matter of facts, the X333ES had to bigger brothers (or sisters depending how you see it), the X555ES and the X777ES which shared the exact same conversion. They were the second iteration of Sony to host this 1bit conversion (after the CDP-X33/55/77ES).
By the way, checking the CXD2552 datasheet, I saw it was including Noise shaping processing too. Which one of the two is active in this player between the CXD1244 or the CXD2552, I can't tell.
This is a beautiful player, me thinks, and I would have not guessed it's nearly 35 years old. Yes, it was released in 1991, and is the lightest (11kg "only"), cheapest of the "ES" range. Whatever ES means, nobody is sure, says the web, but at the end of this presentation, I think you'll agree with me that "Excellence Series" suits it well.
The front face is busy with the nice pad to jump directly from track 1 to 20, I love it, so practical. The drive is faster than light, as usual with the magnetic KSS laser heads, it's addictive.
The back is more classical:
Fixed and Variable RCAs, very common at the time, and only one optical SPDIF output.
The inside shows nothing fancy, but it's still nice to look at:
We see only one power transformer, but some (early?) models have two, one for digital and analog sections.
The servo board is on top of the conversion board. The latter offers full dual mono from after the oversampling filter, which means there are two CXD2552 DACs, of course:
KSS-271A head (theoretically only found in the higher x555ES version, but hey, here you go):
The drawer is much slower to open and close than the previously reviewed Sony CDP-557ESD, but the mech itself is equality fast to skip several tracks, FFW, etc...
I've read so many (random) comments about the sound quality of the Sony 1bit DAC. The review of the X555ES (with the same conversion) that you can find on Stereophile is epic in that way, talking about the absence of rendering space, quite bright sterilized music without reverb and with neutered front to back (squashed depth), being no more than a $300 sounding machine (at best) with a phasing anomaly in the mids. Funny, I say.
I certainly had to make an opinion for myself, by measuring it, as I should find so much defects never seen before. And so, of course, I was tempted
Sony CDP-X333ES - Measurements (Analog out)
All measurements performed with a Cosmos E1AD (grade G) and the Cosmos Scaler (100kohms from unbalanced input).
I am now consistent with my specific measurements for CD Players, as I described them in the post “More than we hear”, and as I reported them for the Onkyo C-733 review. Over time, this will help comparing the devices I reviewed.
The Sony outputs 1dB more than the standard 2Vrsm form the fixed RCA outputs. The two channels were very well balanced (less than 0.02dB), which is very good. The single-ended outputs are non-inverting.
----
As usual, let's start with my standard 999.91Hz sine @0dBFS (without dither) from the Test CD (RCA out):
This is a reasonable amount of distorsion for the time. The right channel has more H3 but same THD overall. SNR is good (in presence of signal) and at the limit of the CD audio format for a full scale signal. Very good.
Next is the same view but at -6dBFS, as I usually show:
Even better as the signal goes down. The distorsion components are very limited. Note that the sine is at the exact 999.91Hz, which means we have a precise clock here, and I'll be back to that later in the review. Last note, the dashboard above shows the right channel results, left one has 10dB less distorsion.
I think you already spotted that this is a silent player, but let's have a closer look:
Sorry, the print screen says linear scale, but wrong, it's logarithmic. Anyways, nothing to see here, despite the FFT length at 512k (capture is done at 192kHz sampling rate) which lowers a lot the random noise in the view. So despite only one power transformer, this is a nice silent CD player which did not need two
----
Next in the list is the bandwidth:
You can see the two channels matching at 0.02dB, and this is very flat (-0.2dB at worst). I’d love to that precision more often.
And since this is the job of the oversampling filter, let's have a look at the rest of what it does:
You can appreciate the filter response with very limited increase of noise beyond 30kHz which is due to the noise shaping. The attenuation of the two tones 18kHz + 20kHz is more than 90dB, which was very good for the time, much better than Philips chips. It was the SOTA filter of the time. If you're not convinced, have a look at the much more recent Myryad Z210 CD Player I recently reviewed.
----
Let's have a look at the multitone test that a lot of you like very much:
This is as good as it can get, and that is 18bits+ free of distorsion. It's not so often I see that, especially for a very old CD Player.
----
Let's move on to the jitter test:
Nailed! This is again a very beautiful trace, especially for the era. The red trace is the WAV file analyzed by the computer. The blue trace is the Sony. Not lateral spikes, that means no jitter.
