Bottom Fishing for CD players

After all the talk of different CD players and my comment that I didn't want to scratch my mint Sony
I decided to go bottom fishing on Ebay looking for the cheapest brand name CD player I could get.
At the same time I avoided the ones with permanently enabled Anti Shock. I have a Panasonic like that which, makes CDs sound like a poor quality MP3.

I have found it is often worth looking at the Buy it Now and found this working Aiwa XP-V30 for just £10
including postage.

It arrived yesterday, nicely packed and in full working order. It sounds pretty good on vocals and Piano.


One of the few articles I found out about it was in Japanese which gives it a favourable review.
https://blog.goo.ne.jp/domzoh/e/b820578c2ed4a895d8e08b9444b354e2
From the photos it appears to be based on a Sony main IC and doesn't have
any of the failure prone surface mount capacitors in it.

The second photo is of it playing a CD from a late Christmas present to myself.

A Dollar box set that cost almost five times what I paid for the player.
38 years after I got the Dollar Album on LP for Christmas. The set includes a DVD which as an
extra has the TV advert for it.

Something I realised thinking about the discussions on sound quality is that all their Albums
actually pre-dated CD so when the tracks were originally being mixed the producer (Trevor Horn for the last one) would have been thinking how it would sound on Vinyl or cassette

I now have my eye on a cheap Technics so this thread might continue.
Any opinions on an SL-XP150 ?

 

"At the same time I avoided the ones with permanently enabled Anti Shock. I have a Panasonic like that which, makes CDs sound like a poor quality MP3." — usually, they have a setting allowing to switch between longer and shorter buffering time. If I understand correctly, with shorter buffering time there is no compression, just buffering. For a longer buffering time they recompress PCM into a lossy format. OTOH, 320 MP3 is perceptually lossless, so I would not care much about recompression. As long as there is no hiss, wow and flutter, I am a happy camper.

"Any opinions on an SL-XP150 ?" — Seems that Panasonic was still making old-school pre-1bit DACs in 1993. I would not want it, I am strictly in a 1-bit DAC camp. My component Technics player is from 1991 and has a 1-bit DAC. The Aiwa XP-V30 above does have a 1-bit DAC though.

 

@Longman if you are happy with Aiwa then no need to spend more on SL-XP150. To my ears this Technics sounds better than most 1-bit Discmans but as @Radio Raheem once said: "just enjoy the sound you have, it doesn't matter if anybody else doesn't like it as it's for you're ears only lads" - Priceless!!!
The benefit of Discmans is they all go 20-20k, with zero W&F, and the only hiss you hear is from the original Master Recording. And if you do not care about the differences between 16/44.1 and MP3 at 320kbps (often marked as "Insane Quality" - some folks do not care about the differences between 192kbps and CD or hi-rez) then you get yourself a lot of extra $$ to buy more music But do not confuse MP3@320 with ALTRAC compression used in the Discmans... you hear the same thing as me here - with skip-protection engaged Discmans sound Insane to my ears
I try to stay away from 1-bit DACs/Discmans (with an exception of D-99/D-311/D-515/D-777) BUT last week I saved myself $150 by buying Schiit Modi 3 instead of its 'multibit' brother. The "savings" were used right away to buy yet another set of headphones

 

Did you mean ATRAC? I doubt you will hear any difference though. Also, isn't it Sony's proprietary codec? So other manufacturers would use something else... like MP3? Or maybe AAC or AC3.

This is what the documentation for my Panasonic CD player says:
POS1 - 10-second anti-skip memory
POS2 - 45-second anti-skip memory
128 kbps MP3 - 100-second anti-skip memory

So, if the buffer is enough to fit 100 seconds of 128 kbps MP3 then its size is about 12800 kb or 1600 kB or about 1.6 MB (I suppose there is a 2MB RAM chip in there).
CD-DA is 1411 kbps, calculating 12800 kb/1411 kbps = 9 s, almost 10 s, which means that POS1 does not use re-compression, it is strictly buffering.
For 45 s buffer we have (10/45) * 1411 kbps = 313 kbps or so, which is perceptually lossless, and I would not hesitate to use this setting for most of my listening.

Isn't it funny that someone who would not accept compression artefacts would rather accept decoding artefacts from a pre-1bit DAC?

 

All tunes on my iPhones are kept at 192kbps (as folks at Linn forum say: "shit in - shit out"), so I am not sure if this post above was addressed at me. @CDV - as you warned in another thread you do sound borderline insulting; but as long as I made you laugh... we are buddies! No need to ask where you got the idea of Sigma-Delta DACs superiority because you've already posted a few links. This opinion of yours is not wrong, but presenting it as a fact is. My neighbor Jason Stoddard is the latest player in hi-end audio, reading his book "Schiit Happened" (also as a blog at head-fi.org) gives a very different view. All I had to say on the subject was said HERE, and my personal opinion on general inferiority of 1-bit Discmans was formed not from reading or comprehending the differences between noise-shaping, oversampling and what-not but from owning dozens (maybe over a hundred) Discmans between 1994 and today. Photo of my previous Discman collection is my intro post at S2G, unfortunately right now I own only my favorite models: D-10, D-50 MkII, D-555, Technics SL-XP5, Denon DCP-100.

