|
|
Post by palace on Sept 16, 2024 12:45:08 GMT
Gents an observation, Philips were the company to introduce the CD player to the world, alongside Sony, in 1982, I bought a Philips CD 104 Red Book standard which uses uncompressed pulse code modulation (PCM) with 16-bit resolution and a sample rate of 44.1 kilohertz (kHz) this must have been in late 1984 it sounded quite good with the right CD, a lot of early CD's unfortunately were like fingernails on a blackboard after a couple of years the CD 104 became intermittent in use. The CD104 used/uses a CDM-1 transport that Philips developed by themselves. The basis of this is a cast-iron form which holds a sophisticated swing-arm laser paired with six Rodenstock glass-lenses. In terms of musicality, the CDM-1 is considered to be the best transport ever made if a little slow in action. Following the audiophile rule of “garbage in = garbage out”, a flawless reading of the source material is the basis for musicality. The CD104 was designed initially as 14 bit resolution 44.1 Khz sample rate using 2 x TDA1540P mono multi-bit DACs, 1 per channel. Philips discovered that Sony and others were coming to the market with 16 bit Players Philips pushed it's engineers into adding oversampling in the guise of a SAA7030 error correction chip, making the CD104 16/44.1. If the 14-bit DAC was ever considered to be a handicap by customers, I can tell you that no failing is audible at all. In fact, the later Philips 16-bit TDA1541 DACs were used in Sony’s High End players well into the 1990s, which says a lot about what Sony thought of the Philips DACs. A good few years ago I did some research on the internet, it appears that most problems with this CD player result from "Griplets" a type of hollow rivet that goes through the double sided PCB joining the ground planes. The "Griplets" fail, simply re-soldering does not work even if a meter shows continuity! as strongly suggested on various forums using a fine bit, drill the "Griplets" easily identifiable as blobs of solder were drilled through then silver wires I had some spare, were pushed through and soldered on each side. This player has fixed signal/phono leads so a pair of my DIY leads were soldered in place, though we had chassis RCA sockets to hand this saved on soldered joints. The result was even better than I remembered & functioned flawlessly. This work was done by my son then at technical college, after further reading I saw the NOS conversion ie no overall sampling where the SAA7030 error correction chip is removed & 4 wires soldered across the resulting gap on the PCB out came my son's soldering iron, he unsolderd the SAA7030 Over-sampling chip & fitting a 24 pin IC socket, rather than the 4 wires he plugged in a "No Overall Sampling PCB with flip/flop IC" I bought from Bulgaria it reclocking the CD104 my son also soldered a Ik resistor from pin 18 to pin 16 cutting a track on the SAA7000 to get 5v and make the chip recognise 14 bit rather than 16 bit. My son had this player for his system. I was so impressed with it that I bought another immaculate sample off the "Bay" it's also an early grey case that came in its original box my son duplicated his previous endeavours plus this CD104 came with phono sockets he replaced the 3" wires with the silver wire I had left over from making phono leads. As to 14 bit the CD104 after conversion has more detail strange but true, better bass indeed very analogue sounding; it sounds so natural I prefer it to the Cambridge CD4se (16/44.1) that I replaced the CD104 with, I now use for convenience the CD4se as a transport as it has a rear BNC rear digital output that the CD104 lacks. I have not asked my now married son with our granddaughter to create a digital output as lampizator shows on the internet. My convoluted point is, if 14 bit can work so well with well recorded CD's is it the playing mechanism including streaming or the CD including streaming or finding the perfect match between them. What is interesting is that there is for sure a lot of audiophiles that is going back to old philips chipset, and old digital technologies, claiming that they have some improvements in certain areas. In my mind it makes sense that lesser technologies has natively less inherent noise, since it is less processing going on on the DAC chip itself. We use trickery to remove noise now a day´s, like jitter reduction methods. But such processes will also create noise. The idea with these older "purer" chips is that they could potentially have a lower inherent noise floor and by that an even lower noise floor potential, if fed with very clean signals. Having said this i don´t want to say that the old way is necessary better, but i think there are actually aspects that can be amazingly good with the older, less process heavy technology. Tobias thank you I was attempting to write a coherent reply you posted before me, my likening the loss of something from 16 bit in the signal that I compared to RFI/EMI being removed allowing more perceived detail to come through was not it would seem so far off the mark. As a great believer in K.I.S.S. your answer pleases me, & no it is not the best player out there. Whether the "No Overall Sampling PCB with flip/flop IC" that apparently reclocks the CD104 helps but it is better than the 4 wires alone. Do later CD's contain an extra something causing noise in the 16 bits' ignored by the NOS converted CD104's? |
|
bencat
Rank: Quartet
Posts: 353
|
Post by bencat on Sept 16, 2024 13:09:42 GMT
For anyone interested I am taking a set of the original Phillips chip set to the Maverick Show and they will be for sale with all monies earns for the show charity . So those skilled at DIY can make there own NOS DAC .
