Price, J., delivered the opinion of the Court in which Meyers, Mansfield, Holland, and Johnson, JJ., joined. Womack, J., filed a concurring opinion in which Keller and Keasler, JJ., joined. McCormick, P.J., dissented.

Appellee filed a motion to suppress evidence that alleged no probable cause for his arrest. At the hearing on the motion, the only witness to testify was the arresting Texas Alcoholic Beverage Commission (TABC) agent. Following the testimony, the judge granted the motion and did not file findings of fact. On appeal, the Court of Appeals found that the agent’s testimony, if believed, showed reasonable suspicion for the initial detention and probable cause for the eventual arrest.¹ However, the Court of Appeals held that the trial court, sitting as trier of fact, was the sole judge of credibility and could disbelieve any or all of the agent’s testimony.² The Court then applied the “almost total deference” standard of review and affirmed the trial court.³ We granted review to determine whether “an appellate court may uphold a trial court’s decision to suppress evidence as within its discretion, instead of de novo, because the trial court might have disbelieved some or all of the State’s uncontroverted evidence.” Because the trial court is the sole judge of credibility of the witnesses and we will uphold a trial court’s ruling on any theory of law applicable to the case, and because this involves an application of law to facts, the outcome of which turns on the credibility and demeanor of the witness, we will affirm the Court of Appeals.

BACKGROUND

Agent Darnell testified to the following events at the motion to suppress hearing. At approximately 12:50 a.m., on November 2, 1997, Agent Darnell and two other TABC agents finished inspecting a local bar. When they walked out the front door, Agent Darnell noticed the windows of a nearby pickup truck were fogged up. He looked inside the truck and saw two young children sleeping, without any blankets or coverings, in the front seat. Because the night was cold and someone could have broken into the truck, Agent Darnell sent another agent into the bar to find the owner of the pickup.⁴ When the owner of the pickup, Ross, came out, Agent Darnell noticed that he held on to things for balance and support, spoke with a thick tongue, had bloodshot eyes, and his breath smelled of alcohol. Agent Darnell asked appellee how he was going to get the children home, and Ross replied that he was going to drive them home. Agent Darnell then arrested appellee for public intoxication.

Ross filed a motion to suppress all evidence surrounding his arrest due to a lack of probable cause. At the hearing on the motion, Agent Darnell was the only witness. Ross presented no evidence himself, although Agent Darnell was cross-examined. The trial court granted the motion to suppress without entering findings of fact or conclusions of law.⁵ The State appealed.

The Court of Appeals affirmed, holding that it was within the trial court’s discretion to grant the motion to suppress if the trial court did not believe any material portion of the State’s evidence.⁶ However, the Court noted an apparent conflict in precedent on the standard of review for suppression hearings. Under Guzman⁷ and Maestas,⁸ mixed questions of law and fact that do not turn on an evaluation of credibility and demeanor are reviewed de novo.⁹ A footnote in Maestas indicates that an issue does not turn on an evaluation of credibility and demeanor where the evidence is uncontroverted because the trial court does not have to decide which conflicting testimony deserves more weight.¹⁰ The Court of Appeals found that the State’s evidence was uncontroverted and showed both reasonable suspicion for appellee’s initial detention and probable cause for his arrest.¹¹ Because granting the motion on the grounds that the facts simply did not add up to probable cause would be a misapplication of the law to the facts, the Court of Appeals reasoned that the trial court sustained the motion based on the credibility and demeanor of the witness.¹² Since the trial court was the sole judge of credibility, it was within the discretion of the trial court to accept or reject Agent Darnell’s testimony and therefore within the discretion of the trial court to grant the motion to suppress.¹³

In its petition to this Court, the State argues that there is no indication that the trial court actually disbelieved the testimony of the Agent Darnell. As such, the appellate court should assume the trial court found the facts testified to were legally insufficient to show reasonable suspicion and probable cause. Because the evidence was sufficient to prove probable cause, the State argues that the trial court should be reversed under a de novo review. Furthermore, the State argues that if we affirm the Court of Appeals’s decision, we will create a way for trial courts to completely insulate their motion to suppress rulings, and thus limit the State’s right to appeal under article 44.01(a)(5) of the Code of Criminal Procedure.