@CDV - could you post your favorite Discman, h/p amp (if any), and headphones? And/Or your main stereo? I keep saying this at Audiogon crowd (but no one ever listens to anyone there) that seeing other folks stereos (and fav tunes) explains a lot about their posts and preferences

 

I do not use any Apple products except for a Macbook that my employer gave me. I have an Android phone with a memory card slot and with a proper access to file system, so I can upload any files I want with any bitrate I want, I am not limited by iTunes. I have many CDs with MP3s having bitrate of 256 or 320 kbps. These CDs have been sitting mostly in folders for the last 15 years or so.

I did not expect my my remark about pre-1bit DACs to be considered offensive, it is a half-joke after all. There are many good ones, although the quality is achieved by brute force: high-quality components, laser trimming, oversampling, several DACs per channel with averaging, and of course adjustment pots. 1-bit DACs solved the inherent instability of traditional DACs by converting high-resolution low-frequency digital signal into a lower-resolution high-frequency signal. Turned out it is much simpler to design a reliable high-frequency generator than to trim resistors with a laser. So, 1-bit DACs opened the gates to high-quality sound to everyone, not just to audiophiles who can spend pretty penny on expensive gear, how bad can that be? Great and consistent sound quality for every Joe and Jane, cheap.

 

I am getting bored by this circular argumentation: if 1-bit DAC players looks good on the paper, can be programmed by a cleaning lady, cheap, and Not made of high-quality components...... then this is definitely what every Joe must get. More $$ left to buy something nice for Jane. The only danger for Joe might be in going out and actually listening to some R-2R or Multibit players/converters. Nowadays they are easy to find: following 20-year-long 1-Bit mantra they made a serious comeback

- this is simply not true. Both the first (D-99) and the greatest (D-515) 1-bit Discmans have adjustment pots. those are for laser adjustment, not DAC. Later 1-bit cheapies have self-calibrating laser mech, thus no trimpots

this below is a setup from 1996 which traveled with me around the Globe quite a few times Unlike its older brothers, this Panas does not have MASH stamped on its lid. For the first few years Panasonic made a point about their superior way of dealing with noise shaping, that is how we've learned that Sigma-Delta had/has its own serious issues Anyway, my travel setup was quite good for what it was, but before saying that this or later Panas model sounds better than Sony D-90 you must show me the proof of you actually listening to both. Only then I will admit that your hearing/preferences are different from mine

 

How about I go back and remove the last line of post #4 and we call it a day. The main point of that post was calculating the buffer size and making an educated guess that buffering does not necessarily mean re-encoding to a lower-bitrate format.

Regarding your purchase of Schiit Modi 3 instead of Schiit Multibit, looks like you did well.

 

Before I lose it I have found an interesting thread on Tapeheads about Anti-Skip buffering
https://www.tapeheads.net/showthread.php?t=23201
I am surprised that they were even considering using 1MByte of memory back then. DRAM would be too power hungry and SRAM was expensive.

Here is the Panasonic player I have mentioned before, another bottom fished player I bought a couple of years ago.
An SL-SX390 old enough to be made in Japan

I was hoping the Mode button would allow me to switch buffer lengths like someone mentioned in that thread.
but it just changes the Play mode i.e Normal Repeat or Random. It seems like I am stuck with 48 seconds Anti Shock. I did check the instructions when I got it and couldn't see if I could change.
Great if you want to use your player on a Fairground ride. Not so great for quiet listening at home.

I will discuss DACs when I have more time as it is almost time for bed here now.

 

The picture from the tapeheads thread, which corresponds to what I wrote above (although I have a slightly different model). Instead of discussing the numbers, the forum members preferred to focus on what POS could stand for.

 

I said I would write about DACs so here goes

When CD was released the Japanese players used 16 bit DACs (often a single one shared between both channels).
Philips were having difficulty making 16 bit DACs so used two 14 bit DACs. To improve quality they oversampled the DACs at four times.
Many people thought that the Philips players sounded better than most of the Japanese ones.

Manufacturers realised that oversampling could improve sound quality. They also made the players easier to make, not needed such powerful
filters on the outputs. By the second generation of players 4 times Oversampling was the norm. Where would this lead ?