|
|
Tobias
Rank: Quartet
Posts: 320
Member is Online
|
Post by Tobias on Sept 16, 2024 13:26:43 GMT
What is interesting is that there is for sure a lot of audiophiles that is going back to old philips chipset, and old digital technologies, claiming that they have some improvements in certain areas. In my mind it makes sense that lesser technologies has natively less inherent noise, since it is less processing going on on the DAC chip itself. We use trickery to remove noise now a day´s, like jitter reduction methods. But such processes will also create noise. The idea with these older "purer" chips is that they could potentially have a lower inherent noise floor and by that an even lower noise floor potential, if fed with very clean signals. Having said this i don´t want to say that the old way is necessary better, but i think there are actually aspects that can be amazingly good with the older, less process heavy technology. Tobias thank you I was attempting to write a coherent reply you posted before me, my likening the loss of something from 16 bit in the signal that I compared to RFI/EMI being removed was not it would seem so far off the mark. As a great believer in K.I.S.S. your answer pleases me, & no it is not the best player out there. Whether the "No Overall Sampling PCB with flip/flop IC" that apparently reclocks the CD104 helps but it is better than the 4 wires alone. Do later CD's contain an extra something causing noise in the 16 bits' ignored by the NOS converted CD104's? Hehe, i had to lookup K.I.S.S (i am not natively English speaking) and now i have learnt something! I am definitely starting to see that there could be merits in keeping it simple, especially if we manage to get down to a really low digital noise floor, where complexity also can become a catch 22 hurdle, at some point. Keeping things at a lower resolution is also (likely) a step towards even lower digital noise floor potential, avoiding down sampling ideally (to try to tie this to the topic) |
|
|
|
Post by palace on Sept 16, 2024 13:41:18 GMT
Tobias thank you I was attempting to write a coherent reply you posted before me, my likening the loss of something from 16 bit in the signal that I compared to RFI/EMI being removed was not it would seem so far off the mark. As a great believer in K.I.S.S. your answer pleases me, & no it is not the best player out there. Whether the "No Overall Sampling PCB with flip/flop IC" that apparently reclocks the CD104 helps but it is better than the 4 wires alone. Do later CD's contain an extra something causing noise in the 16 bits' ignored by the NOS converted CD104's? Hehe, i had to lookup K.I.S.S (i am not natively English speaking) and now i have learnt something! I am definitely starting to see that there could be merits in keeping it simple, especially if we manage to get down to a really low digital noise floor, where complexity also can become a catch 22 hurdle, at some point. Tobias: I'm sorry about K.I.S.S. My Swedish neighbour & American wife unfortunately moved out yesterday never got any of the promised Surstromming (salted fermented Herring in a blown tin can). In my experience Scandinavians speak better English than some Brits. but colloquialisms are a different thing. After the recent help from this forum re removing/shielding from EMI/EFI is what made me think of the comparison. bencat: That is a kind offer I see from the internet they are making good money these days, however many moons away from my skill set. |
|
Tobias
Rank: Quartet
Posts: 320
Member is Online
|
Post by Tobias on Sept 16, 2024 15:08:29 GMT
Do you mean the album My Method Actor from 2024? It's in 24/44, an unusual hi-res format which donwsamples relatively easily to 16/44. It's quite compressed at -8.43dB so not using all the usable dynamic range. I can't comment much more, the music is not to my taste. I saw this comment now. Yes, its from this Friday and was my most anticipated album this year, since i loved her last one.  Ok, that is a good finding since then it makes sense that the downsampling did not actually cause much processing activity from the streamer (in theory), at least less than if it was 192... I was planning to listen a lot more carefully today but will not get to it today unfortunately. |
|
Tobias
Rank: Quartet
Posts: 320
Member is Online
|
Post by Tobias on Sept 20, 2024 12:29:32 GMT
Spent some more time listening and comparing 16/44 fixed from my streamer VS no downsampling.
Short story is that I still very much prefer 16/44 in all cases. It is however clearly less resolving (obviously) but also less bright, harsh and flat sounding. I think the big difference is that i want to play louder, due to the more relaxing sound. Louder volume gives more of everything and it sounds fuller, denser, more textured, more harmonic richness and just more musically pleasing.
Regardless of what resolution the source material has (if more than 16/44) it is a much more enjoyable and musical/analogue type of sound when it is downsampled.