DISCUSSION

In a motion to suppress hearing, the trial court is the sole trier of fact and judge of the credibility of the witnesses and the weight to be given their testimony.¹⁴ Accordingly, the judge may believe or disbelieve all or any part of a witness's testimony,¹⁵ even if that testimony is not controverted.¹⁶ This is so because it is the trial court that observes first hand the demeanor and appearance of a witness, as opposed to an appellate court which can only read an impersonal record.¹⁷

Furthermore, when the trial court fails to file findings of fact, we view the evidence in the light most favorable to the trial court’s ruling and assume that the trial court made implicit findings of fact that support its ruling as long as those findings are supported by the record.¹⁸ If the trial judge's decision is correct on any theory of law applicable to the case, the decision will be sustained.¹⁹

While discussing the appropriate level of appellate review, we stated in Guzman v. State:

[1] [T]he appellate courts, including this Court, should afford almost total deference to a trial court’s determination of the historical facts that the record supports especially when the trial court’s fact findings are based on an evaluation of credibility and demeanor. [2] The appellate court should afford the same amount of deference to trial courts’ rulings on “application of law to fact questions,” also known as “mixed questions of law and fact,” if the resolution of those ultimate questions turns on an evaluation of credibility and demeanor. [3] The appellate courts may review de novo “mixed questions of law and fact” not falling within this category.²⁰ [citations omitted]

As we pointed out in Carmouche, the typical motion to suppress case alleging a lack of probable cause will be reviewed through a bifurcated standard of review combining the first and third categories of Guzman: we will give almost total deference to a trial court’s express or implied determination of historical facts and review de novo the court’s application of the law of search and seizure to those facts.²¹ This is because in most instances, the appellate court will have facts established by the trial court, either impliedly or expressly, to which the law may be applied.²² But, not every probable cause case will be reviewed under this bifurcated standard.

In a motion to suppress hearing where the only evidence presented is the testimony of the arresting officer (which, if believed, adds up to probable cause) and the trial court grants the motion without any explanation, there is not necessarily a “concrete” set of facts that can be implied from such a ruling. The trial court may have disbelieved the officer on at least one material fact, or the trial court may be in a situation in which it does not know what exactly the facts are, but it does know (on the basis of demeanor, appearance, and credibility) that they are not as the witness describes. In this situation, the appellate court does not necessarily have a set of historical facts to which it may apply the law. The determination of probable cause rests entirely on the credibility of the lone witness. This scenario is a mixed question of law and fact, the resolution of which turns on an evaluation of credibility and demeanor. The proper standard of review is therefore the second category of Guzman, “almost total deference” to the trial ruling.²³

In applying these principles to the case at bar, we will review the trial court’s ruling under the “almost total deference” standard. As no findings of fact were filed, we must view the evidence in the light most favorable to the trial court’s ruling and will uphold the ruling on any theory of law applicable to the case.²⁴

We see two possible theories upon which the trial court’s ruling may rest, either the testimony of the agent was credible, but the facts established by that testimony do not constitute probable cause (or reasonable suspicion for the initial detention); or the trial court did not find the testimony of the agent to be credible. Because we agree with the Court of Appeals’s conclusion that Agent Darnell’s testimony (if believed) added up to reasonable suspicion for the initial stop and probable cause for the eventual arrest, we will not presume that the trial court impliedly found that the facts simply did not add up to reasonable suspicion or probable cause; to do so would be to presume error. The trial court, however, was free to disbelieve all of the agent’s testimony. As the sole trier of fact and judge of credibility, the trial court was not compelled to believe the agent’s testimony, even if uncontroverted, based on credibility and demeanor. Because the evidence, if believed, would compel a denial of the motion to suppress, the record supports the second theory that the trial court did not find the agent’s testimony to be credible based on demeanor, appearance, and tone.²⁵

“Given the absence of any factual findings, the appellate presumption of the regularity of a trial court’s judgment, and which party had the burden of proof, the trial court’s implied factual findings were dispositive of the ultimate constitutional question of probable cause.”²⁶ We hold that the trial court was within its discretion in disbelieving the testimony of the agent and granting the motion to suppress. We further find that this case falls within the second category of review under Guzman and affirm the trial court’s use of the “almost total deference” standard of review.

