Converting DSD Files to FLAC

QuadraphonicQuad

Help Support QuadraphonicQuad:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.
So are you saying that on the fly conversion (as would be done in a disc player or AVR at the time of playback ) vs file conversion from DSD to PCM may yield different outcomes? And that the on the fly conversion would be the one that suffers?
Yes, I don't have a firm opinion yet about the conversion done by Foobar vs native DSD, I haven't done any comparisons. Since discovering the difference with the Oppo's analogue output I've decided to keep DSD as DSD wherever possible.

If you look in Foobars settings there are a number of different options for conversion. One setting is for the type of filtering to be applied, I suspect that is were degradation is happening Excessive filtering could possibly remove some of the audible signal causing a softening of the sound. Foobar will play .dsf files on the fly by converting to PCM, but again I haven't yet done any real listening comparisons that way, instead I drop the files to a flash drive and let the Oppo play them.
 
If you look in Foobars settings there are a number of different options for conversion.

This is the only screen I see in Foobar. I don't see any option for different filters. Is there another adjustment screen I don't know about?

See below
 
Last edited:
Sorry. I posted the wrong photo. This is the SACD screen. I still don't see filter options. I also dont know what the Mode selection is about. Can anyone provide an explanation. At the time of conversion there is also an option to add dither, which i never do.

1639932710576.png
 
Last edited:
Sorry. I posted the wrong photo. This is the SACD screen. I still don't see filter options. I also dont know what the Mode selection is about. Can anyone provide an explanation.

View attachment 74496
I notice you’re giving +4 db boost on the PCM vol.; I’ve been considering some on mine. Typically though, I just leave them as dsf. and use my volume knob. But if you’re shuffling from other sources, other stuff is too loud. Just not sure how much to give it. Maybe I could see what the difference is between the DVD-A & SACD of Goodbye Yellow Brick Road; but have to research if the mastering was the same on that one :unsure:
Does any of that make sense?
 
I notice you’re giving +4 db boost on the PCM vol.; I’ve been considering some on mine. Typically though, I just leave them as dsf. and use my volume knob. But if you’re shuffling from other sources, other stuff is too loud. Just not sure how much to give it. Maybe I could see what the difference is between the DVD-A & SACD of Goodbye Yellow Brick Road; but have to research if the mastering was the same on that one :unsure:
Does any of that make sense?
Click on Mode and you will see all the different filter options. Which one is the best, I haven't got a clue! If the level is set too high you will get clipping, if too low you may have to normalise after conversion. Options for sample rate as well, do higher rates produce better results? Most say that 88.2 is sufficient. So you see that there is a lot of room for experimentation within the conversion process.
 
I notice you’re giving +4 db boost on the PCM vol.; I’ve been considering some on mine. Typically though, I just leave them as dsf. and use my volume knob. But if you’re shuffling from other sources, other stuff is too loud. Just not sure how much to give it. Maybe I could see what the difference is between the DVD-A & SACD of Goodbye Yellow Brick Road; but have to research if the mastering was the same on that one :unsure:
Does any of that make sense?
I'm not even sure if that +4dB is the default or not, or where it came from. Mabe someone else suggested it. In practice for me though, i tag everything with replay gain, so it all plays back at similar percieved loudness.
 
Click on Mode and you will see all the different filter options. Which one is the best, I haven't got a clue! If the level is set too high you will get clipping, if too low you may have to normalise after conversion. Options for sample rate as well, do higher rates produce better results? Most say that 88.2 is sufficient. So you see that there is a lot of room for experimentation within the conversion process.
I'm not an expert (if anyone is please speak up), but the first two options under mode (32fp and 64fp) have something to do with the precision of the conversion, not the filter type AFAIK. I'm unsure of the rest of the options.

I have seen references to different filter options in DACs when they are tested in Audio Science Review forum, but i dont see those options in foobar.

At any rate this is leading to a conclusion that says the Oppo is doing something sub par when it makes a conversion from DSD to PCM at the time of playback.
 
I'm not an expert (if anyone is please speak up), but the first two options under mode (32fp and 64fp) have something to do with the precision of the conversion, not the filter type AFAIK. I'm unsure of the rest of the options.

I have seen references to different filter options in DACs when they are tested in Audio Science Review forum, but i dont see those options in foobar.

At any rate this is leading to a conclusion that says the Oppo is doing something sub par when it makes a conversion from DSD to PCM at the time of playback.
I suspect that you are correct that the Oppo is doing something sub par in the conversion process. That calls into question all other conversion processes as well. Here is an interesting page about such conversions. ANALYSIS: DSD-to-PCM 2015 - foobar SACD Plug-In, AuI ConverteR, noise & impulse response...
 