To give an example i never liked the sound of the opening track on the OK Computer album, by Radiohead. This is otherwise a favorite album but that opener hasn´t been very nice sounding but gotten much more acceptable as i improved my setup. Now it turned into something very enjoyable with a much more textured and rich sound that was rather amazing.
This seem to be something related to my DAC (chip?) since i found an article where he said that the 16bit sound is much preferred over the 32 bit sound, and their sound description matches my experience exactly.
Hmm, interesting... I will now stick to 16/44 but will probably compare again after having done updates in other area.
|
|
|
|
Post by John on Sept 20, 2024 12:37:24 GMT
Glad this is working best for you.
I think it just goes to show how different chips and implementations play a part in sound.
|
|
|
|
Post by MartinT on Sept 20, 2024 14:04:17 GMT
It is however clearly less resolving (obviously) but also less bright, harsh and flat sounding. This is classic issues around digital noise. I spent four years getting hi-res to sound better and it isn't easy - but when you've cracked it, all sample rates & bit-depths sound the same in balance, just with increasing resolution. |
|
Tobias
Rank: Quartet
Posts: 320
Member is Online
|
Post by Tobias on Sept 20, 2024 15:05:00 GMT
It is however clearly less resolving (obviously) but also less bright, harsh and flat sounding. This is classic issues around digital noise. I spent four years getting hi-res to sound better and it isn't easy - but when you've cracked it, all sample rates & bit-depths sound the same in balance, just with increasing resolution. Yes, i have to agree on that in general. I was very skeptical about downsampling since it should only make things worse, since that should rather add processing noise in my streamer, as you have also said before. But the result was not what i was expecting and it sounded like how less digital noise usually positively impacts the DAC... I still think there is a DAC chip variable in play as well, where some older chipsets can technically handle hi-res but might not ultimately perform at their best with it. (but i could be wrong) |
|
|
|
Post by MartinT on Sept 20, 2024 16:17:26 GMT
I still think there is a DAC chip variable in play as well, where some older chipsets can technically handle hi-res but might not ultimately perform at their best with it. (but i could be wrong) Yes, that's entirely possible. My experience with DAC chipsets is the ESS 9038 Pro (2 devices, the current one much better in terms of lower harshness) and a venerable Wolfson chipset before that (which sounded good but was limited). |
|
Tobias
Rank: Quartet
Posts: 320
Member is Online
|
Post by Tobias on Sept 20, 2024 17:42:00 GMT
I still think there is a DAC chip variable in play as well, where some older chipsets can technically handle hi-res but might not ultimately perform at their best with it. (but i could be wrong) Yes, that's entirely possible. My experience with DAC chipsets is the ESS 9038 Pro (2 devices, the current one much better in terms of lower harshness) and a venerable Wolfson chipset before that (which sounded good but was limited). Martin, may I ask you about your general thought on HighRes re-releases? I understand that it is more variables then the resolution, since it can be mastered differently, but do you generally prefer the highRes version of albums that was initially 16/44? I mean, if we take away the DAC from the equation, if we can do that... For example, if you are going to listen to a classic album. Will you then go for a HighRes version of the original version, in general (if streaming)? |
|
|
|
Post by MartinT on Sept 20, 2024 18:53:53 GMT
In general, I always prefer the hi-res version.
There are a very few anomalies in mastering, but it's rare. One real oddity is Steely Dan - Gaucho: available on Qobuz in 24/192 and in a different mastering in 24/96 (both dated 21 November 1980). In this particular instance, the 24/96 recording sounds far more dynamic and vivid. That, to me, is a different mastering rather than the fault of the format itself.
|
|
Tobias
Rank: Quartet
Posts: 320
Member is Online
|
Post by Tobias on Sept 20, 2024 19:04:43 GMT
Thanks for sharing, much appreciated.
I have sometimes felt that some dynamics is lost in HighRes, without knowing the actual reason for that of course.
Was playing around with the thought that HighRes will be more noisy just because it will load the DAC with more data to process. (Thinking about what Sony said about Red Book being the sweetspot mathematically)
This would mostly be noticed if you have a really low noise floor, since then the dac´s own inherent noise floor, at operation, starts to matter and eventually become the bottle neck (I was thinking).
But then that doesn´t go hand-in-hand with your experience.
|
|
Tobias
Rank: Quartet
Posts: 320
Member is Online
|
Post by Tobias on Sept 21, 2024 19:11:05 GMT
I did a mistake!
Part of my testing was done with upsampling turned on, which i missed and didn´t understand...
In general, i still prefer 16/44, but now by a very narrow margin.
Some things i said, where there was a very big difference, was when my streamer upsampled to 24/192, which has a big negative impact since it has all the negatives that I associate with noise, which makes a lot of sense.
Sorry for sharing some information that was partly wrong.
|
|