Before addressing the State’s argument, we pause to address the apparent conflict as noted by the Court of Appeals concerning Maestas.²⁷ Maestas involved a defendant’s motion to suppress a confession alleging a violation of the Fifth Amendment.²⁸ The officers’ testimony was uncontroverted and the trial court denied the motion to suppress.²⁹ It appears that findings of fact were not filed. We found that a de novo review was appropriate as the resolution of the issue did not turn on an evaluation of credibility and demeanor.³⁰ In footnote eight, we said this was because the officers’ testimony as to what transpired was uncontroverted.³¹ We stated, “Under different circumstances, where the police and the appellant provide conflicting testimony, resolution of this question would involve an evaluation of credibility and demeanor because the trial court would have to decide which testimony deserved more weight.”³²

Footnote eight is a correct statement of the law in that when faced with conflicting testimony, the trial judge will have to determine which testimony is the more credible. However, a de novo review in Maestas was appropriate because the dispositive facts were uncontroverted and there was no indication that the trial court did not believe that testimony. In other words, the trial court’s ruling was consistent with the uncontroverted testimony. Because it suggests that an issue turns on credibility and demeanor only when the testimony is controverted, we expressly disavow footnote eight of Maestas.³³ As former Presiding Judge Onion recently explained:

Nothing in Guzman [or its progeny] should be interpreted as catapulting an appellate court into the seat of a second trier of fact determining the credibility and demeanor of witnesses, in whole or in part, from a cold record. This is true whether the resolution of the ultimate question falls within the second or third category of Guzman or whether the standard of review is . . . an abuse of discretion or a de novo review.³⁴

Having determined that the Court of Appeals’s resolution of this case was correct, we address the State’s arguments. The State argues that because the testimony was uncontroverted and sufficient to show probable cause, and because there is nothing in the record that indicates that the trial court “actually” disbelieved the testimony, reversal of the trial court through a de novo review is proper.³⁵ We decline to do this.

The theory that the trial court disbelieved the agent’s testimony, either in whole or in part, supports the ruling of the trial court. The State is asking us to disregard the theory that supports the trial court and adopt the theory that results in error and leads to reversal. We will not adopt a rule that presumes error when there is a viable legal theory that supports the trial court’s ruling.

Finally, the State argues that by not conducting a de novo review in this case, we will be limiting appeals under article 44.01(a)(5),³⁶ and give trial courts a way to insulate their rulings on motions to suppress from review by merely not filing findings of fact. We do not accept these conclusions.

Our holding does nothing to limit the State’s right to appeal under article 44.01(a)(5). The State is just as free to make the appeal today as it was before the decision today. If a non- prevailing party wishes to avoid the effects of these appellate presumptions, then it should attempt to get the rationale for the trial court’s ruling on the record through either a verbal explanation at the hearing or express findings of fact and conclusions of law.

CONCLUSION

In most motion to suppress cases where probable cause is challenged, there will be historical facts found, either expressly or impliedly, to which a Guzman/Carmouche style de novo review will be applied. It is only in cases, like this one, when no findings of fact are filed, the uncontroverted testimony of the witness adds up to probable cause if believed, and the motion is nevertheless granted, will Guzman’s second category of review be applied. In other words, the resolution of the issue in such cases turns completely on the credibility of the witness.

The judgment of the Court of Appeals is affirmed.

Delivered November 29, 2000.

Publish.

¹See State v. Ross, 999 S.W.2d 468, 472-73 (Tex. App.–Houston [14th Dist.] 1999).

²See id. at 470.

³See id. at 474.

⁴Agent Darnell testified that he was concerned that the offense of endangering a child was being committed.

⁵The trial court gave no indication as to the basis for its ruling. The transcript merely reads, “The Court grants the motion to suppress.” Ct. R. at 36.

⁶See Ross, 999 S.W.2d at 474.

⁷Guzman v. State, 955 S.W.2d 85, 89 (Tex. Crim. App. 1997).

⁸Maestas v. State, 987 S.W.2d 59, 62 n.8 (Tex. Crim. App. 1999).

⁹See Guzman, 955 S.W.2d at 89.

¹⁰See Maestas, 987 S.W.2d at 62 n.8.

¹¹See Ross, 999 S.W.2d at 471-73.

¹²See id. at 473-74.

¹³See id.

¹⁴See State v. Ballard, 987 S.W.2d 889, 891 (Tex. Crim. App. 1999).