Click on Mode and you will see all the different filter options. Which one is the best, I haven't got a clue! If the level is set too high you will get clipping, if too low you may have to normalise after conversion. Options for sample rate as well, do higher rates produce better results? Most say that 88.2 is sufficient. So you see that there is a lot of room for experimentation within the conversion process.
I'm not even sure if that +4dB is the default or not, or where it came from. Mabe someone else suggested it. In practice for me though, i tag everything with replay gain, so it all plays back at similar percieved loudness.
I always go with 88.2 for mine.

I'll need to check my Foobar2K settings today and see where I'm at; but I believe the initial settings for PCM are zero.

Typically I just play dsf files from an SSD to the oppo; but sometimes a situation comes up where I convert to flac. Remember the pop that sometimes occurs on those Sony 7" SACDs at the front of the first song? Well to get rid of that I have to convert to flac so I can open in a DAW and mitigate the pop, as my DAWs wont open dsf.

As to the Elton GBYBR, I've got both the SACD and the DVD-A and thought maybe I could use that as a base line adjustment, and hear what happens.
 
That's looking like a deep dive there, but I'll get through it eventually.

It would be nice if we could formulate a SOP for these DSD-FLAC conversions. Maybe keep the final recipe as a sticky post up top.

I'm not willing to give up my DSD-PCM conversions. Being able to use room correction DSP far outweighs any arguably inaudible side effect of the conversion process. There is no mistake knowing if the EQ routine is being applied.

It looks like the +4 dB boost is a good choice based on the first page of this thread. It gets uglier for determining the LFE level due to a poor standard.

EDIT: by the way, the LFE issue even affects standard playback IIRC.
 
Last edited:
So are you saying that on the fly conversion (as would be done in a disc player or AVR at the time of playback ) vs file conversion from DSD to PCM may yield different outcomes? And that the on the fly conversion would be the one that suffers?
Absolutely! The on the fly conversion needs to be compromised if the full quality conversion would take longer than 1x real time. And these conversions do.
 
I remember hearing a couple DSD rips done incorrectly in the early days of messing with this. Output to not only 44.1k but also 16 bit! (Probably done with Foobar or some other Windows app.) Very very apparent distortions! So that's subjective... Just saying that I've heard gross damage with errant settings before.

Some conversion apps are a bit malicious compliant. "You set me to do what now?! Sure! Here ya go!"

You have to do significant damage to audio before you hear any fallout on the cheapness stuff sold at Worst Purchase, etc. Convenience wins. If you chase fidelity and sound that's free of generation loss it would behoove you to be critical with conversions with this stuff. DSD is really truly just as capable of full fidelity audio as PCM. It should absolutely be thrown away because it was a pathetic bullying attempt at competition with a different digital language for the same end result. But there's music out there and the format is full fidelity.

At the end of the day, you just might have better PCM DA converters than might even be available for DSD DA converters. The transcode to PCM will sound truer to the source on your system that native with lesser DSD converters.
 
Here's a link from a member that compared the GBYBR (single song anyway) for mastering, and feels (also seems to be a general consensus on the Poll thread) they are the same mastering on the SACD & DVD-A; especially since IINM there was the version with both formats in it.
https://www.quadraphonicquad.com/fo...-road-sacd-dvd-a-bda.3646/page-11#post-439965
So I'm going to pull my files for a song and see what the levels are in comparison.
 
Here's what I found for levels from the DVD-A 96kHz and the SACD 88.2 kHz (flac with 0 offset db conversion)
*this is just for the song GBYBR!

DVD-A
F - 0.5
C - 0.5
LFE - 0.5
RL - 0.5
RR - 0.5

SACD
F - 5.197
C - 6.329
LFE - 6.468
RL - 6.154
RR - 6.282

I'm just thinking now, that I see this, that my DVD-A rip may not be correct that I have, would everything be level matched at 0.5 dbs ???

Can someone confirm that those numbers are correct please on the DVD-A?

If so seems there may be a slight difference in mastering for the Front channels :unsure:
 
No sir, you are hearing expectation bias.
Both are capable of returning a perfect rendition of the mic's feed. The only real difference that may exist are being done in the mastering.
Wrong, wrong, wrong, read back to the Holy Cow thread. I never expected there to be any difference in sound so no expectation bias!!! I was blown away by the difference in sound that I found!! I was caught quite off guard by the sound quality!

And how many times do I have to repeat myself? Both PCM and DSD can produce excellent results, the problem appears to be in the conversion process from DSD to PCM!!!!!!
 
Last edited:
I never expected there to be any difference in sound so no expectation bias!!! I was blown away by the difference in sound that I found!! I was caught quite off guard by the sound quality!
Bias can be triggered in many ways you'd never expect. As long as you know what's playing it can rear it's ugly head.
 
Back
Top