Panasonic and Philips both started experimenting with even higher levels of oversampling.
It is interesting that the SL-XP150 portable CD player proudly boasts 8 x Oversampling but doesn't say how many bits.
I suspect they didn't want to say. At the time, anything less than 16 would have been considered inferior and been a turn off to potential buyers.
Soon afterwards Panasonic / Technics (both same company) launched what they called MASH "Multi - stAge Noise SHaping".
By this time the buying public were seeing through the marketing hype of using 18 bit DACs to play a 16 bit source (which was happening at the time)
and were ready to embrace new ideas providing they sounded good.

MASH used 4 bit DACs (curiously referred to as 3 1/2 bits in some early reviews) but a greater level of oversampling than had ever been used before.
The consensus was that this approach sounded pretty good.

At around the same time Philips launched the first true one bit DAC. They soon used it in their top of the range CD850 HiFi CD player
to produce the MkII calling it Bitstream
This review shows people were impressed https://zstereo.co.uk/2013/04/12/philips-cd850ii/

However, their next step was a step backwards, coming up with a different one bit design they called Bitcheck.
Whether this was to reduce costs, reduce power consumption for portable players, or simply ensure that there was still a market for high end players I don't know.

Here is a review of a player using it https://www.tnt-audio.com/sorgenti/cd723_e.html
It certainly isn't as positive as the Bitstream reviews

Anyway my conclusions are:

You can't necessarily expect modern players, often built down to a price, to be as good as expensive ones designed for audiophiles when the manufacturers were trying to prove that CDs could be better than vinyl.

Secondly you only have to look at the popularity of valve (tube) amplifiers to realise that distortion comes in different forms. Most valve (tube) amplifiers cause significant distortion. However, unlike transistor amplifiers the distortion is musically pleasing. That is why electric guitarists love them.
Something I saw at a HiFi show before 1989 was an early attempt to combine the convenience of CD with the warm sound of valves.
It seems that it is an idea that has never gone away. https://www.rega.co.uk/products/valve-isis

Anyway as "Jorge has said the main thing is simply whether you enjoy the sound. Annoyingly the only MASH CD player I currently have is in a CD Radio Cassette which I recall having a very noisy headphone output. It was after discovering that, that I bought my first Sony portable CD player.

I'm still keeping an eye on eBay, especially as most of the players are still a fraction of what they cost new and it is interesting to compare them. I would expect to get three of four used but working players for what Jorge paid for his DAC.

 

Interesting, I thought MASH was a 1-bit DAC. Do you have a link? This is what surprised me: both "MASH" and "4DAC" on my SL-PG100. I searched for this discrepancy, but to no avail. Elsewhere I've heard that "4DAC" was a remnant of earlier models, which indeed used two DACs per channel, and Panasonic did not remove this label either because they was too late in the production process, or because they were afraid of consumers' response to "1-bit" DAC after being told for years that more bits was better.

Audiophiles? Ah, I've heard of them — they think that good equipment must be heavy, expensive, not easily accessible to common folk, and should require expensive and regular maintenance by specially trained technicians.

I do agree with you that development cycle usually goes from improving specs and reducing size to just making it cheap, so products made sometime in the middle of the cycle are usually the best. The latest CD players may have a DAC as good as or better than "audiophile" models of yore, but they have crappy transport made of cheap and noisy parts, which breaks apart within a year. Older devices, on the other hand, are usually mechanically more sturdy and reliable, but their electronics is not so advanced.

Anyway, the CD player that I got on a garage sale for $15 is exactly from the middle-of the cycle timeframe I am quite happy with it. Sadly, it does not have digital output. I think in the early 1990s very few models had it, because the standard has not yet been fully agreed upon.

 

Someone is claiming it here
https://audiokarma.org/forums/index.php?threads/technics-panasonic-mash-cd-players.825663/
Another forum said that they are 2.5 bits with the explanation that the output can go to five levels.
Of course with what is effectively a tradenames the manufacturers could change things as much as they liked internally for each new device.
It is even more difficult to find any explanations of what Panasonic were doing than Philips.
I just found a Philips datasheet that clearly states that one of their Bitstream DACs was 5 bit with 96x or 128x oversampling
http://www.acoustica.org.uk/t/naim/data/TDA1305.PDF
p.s I just checked and that datasheet is from 1995 so back when Pjilips were still trying to "Lets make things better".

 

Just one guy off-handedly mentioning 4-bit DAC... it is quite thin. But thanks for taking time to look it up. I'll dig further.

Thanks! I learn something every day. Well, not exactly learning, because most of this stuff is Greek to me. I think it is less important whether they send the signal in one lane ("1-bit") or multiple lanes, but that the bitstream principle is still employed: converting high-resolution low-frequency digital signal into a lower-resolution high-frequency signal. From what I've gathered, most (all?) modern bitstream DACs are multibit ones.