¹⁵See Garcia v. State, 15 S.W.3d 533, 535 (Tex. Crim. App. 2000); Allridge v. State, 850 S.W.2d 471, 492 (Tex. Crim. App. 1991); Meek v. State, 790 S.W.2d 618, 620 (Tex. Crim. App. 1990). See also, State v. Fecci, 9 S.W.3d 212, 221 (Tex. App.–San Antonio 1999, no pet. h.).

¹⁶See Mattias v. State, 731 S.W.2d 936, 940 (Tex. Crim. App. 1987).

¹⁷See Garcia, 15 S.W.3d at 535; Villarreal v. State, 935 S.W.2d 134, 138 (Tex. Crim. App. 1996); Romero v. State, 800 S.W.2d 539, 543 (Tex. Crim. App. 1990).

¹⁸See Carmouche v. State, 10 S.W.3d 323, 328 (Tex. Crim. App. 2000).

¹⁹See Romero, 800 S.W.2d at 543.

²⁰955 S.W.2d 85, 89 (Tex. Crim. App. 1997).

²¹See Carmouche, 10 S.W.3d at 327; Guzman, 955 S.W.2d 88 n.3 & 90-91.

²²For example, even if there is only the testimony of the officer, which adds up to probable cause if believed, and the trial court denies the motion without express findings of fact, there will be two implied fact findings: the officer’s testimony was credible and the facts are as the officer testified. See e.g., Guzman, 955 S.W.2d at 88 n.3 & 90-91.

²³See Loserth v. State, 963 S.W.2d 770, 772-73 (Tex. Crim. App. 1998); Guzman, 955 S.W.2d at 89 (“Our decision is also meant to reaffirm the longstanding rule that appellate courts should show almost total deference to a trial court’s findings of fact, especially when those findings are based on an evaluation of credibility and demeanor . . .”).

²⁴See Romero, 800 S.W.2d at 542.

²⁵We reiterate that the nature of demeanor and credibility is not such that it would readily appear in the record. Demeanor and credibility assessments are based in large part on visual and audio observations and are not usually found in the record. A presence or absence of contradiction or inconsistency in the record is not therefore dispositive.

²⁶Villarreal v. State, 935 S.W.2d 134, 140 (Tex. Crim. App. 1996).

²⁷987 S.W.2d 59 (Tex. Crim. App. 1999).

²⁸See id. at 61.

²⁹See id. at 61 & 62 n.8.

³⁰See id. at 62.

³¹See id. at 62 n.8.

³²Id. In its entirety, the footnote reads:

This is because the officers' testimony as to what transpired from Appellant's arrest up to Appellant's confession is uncontroverted. Appellant did not provide conflicting testimony or contradict the officers' presentation of the facts in any way. In addition, Appellant did not present any complaints, for example complaints concerning the conditions of her jail cell, or complaints that she was coerced, intimidated, or forced to speak with police. Under different circumstances, where the police and the appellant provided conflicting testimony, resolution of this question would involve an evaluation of credibility and demeanor because the trial court would have to decide which testimony deserved more weight. In that case, our review would “afford almost total deference” to the trial court's determination, in accordance with Guzman.

³³We note that in Oles v. State, 993 S.W.2d 103, 106 (Tex. Crim. App. 1999), we stated that “the instant case presents us with a question of law based on undisputed facts, thus we apply a de novo review.” Oles involved a motion to suppress test results from clothing. The defendant was lawfully arrested but his clothing was tested for an unrelated crime and there was no visible evidence on the clothing that would link the defendant to the second crime. The trial court denied the motion. The testimony was uncontroverted and the trial court’s ruling was not contrary to the testimony. Oles simply did not state the implied premise that the trial court found the testimony to be credible. Because the trial court found the testimony credible, it made its decision based on the facts testified to. As this gave the appellate court a “concrete” set of facts, through the presumption that arises when no findings of fact are filed, the appellate court could do a de novo review.

³⁴State v. Fecci, 9 S.W3d 212, 222 (Tex. App.–San Antonio 1999, no pet. h.)

³⁵State’s Brief on the Merits at 5 (“The Court of Appeals should therefore have assumed that the trial court believed the State’s evidence, conducted a de novo review, and held that the trial court’s granting of Ross’s motion to suppress could only have resulted from misapplying the law to the facts.” ).

³⁶Tex. Code Crim. Proc. Ann. art. 44.01(a)(5) (Vernon 2000).