----
Started with the Teac VRDS-20 review, and on your request + support to get it done (more here), I'm adding now an "intersample-overs" test which intends to identify the behavior of the digital filtering and DAC when it come to process near clipping signals. Because of the oversampling, there might be interpolated data that go above 0dBFS and would saturate (clip) the DAC and therefore the output. And this effect shows through distorsion (THD+N measurement up to 96kHz):
| Intersample-overs tests Bandwidth of the THD+N measurements is 20Hz - 96kHz | 5512.5 Hz sine, Peak = +0.69dBFS | 7350 Hz sine, Peak = +1.25dBFS | 11025 Hz sine, Peak = +3.0dBFS |
| Teac VRDS-20 | -30.7dB | -26.6dB | -17.6dB |
| Yamaha CD-1 | -84.6dB | -84.9dB | -78.1dB |
| Denon DCD-900NE | -34.2dB | -27.1dB | -19.1dB |
| Denon DCD-SA1 | -33.6dB | -27.6dB | -18.3dB |
| Onkyo C-733 | -88.3dB | -40.4dB | -21.2dB |
| Denon DCD-3560 | -30.2dB | -24.7dB | -17.4dB |
| Myryad Z210 | -70.6dB (noise dominated) | -71.1dB (noise dominated) | -29.4dB (H3 dominated) |
| Sony CDP-x333ES | -30.5dB | -24.8dB | -16.3dB |
I kept some references and will keep the same for other reviews, so you can quickly compare. The results of the Sony CDP-X333ES means that it has no headroom in the oversampling filter. In fact, it starts saturating as soon as with +0.01dB overs. It's not "hot Masters" friendly.
----
Let's continue with the good old 3DC measurement that Stereophile was often using as a proof of low noise DAC. It is from an undithered 997Hz sine at -90.31dBFS. With 16bits, the signal should appear (on a scope) as the 3DC levels of the smallest symmetrical sign magnitude digital signal:
That is a good trace, even if the Sony CDP-557ESD with R2R conversion did better. On top of that, I must mention the linearity which is very good for the time, again. For instance, this is a -100dBFS sine (with rectangular dither) that is perfectly represented:
The total absence of distorsion spikes is to be noted.
----
Other measurements (not shown):
- IMD AES-17 DFD "Analog" (18kHz & 20kHz 1:1) : -96.7dB
- IMD AES-17 DFD "Digital" (17'987Hz & 19'997Hz 1:1) : -97.7dB
- IMD AES-17 MD (41Hz & 7993Hz 4:1): -98.2dB
- IMD CCIF (19kHz & 20kHz 1:1) : -97.6dB
- IMD DIN (250Hz & 8kHz 4:1) : -92.9dB
- IMD SMPTE (60Hz & 7kHz 1:4) : -93.3dB
- IMD TDFD Bass (41Hz & 89Hz 1:1) : -116.8dB
- IMD TDFD (13'58Hz & 19841Hz 1:1) : -111.2dB
- Dynamic Range : 96.7dB (without dither @-60dBFS)
- Crosstalk: -125dBr (100Hz), -125dBr (1khz), -122dBr (10kHz)
- Pitch Error : 19'997.05Hz (19'997Hz requested) ie 2.5ppm
- Gapless playback : Yes
The crosstalk is very good because the DAC not receiving any signal shuts down. I need to update my test file to prevent that. This is also how Sony could justify an SNR of -118dB, which is indeed precisely what I measured when no signal goes into the DAC. But that is a trick, of course. All that said, it means that 35 years later, it performs as out of the factory...
The pitch error test shows a very precise clock with only 2.5ppm deviation, exceptional for the time and the age, and many times better than what our ears need.
----
You might know that it is now possible to go below the theoretical resolution of 16bits of the CD Audio, using a technique that is called noise shaping. This a mathematical process to reduce the quantization errors due to the format (only 16bits) and spread these errors in a smart manner so the actual resolution can be increased where it matters for our ears. The below is an FFT of the same 999.91Hz Sine test tone at full scale (0dBFS), with shape dither. To process and calculate what the player is capable in terms of resolution, I reduce the span to 20Hz - 6000Hz, instead of the standard 20Hz - 20kHz, to prevent adding the noise generated at the higher frequencies. I added the result with the Denon DCD-900NE as a reference:
Reducing the span allows the software to compute a theoretical ENOB, which is very real to our ears. With shape dither, the Denon can achieve 18.2bits of virtual resolution, while we see the true limits of the Sony CDP-X333ES at 17.1bits. The higher noise floor of the Sony is probably due to the noise shaper which might add some digital noise because its oversampling filter is an 18bits one, but also maybe because it lacks power and/or the noise shaper is less agressive than the one of the Denon. That I don't know. But this is a very good result, also because the trace is very neat
----
Last and not least, I like to run a THD vs Frequency sweep at -12dBFS as it shows how the conversion has evolved over time. I am currently using the beta version of REW and I discovered that this sweep gives better and more reliable results than before. So I decided to measure a number of CD players I still have on the bench and overlay their results. The CD players are:
- Marantz CD-73
- Sony CDP-337ESD
- Teac VRDS 25x
- Sony CDP-557ESD
- Denon DCD-3560
- Sony X333ES
- Denon DCD-900NE
As you can see, the Sony X333ES is in par with the much more recent Denon DCD-900NE and its ESS conversion. It shows the advantage of the 1bit and delta sigma conversion over the best R2R converters of the past (the Denon DCD-3560 with its dual BurrBrown PCM58K in pseudo co-linear mode exhibits a crazy good result here, for an R2R conversion). The worst trace is from the first Philips 14bits conversion (in the Marantz CD-73), hence the highest distorsion.
