Philips CD723 measurements

[audio_tony]

Measurements of a Philips CD723.

From memory, this was a bottom of the line (or close to) player. It mostly measured as expected, however it had some real issues with the IMD, multitone and jitter measurements, all of which raised the noise floor considerably.

I do seem to recall that this player was quite popular with 'modders' back in the day however, so it must have had something going for it.

It is of course possible that my example is faulty / requires recapping - it is rather old now, and never been opened as far as I can see.

NOTE: The caption in the first graph should read -0.1dB not -0.06.

My test setup: Asus Xonar STX sound card with a self built buffer / gain box in front of that. Host PC is an Intel i7 running Windows 10 with 32G RAM and SSD - no mechanical disks or CDROM drive present.
I also use an Altor Audio Olivine-2 ADC, however the Asus Xonar was used these measurements.
Sound card input was set to 24bit/96kHz
REW was configured as per the recommendations in the post from @NTTY

 

[DSJR]

I had one of these some years back and 'modded it' with suggestions by Lampizator I recall - easy to work on the small main board . Not sure it made the slightest subjective difference at all, but it was usable as transport, as I didn't butcher the digital output as also recommended. When I was gifted a nicer Denon DCD-1015 (a MUCH better player for the senses all round I discovered), I gave the Philips away to a pal who I think still uses it occasionally.


No idea if anyone else feels this, but a touch of noise at or below -75dB doesn't necessarily hur the listening experience. In face, I recall it said that said once that noise can enhance the listening experience. No idea at all if ano of you lot would even begin to agree and it matters not to me now, as I can listen to UK FM BBC Radio 4 and 3 with no issues regarding background noise (the other BBC stations suffer hugely from compression in comparison). U thirteen bit FM is around -70 to -75dB S/N isn't it?

 

[AnalogSteph]

Holy moly, that thing has some jitter issues. I guess the fact that it has the processor/filter on one board and the (multibit) DAC on another wouldn't exactly be helping in that regard, but this stinks. I wonder what a scope with a 10X probe would say to the BCLK, WCLK and DATA signals, and whether there's any modding potential at all (not the Lampizator nonsense, obviously). Oh, and it uses the infamous SAA7378 processor, too (the one that applies a small level reduction crudely at 16 bit with truncation even to digital output). What a steaming pile.

BTW, the FR is just about useless like that. Try 1/48 smoothing and a bunch more averaging with like a 10 dB vertical scale instead, and nobody ever got hurt by including some of the out-of-band filter response either.

 

[audio_tony]

BTW, the FR is just about useless like that. Try 1/48 smoothing and a bunch more averaging with like a 10 dB vertical scale instead, and nobody ever got hurt by including some of the out-of-band filter response either.

I'll bear that in mind.

I actually have a proper analogue sweep (~10Hz to 45kHz) that I generated from my signal generator and digitsed. I should probably start using this instead.

Incidentally I suspected that something was wrong with this player so I took a look at the FFT and it's jumping all over the place, which due to the smoothing etc I'm using in REW I didn't see.

 

[AnalogSteph]

I actually have a proper analogue sweep (~10Hz to 45kHz) that I generated from my signal generator and digitsed. I should probably start using this instead.

Sine sweeps are kind of notorious for requiring very precise clock alignment, but it's certainly worth a try. Probably depends on what you're using for an algorithm. (RMAA uses MLS sweeps, which may be one of the best but are fairly timing-critical as well.)

If it were me I'd be tending towards using pink noise to have the same energy in every octave (check "Adjust RTA levels" in REW), otherwise the bottom end is going to be relatively poorly-defined on a log frequency scale.

 

[audio_tony]

Sine sweeps are kind of notorious for requiring very precise clock alignment

This is what I get if I play the sweep back in WaveSpectra. FFT is 'Flat top' and size is 4096.

If I try to narrow down the amplitude to improve the scale is does go a bit awry though!