 

Digging deep I did find where the 3.5 bit number I had stuck in my mind had come from in an ETI article from September 1989

Before we had the internet I was an avid reader of technology magazines, getting Electronics Weekly at work so probably read it there.
A 33MHz IC was cutting edge stuff back then comparable to the fastest CPUs
The promised comparison between the Philips and Panasonic technologies never seemed to materialise, with just one more "Playback" article discussing the latest Sanyo DACs that seemed to be a combination of interconnected 9 & 4 bit converters with oversampling. There are probably literally 101 ways to implement an audio DAC IC.

However, back then I certainly had no intention of changing my CD player having only bought it, second hand, a couple of months earlier. Interestingly the highly paid manager I bought it from was already upgrading, although I don't know what he bought to replace the first generation Hitachi I bought from him.

In the next ETI issue I did find someone complaining about the quality of CDs.

I wonder what the writer would have made of MP3 ?

 

Whoa, thanks, this is very insightful! Thanks a lot!

They were smart enough to blame not the technology but the recording and mastering process that did not take full advantage of DDD.

MP3 and other lossy formats is good technology. It is accepted that 128 kbps MP3 is as good as analog compact cassette or better (W&F is clearly miles better), and 320 kpbs is perceptually lossless. If this technology, both the algorithms and the hardware, existed in the late 1970s, we might have not seen CD-Audio at all. We are lucky that Philips and Sony created a standard that is, basically, a golden mean between LPs and tape on the one hand and hi-res digital on another.

Maybe you can help me understanding the charts in the 2008 Stereophile article. The very first chart is supposed to show "the familiar 96dB signal/noise ratio", but the top is at -20 dB, and the noise is at about -130 dB, so should it be 110 dB, not 96 dB?

The LPCM chart on the second page is easier to understand: top at -40 dB, bottom at -130 dB, the difference is 90 dB.

For 128 kbps MP3 the noise floor has risen to -80 dB, so I suppose the dynamic range is 80 - 40 = 40 dB, is this correct?



If I am reading it correctly, this is worse than a Type I tape without Dolby, but why I don't hear any hiss when listening to MP3s? Does the noise have a different character? Or I should not simply subtract 40 from 80? Or I cannot read the graph? Or 40 dB is more than enough for casual listening?

 

This video on Youtube explains why the technology didn't exist in the 1970s and only in research labs in the 1980s

I guess the reason is in the Noise Shaping part of the MASH name and similar techniques. If there is no sound at a point on the CD everything should be static and there shouldn't be any noise. Dire Straights "Brothers in Arms" was described as the CD that sold a million CD players, and was a favourite for demonstrating them. Why? Listen to how much silence there is in the track and then imagine that on hissy tape or a noisy LP.

Unlike tape lossy compression can't be tested just by using test tones although I suppose you might get noise in them. JPEG also uses lossy compression. Here is a 200 pixel high file I created using Paint

I then dropped the file height down to 50 pixels so that is a 16 to 1 overall compression then re-expnded it back to the original size

The white areas are still noise free, but look what has happened around the text: noise.
The text is still readable but it certainly doesn't look as good.
That is similar to the noise audio compression introduces.
Now imagine that happening with five instruments in a band.
I still find it incredible that the designers of MP3 used Susan Vega "Toms Diner" as their main reference

as it is basically her voice and not much else. I'm sure we all know what happens when you try to play music through a GSM phone (which admittedly can be using a 32Kit/s channel).

p.s Youtube must know my tastes as it went straight from Susan Vega to Tiffany. .
Guitars, Drums, Keyboards, and multi-tracking it has the lot.

 

Um, thanks, but what about the graph above? Does it show 40 dB dynamic range or I am reading it wrong?

Although CDs are miles better than cassettes, I have Brothers in Arms both on CD and on cassette, and the cassette version sounds just fine with Dolby turned on.

As I understand it, noise shaping, which usually is used in combination with dithering, allows to flatten spurious noise peaks and to move away the noise beyond audible range, where it can then be cut by a filter.

This video describes dither and noise shaping starting from about 9-minute mark:

 

To be honest I am not sure what the chart is trying to show. It looks to be about 40 equally spaced test tones so the overall power of them will be much higher (16dB if I am not muddling power and voltage). The noise also seems to drop significantly at 8KHz and 16KHz.

As I said a single test tone isn't a good test. 40 test tones will be more representative of an Orchestra or Tiffany and her band.

 

This is exactly what they did. The signal "simulates a musical signal by combining 43 discrete tones with frequencies spaced 500Hz apart. The lowest has a frequency of 350Hz, the highest 21.35kHz."

It shows peaks and valleys and the noise floor and the dynamic range.

Ah, so I should not simply look at the tops and the bottoms, I need to take into account how many of these tops I have? Thanks, that makes it somewhat clearer. I need to look for a formula I guess.