Wondering if any of you use a PS3 as an SACD or CD player.
If so, which bitmapping type do you prefer and why?
Is it best to request upsampling from my PS3, or leave things at 44hrz?
Printable View
Wondering if any of you use a PS3 as an SACD or CD player.
If so, which bitmapping type do you prefer and why?
Is it best to request upsampling from my PS3, or leave things at 44hrz?
For SACD playback, the stream is converted to LPCM within the PS3 before being output over HDMI (even with HDMI 1.3 and DSD-over-HDMI capable receivers). In simple terms, the SACD's signal is being converted.
For that case, the higher the quality, the better. Last I checked, the "upsampling" frequency was used to determine SACD re-sampling frequency (though the PS3's audio output settings should really control this and allow up to 192kHz).
Bottom line. You will get the best SACD playback by having "upsampling" turned on.
One caveat here. For my Onkyo receiver, the Audyssey EQ will not function at 176.4 kHz (I believe the max is 96kHz). With flat Ascend speakers and minimal room interference, the auto-EQ does little anyway (plus it raises the noise floor slightly). If certain EQ or DSP effects do not work as desired for higher frequency signals, you may want to limit this to 44.1 or change your audio output to not allow 176.4 (so that upsampling only goes to 88.2).
Now. Onto CD's where this setting actually means "upsampling."
Whenever you resample an audio or video signal (especially with downsampling) you want to dither the signal to spread (and hopefully reduce) overall error. This is obviously needed with downsampling where you are inherently losing information. BUT, with upsampling, dithering can help reduce artifacts, distortion, and noise.
What do they do?
Type 1: Simple dithering. The result will reduce artifacts and potentially help distortion. It will, however, inherently add white noise to the signal increasing the noise floor. This can be helped, which is where Type 2 comes in.
Type 2: Noise shaping. This technique dithers the signal, but also quantizes error in a feedback loop, which is used to direct the dithering. The final step is to push the error to audio frequencies less perceptible by the human ear. The overall error will be the same as dithering, but the error will be less noticeable (to humans). This is how you can have a 24-bit remaster on CD. Noise shaping can effectively maintain the dynamic range and noise of a 24-bit signal for the 1kHz-4kHz range in a 16-bit signal.
Type 3: "Use a technique developed specifically for the PS3 system to enhance audio playback." Obviously a bit of marketing there. It may have been produced by the PS3 team within Sony, but it's just a dithering technique performed on an audio signal, which would work just the same for any device working with digital signals. I'm not sure what technique this is. Just as Type 2 is a more directed and "smarter" extension of Type 1, I'd assume that Type 3 either adds to what Type 2 does or perhaps is a different approach, but still based on Type 1 dithering.
So, Type 2 is certainly better than Type 1. Also, it is probably safe to assume that Type 3 is better than Type 2. If you are worried about the unknown, stick with Type 2.
In any case, we are dealing with upconversion here, where dithering is less important than downsampling. I would not be worried about any Type ruining your audio signal.
I am not sure if this setting comes into play for SACD playback. In that case, not only are you resampling, but going from a high frequency 1-bit DSD stream to a lower frequency multi-bit LPCM stream. Dithering would certainly be important there. However, the conversion is taking place between two very high quality streams, so error is actually rather minimal.
As for what I use: Type 3, upsample to 176.4.
Wow, that was a very useful reply!
I may be crazy, but after experimentaion I find that, when listening to very complex music (symphonies) that type 3 seems to smooth out too many details. Not sure why, or maybe I'm imagining it. (?)
Here is something I have thought about: if you produce a flac from a regular CD, and play that on your compu via an optical connection, might there be even less error induction than straight CD play?
It would be nice to be able to A/B Type 2 and 3, but with only 1 PS3, it takes time to switch. Having said that, I will have to do some listening between Type 2 and 3 now that you say that. Not knowing what Type 3 actually does, it could indeed smooth at least certain portions of audio signals.
Flac is lossless compression. If you rip flac from CD, you have the same bits (unless the CD is really scratched and/or is read or was burned at a high speed, in which case it is possible to read an incorrect signal).
Assuming a good/normal/playable CD and a proper flac encoding, the bits going out the optical connection (and the bits reaching the other end for that matter) will be identical to those coming from a CD.
If you compared that to "upconverted" CD playback from the PS3, yes you'd have different results, but the flac transcode alone will not really help anything.
A CD contains information to correct errors. Uncorrectable errors almost never occur.
You can argue timing alignment errors during digital to analog conversion (known as jitter), but having a computer or PS3 play the audio likely removes jitter from the disc end. As long as the audio information is buffered from the disc without loss, matching the digital output to proper clock timing should not be an issue. Using a more advanced connection than a simple stream, such as HDMI, likely also removes jitter from the transmission step. In any case, jitter has been found to vary more drastically between different masterings of CDs than it does between cheap players and ones costing thousands.
Wow! Very interesting. There really are some very smart people here on this forum. I guess it's the kind of crowd that Ascend attracts. :)Quote:
Originally Posted by Fantom
Which would sound better though? The PS3 or the flac via optical to the receiver? Which would be more accurate?
I just thought that the flac encoder would correct all errors, and so my sound card would output, via flac, an error-free version of the CD track. It seems to me though, that the sound output from my computer's optical out, on flacs, is just a tad shrill. Would a better computer card help? Is this common?
Playing a CD or ripping a CD to flac should sound EXACTLY the same.
Any CD player made to spec will "correct" errors in the exact same way. The extra information on the CD is used to make sure the incoming bits are correct. If they are not correct, that extra information can be used to make them correct.
If a CD is Red Book compliant, all the errors on the disc will be correctable, meaning the audio read from the disc will always be 100% accurate.
If an error cannot be corrected, then the software or device will interpolate the result. I am not positive on how this is done, but it is probably specifically defined in the specification. Of course, not all manufacturers follow every spec, but interpolating uncorrectable errors should happen the same on all players.
As I said, as long as the disc is in decent condition, there should be no uncorrectable errors at all. And even if there are, your computer software will make the same interpolation guesses before flac encode that the PS3 would make.
I guess one argument is that you could make sure the disc is really clean before encoding to flac, and then know that the flac encode will always give the same bits from that point on, whereas the CD could become scratched and dirty over time.
As far as how it sounds, the bits should remain the same. From CD, read by CD drive, into the flac encoder. Once decoded from the flac file, they should again be the same. Also, the digital bits going out the optical connection and the bits reaching the receiver/processor. So it should sound exactly the same.
You may have bass/treble, EQ, EAX environment, ac3filter settings messing with your audio before leaving the optical connection.
Make sure all settings are level or off if you can. Set the digital output settings to S/PDIF wherever you can.
Also, so as not to touch the signal in any way, all volume levels should be at max. Your player likely has a volume control that will actually change the data to make the audio quieter. Your computer's volume setting will also do this. Setting all of these volume levels to MAX will allow the audio stream to reach your processor/receiver untouched. Only adjusting volume at the final step (amplification) will lead to the most accurate audio.
Wow! Again, super-useful, hard to come by information.
Many, many thanks!:)