Sony CDP-X333ES - Measurements (Optical Out)
I've seen several of you reviewing CD players using their digital outputs, in case the results could be improved from an external DAC.
This Sony suffered an unusual issue, and I think the cause is the emitter. Have a look at the 999.91Hz @-0dBFS:
Yes, full of noise. And too bad because the clock is precise and so it's easy to have a good trace. My testing interface is not decoding/correcting anything, just taking the digital input as it comes. And it shows an issue. But when connecting to a standard DAC (an SMSL PS200 in my case), the receiver of the DAC can manage to correct the errors, as you can see with the same sine:
The result is nearly perfect, very close to the original WAV file.
I thought about a way to test the data stream input, without making it a nightmare. So I simply reused my standard 3DC levels test (997Hz @-90.31dBFS) to verify if a 1 would turn out to be a 0 sometimes (or the other way around). Theoretically, and in practice, I should see this:
At this level we must see a scare (lowest symmetrical signal in 16bits). Ringing is due to the Gibbs Phenomenon.
But with the Sony, sometimes, I see that:
Haha, a 1 turned out to be a 0 in that case (0000000000000000 instead of 0000000000000001). So I did not need to deep dive into this, there are indeed some errors in the transmission, probably some aging components (I'll check later, maybe).
But, at the end of the day, it was nice to see that the external DAC had no issue dealing with these errors. When I think some say they can hear a difference between two optical cables
Sony CDP-X333ES - Testing the drive
What would be good measurements if the drive would not properly read a slightly scratched CD, or one that was created at the limits of the norm? The below tests reply to these questions.
| Test type | Technical test | Results |
| Variation of linear cutting velocity | From 1.20m/s to 1.40m/s | Pass |
| Variation of track pitch | From 1.5µm to 1.7µm | Pass |
| Combined variations of track pitch and velocity | From 1.20m/s & 1.5µm to 1.40m/s & 1.7µm | Pass |
| HF detection (asymmetry pitch/flat ratio) | Variation from 2% to 18% | Pass |
| Dropouts resistance | From 0.05mm (0.038ms) to 4mm (3.080ms) | Up to 0.75mm. |
| Combined dropouts and smallest pitch | From 1.5µm & 1mm to 1.5µm & 2.4mm | <1mm. |
| Successive dropouts | From 2x0.1mm to 2x3mm | <1mm. |
The drive was able to consistently continue playing, without generating typical digital clicks, with dropouts of up to 0.75mm. The interpolation effect remained hidden to my ears when it kicked-off but failed to maintain a constant flow beyond 0.75mm as it stopped playing with 1mm dropouts. The Sony had no issue with variable linear velocity and/or track pitch, as well as with HF detection. These results are a little lower than what I measure usually. I guess we have to pay the price for the speed of that drive.
Conclusion
Let's forget the issue with the digital output (which is a not an issue at the output of a DAC), and recognize the quality of this player. Besides to non-resistance to intersample overs, the results are extremely good. There is absolutely nothing bad to mention here. I'm really happy to have had the opportunity to test this first 1bit conversion offering from Sony.
This was SOTA at the time, and is still a very good for a CD player more than 30 years later. I must say I did not expect so good results. There's only the noise floor that limits the real resolution of that player to a still very good 17+bits, which is probably due to the ancient noise shaping (I'll add more information about that later).
This 1bit tech was a preview of the DSD and SACD to come. Sony folks knew what they were doing, no surprise.
Oh yes, it sounds really good to my ears, like all the others
I hope you enjoyed this review.
Have a nice week!
Last edited:
